Author index

Aarts, R.

• Development of Dutch sound locators to detect airplanes (1927 - 1940)
  A. van der Voort and R. Aarts [I]
[A]

From the first world war until the 30's air acoustics played an important role in the air defence. Air vehicles carrying a weapon could not be located from the ground e.g. at night time or under cloudy conditions. As radar was still to be discovered, vision had to be supplemented by hearing using the sound of the engines. At least a dozen different acoustic location equipment's for aeroplanes from different countries were in the 1920's available on the military market. They served as said as supplement to visual means preceding the use of artillery. These equipments contained most times four sound pick-up elements such as curved shells or horns used in pairs. Two elements at a fixed mutual distance were adjusted in the horizontal plane for the determination of the chart angle and the remaining two in the vertical plane for the measurement of elevation. The transport of sound from each element of a pair to a corresponding pair of human ears was carried out by means of metal or rubber tubes. Each pair is adjusted in the correct direction when the two ears receive the sound signals at the same time.

• Estimating the velocity profile and acoustical quantities of a harmonically vibrating membrane from on-axis pressure data
  R. Aarts and A.J.E.M. Janssen [I]
[A]

Power series expansions are derived for acoustical quantities of a harmonically excited resilient, flat, circular radiator in an infinite baffle. These quantities are the sound pressure: on-axis, far-field, and at the edge of a radiator, as well as the reaction force on the radiator and the total radiated power. These expansions are obtained by expanding the velocity distributions in terms of orthogonal polynomials. For the rigid, simply supported and clamped radiators, this results in explicit, finite-series expressions involving (spherical) Bessel functions for both the on-axis and far-field pressure. In the reverse direction, a method of estimating velocity distributions from (measured) on-axis pressures by matching in terms of expansion coefficients is proposed. Together with the forward far-field computation scheme, this yields a method for assessment of loudspeakers in the far-field and of the total radiated power from (relatively near-field) on-axis data (generalized Keele schema). The relation between the radiated power as obtained from near-field measured sound pressure data via the theory on one hand and from a far-field measurement in a reverberation room on the other is highlighted.

Abdennadher, S.

• Survey on common Arabic Language forms from a speech recognition point of view
  M. Elmahdy, R. Gruhn, W. Minker and S. Abdennadher [C]
[A]

Arabic language is the largest still living Semitic language based on the number of speakers which exceeds 250 million first language speakers. Arabic language has two main forms: Standard Arabic and Dialectal Arabic. Standard Arabic includes ''Classical Arabic'' and ''Modern Standard Arabic (MSA)'' while dialectal Arabic includes all forms of currently spoken Arabic in day life and it vary among countries and deviate from standard Arabic to some extent. While there are many forms of Arabic, there still many common characteristics on the acoustic level and the language level. In this paper we investigate the main differences between Arabic forms from a speech recognition point of view to focus on similar and different acoustic properties among Arabic forms as well as the language properties. For standard Arabic we have chosen both classical and MSA forms and for dialectal Arabic we have chosen Egyptian colloquial Arabic as an example since it is the most popular dialect among Arabic speakers. The main purpose of this paper is to summarize all characteristics of Arabic and previous research results in Arabic speech recognition from a speech recognition point of view in one paper so researchers in this field can use it as one reference.

Abel, M.

• Nonlinearities in wind instruments at the example of organ pipes
  M. Abel and K. Ahnert [I]
[A]

Nonlinear effects are typical for sound generation, let it be the aeroacoustical origin as in wind instruments, complicated stick-slip motion with string instruments or the dependence of body vibrations on the excitation amplitude. We demonstrate the intricate interaction of nonlinearities by the synchronisation properties of organ pipes by a high-accuracy experiment where a loudspeaker drives an organ pipe and find that even for minute speaker intensity the organ pipe still follows this control mechanism. Because the principle is very general, the consequences for other instruments and music has to be considered.

Ablitt, J.

• Temporal and Spectral Characteristics of a Marine Piling Operation in Shallow Water
  P. Lepper, S. Robinson, J. Ablitt and S. Dible [I]
[A]

Analysis of the underwater radiate acoustic characteristics for marine piling operations for two pile diameters, 2m and 4.74m, in a relatively shallow water site are presented. Measurements of the entire piling sequence for several piles were conducted at ranges from 10m to 22km for piles in 10-20 m water depth. Variations in the temporal and spectral characteristics of radiated energy are analysed in context of pile size, range from source, hammer energy used and pile penetration depth.

The hammer energy used shows a strong interdependence between mechanical strike 'hammer' energy and underwater radiated acoustic energy. This process appears 'coarsely' linear for individual piling operations although considerable variation in overall gradient were observed between operations. Temporal and spectral variations in radiated energy due to pile penetration are also examined for fixed hammer energy and range.

Simultaneous recordings of radiated energy made at increasing distances from the pile showed evidence of temporal and spectral dispersion effects consistent with relatively shallow water propagation. Correlation of received levels at various ranges in differing seabed topographies were made suggesting complex shallow water modal propagation dependant on both the source and environment characteristics including seabed topography, sediment type and water column acoustic properties.

Abshagen, J.

• Estimation of Underwater Flow Noise by Wave Number Decomposition
  J. Abshagen [C]
[A]

Flow noise limits the capability of a SONAR system operated on any moving platform at sea. Its physical origin differs qualitatively from other underwater noise sources, such as from ambient noise, yet it is difficult to measure. Flow noise results from pressure fluctuations in the turbulent boundary layer which are generated by the movement of the SONAR platform relative to the surrounding sea. In particular at higher frequencies this noise source differs in wave number distribution from sources which originate in wave propagation processes, e.g. from ambient noise. Measurements of a wave number distribution require an array of spatially separated hydrophones. A method based on wave number-frequency analysis is presented that allows to discriminate between flow noise and noisy sound sources. It therefore enables, for example, estimations of frequency dependencies of underwater flow noise. The method is validated by flow noise measurements performed with an equally spaced line array mounted on a towed system.

Agnani, A.

• Vibroacoustic characterization of flexible hoses for air conditioning systems
  A. Agnani, F. Angelini, M. Martarelli and E.P. Tomasini [I]
[A]

Flexible tubes covered with rubber are the only way to damp vibrations and noise coming from compressor in air conditioning systems. Usually they have, as core, a corrugated steel pipe, for increasing the hose stiffness, this causing fluid turbulences generating vortex separation that gives a beginning of a audible and bothersome high frequency whistle (from 2500 Hz to 5000 Hz). The vibroacoustic analysis consists on the experimental and numerical (by Finite Element Methods) characterization of the hose vibrational behaviour and consequent acoustic characterization. The experimental measurement of the hose vibration has been performed in both controlled and operating conditions, the latter with the hose installed in the test bench including all the refrigeration cycle components. The acoustic characterization has been performed via propagation of the measured surface vibration in Boundary Element Codes and validated by means of acoustic intensity measurement. The acoustic measurement in operating conditions is very challenging, being the acoustic emission of the hose comparable to the total emission of the refrigeration bench (in primis the compressor), and therefore the hose must be isolated from the rest of the bench. Consequently the possibility of predicting the acoustic emission via BEM allows to avoid this kind of complicated measurement.

Ahlefeldt, T.

• Broadband Simulation Of Flow-Induced Noise Generation in Air Conditioning Ducts
  C. Spehr, R. Ewert, T. Ahlefeldt and J. Delfs [C]
[A]

Air conditioning systems in cabins are possible noise sources influencing the passenger comfort in civil aircrafts. The current study continues an initiative to predict the generated noise on components of air distribution systems. The approach presented here is based on the numerical inexpensive RANS-simulation. Using a perturbation technique the turbulent flow fluctuations are represented by artificial turbulences injected into the mean flow field. In the first part of the study the time dependent perturbation was induced by injecting individual artificial vortices which allow qualitative comparison of the sound generation of different arrangements on the basis of the overall sound power level. In the current work a broadband turbulence is induced in the source region. The turbulence is thereby generated by filtering white noise to reproduce the autocorrelations and integral length-scales provided by a RANS-simulation. The interaction of the turbulence with AC-components generates pressure fluctuations with broadband character which are compared with corresponding measurements. Arrangements of two consecutive orifices were investigated as typical components in the air conditioning system. The sound generation on the downstream orifices depends therefore not only on its shape, but also on the flow disturbance, which in turn depends on the distance to the upstream orifice.

Ahlersmeyer, T.

• Transfer Path Analysis - a Review of 18 years of Practical Application
  T. Ahlersmeyer [I]
[A]

Transfer Path Analysis (TPA) is a methodology to evaluate the contribution of individual transfer paths to a noise or vibration (NVH) problem. This is a first step to understand to root cause of a complex NVH problem. With the advent of affordable multi-channel digital (ADC) data acquisition systems the TPA has changed from simple detaching tests some 20 years ago to a structured process. However, in the last 15 years the inherent problems of the classical approach (exact force and transfer function determination) have not yet been overcome. Instead, shorter development cycles and shorter availability of less prototypes demand shorter test and analysis cycles and thus streamlining of the TPA approach. In the same time period the structural stiffness of passenger cars has increased significantly which results in higher crosstalk between the paths thus making a separation more difficult. The paper gives an overview how this has impacted the practical application of test-based TPA in the vehicle development of FORD in the last 15 years. It will also give an outlook on and first experiences with new approaches like TPA in time domain and estimation of the transfer matrix by using operational data only.

Ahnert, K.

• Nonlinearities in wind instruments at the example of organ pipes
  M. Abel and K. Ahnert [I]
[A]

Nonlinear effects are typical for sound generation, let it be the aeroacoustical origin as in wind instruments, complicated stick-slip motion with string instruments or the dependence of body vibrations on the excitation amplitude. We demonstrate the intricate interaction of nonlinearities by the synchronisation properties of organ pipes by a high-accuracy experiment where a loudspeaker drives an organ pipe and find that even for minute speaker intensity the organ pipe still follows this control mechanism. Because the principle is very general, the consequences for other instruments and music has to be considered.

Ahrens, J.

• Timbral Coloration in High Resolution Sound Field Reproduction Due to Spatial Bandwidth Limitation
  J. Ahrens and S. Spors [I]
[A]

Sound field reproduction methods based on orthogonal expansions like e.g. higher order Ambisonics always introduce a limitation of the spatial bandwidth of the loudspeaker driving signals. As a consequence, the desired component of the reproduced wave field is similarly bandlimited. This leads to a pronounced sweet area around the centre of the loudspeaker distribution. This area is sweet both in terms of spatial aliasing artefacts as well as in terms of accuracy of the desired component of the reproduced wave field. The sweet area gets significantly smaller with increasing temporal frequency. This means that outside the centre of the loudspeaker distribution, the sweetest point, frequency dependent artefacts arise. In this paper, we objectively investigate the resulting artefacts for circular loudspeaker arrangements with special attention to timbral coloration.

• Binaural Monitoring of Massive Multichannel Sound Reproduction Systems Using Model-Based Rendering
  M. Geier, J. Ahrens and S. Spors [I]
[A]

In model-based spatial sound reproduction a number of input signals is positioned in a virtual space according to a scene description. Using soundfield reproduction methods like Wave Field Synthesis or Ambisonics a scene can be rendered with a loudspeaker array. Suitable loudspeaker arrays typically consist of a huge number of loudspeakers and are often installed in heavily booked venues.

During the creation of an audio scene for playback in such a system it is desirable to listen to a representation of the scene via headphones without needing access to the actual loudspeaker system. This can save production resources because the target system is only needed for final adjustments. Alternatively, the same scene can also be monitored in a studio with much fewer loudspeakers using Vector Base Amplitude Panning.

We present different methods for reproducing a given audio scene via headphones, either by binaural simulation of a loudspeaker-based (possibly virtual) system or by binaural rendering of the original scene. Binaural simulation can be implemented efficiently by combining the loudspeakers' driving functions with Head Related Transfer Functions.

• Spatial Sampling Artifacts of Focused Sources in Wave Field Synthesis
  S. Spors and J. Ahrens [C]
[A]

Wave field synthesis (WFS) is a spatial sound reproduction technique that facilitates a high number of loudspeakers to create a virtual auditory scene. Amongst other interesting properties, WFS allows to reproduce virtual sources that can be positioned in the area between the loudspeakers and the listener. These are known as focussed sources. The physical properties of focussed sources have not been investigated in detail so far. Of special interest in the context of this contribution is the influence of spatial sampling. It will be shown that the resulting spatial sampling artifacts can be perceived as pre-echo artifacts. Their audibility depends on the source and listener position, and the size of the WFS system. This paper will discuss the physical foundations and will present first subjective experiments to investigate the audability of the resulting artifacts.

Al Moubayed, S.

• Noise reduction for media streams
  T. Bisitz, T. Herzke, M. Zokoll, A.-M. Öster, S. Al Moubayed, B. Granström, E. Ormel, N. van Son and R. Tanke [C]
[A]

In the Hearing at Home project the central Home Information and Communication platform is developed to support hearing-impaired persons with easy to configure ``Supportive Audio Signal Processing'' and visual support on a TV screen in home environments. Amongst other, individually configurable algorithms, several algorithms can be used for noise reduction of the media streams without individual fitting to enhance the accessibility of the presented audio materials using the Master Hearing Aid as a real-time audio-signal-processing framework: (1) Single-channel noise reduction according to Ephraim-Malah together with a speech pause detector, (2) stereo noise reduction additionally removing the central signal in a control path, (3) frequency shaping noise reduction by emphasising a frequency range important for speech intelligibility, (4) time-varying attenuation and (5) noise reduction by selection of channels. It will be possible to classify the acoustic situation in the media stream automatically in real-time using six classes allowing to activate the appropriate noise reduction strategies. In user tests the performance of algorithm 1-3 is evaluated: Results of sound quality assessment, listening effort estimation and speech intelligibility measurements with hearing impaired subjects from groups with different types of hearing impairment will be presented.

Albrecht, S.

• On the Inverse Music Sequencer Operation - Detection of Music Components from Wave Table in a Complex Music Signal
  S. Albrecht [C]
[A]

We made an attempt to a novel approach in automatic music transcription (i.e., in detection of pitch, loudness and timing of all sound events in a complex music signal automatically) working without any constraints on the observed music signal in general. It follows the reverse working of music sequencers. That is, we have a wave-table comprising arbitrary sounds as drums, melodic sounds, etc. Given the wave-table with sounds and a piece of a complex music signal for analysis, the sounds in the wave-table (or their modifications) are identified in the music signal. The sound events (as component IDs, their position in time and their truncation) are the output of the identification process. When we try to put what was identified into the track, we should obtain the same or rather similar song up to some point. The core part of the algorithm is the sequential Monte Carlo method. Tests of functionality will be demonstrated.

Algermissen, S.

• Reduction of Interior Noise in an Automobile Passenger Compartment by Means of Active Structural Acoustic Control (ASAC)
  M. Misol, S. Algermissen, H.P. Monner and A. Naake [C]
[A]

Results of the latest research from the automotive industry corroborate the assumption that windshield vibrations occurring in normal driving mode are a crucial factor for interior noise in automobiles. Due to the lack of passive damping methods that are compliant with the objectives of automotive lightweight construction and the specific requirements of the system under consideration, only active methods can be used in order to appropriately address the low frequency noise. Thus the research presented here has been focused on the development of a structurally integrated and compliant ASAC-system for the active reduction of the sound pressure level (SPL) inside the car's passenger cabin. In this work both adaptive feedforward and optimal state-feedback controllers have been designed and experimentally validated on a medium-class car. The use of highly efficient and global system identification methods based on structural sensors and laser-scanning vibrometer measurements allowed the definition of a global performance index for the controller design. In accordance with the numerical predictions on control performance the experimental results showed global reductions in vibration level up to 7dB and reductions in SPL up to 15dB.

Alizadeh, M.

• Bubble-Shock Wave Interaction
  M. Alizadeh, T. Kurz, D. Kröninger and W. Lauterborn [C]
[A]

The interaction of shock waves with a single bubble is studied by means of high speed photography and pressure measurements. The bubbles are produced by optical breakdown using a Q-switched Nd:YAG laser. A piezoelectric shock wave generator is applied as the source of the shock waves. The focus of the shock waves is located at the position of the bubble. In order to capture the very fast bubble dynamics with high resolution, a high-speed camera attached to a long distance microscope is used. The pressure field is measured with a Fiber Optic Hydrophone. Image sequences of the bubble dynamics are obtained at different delay times between the breakdown and the arrival time of the shock at the bubble. They show a reduction of collapse time by the shock wave which is more pronounced when the shock hits the bubble at a time when the bubble is already collapsing. The shock wave produced by the collapse itself is enhanced compared to the case of free oscillation. The shock leads to aspherical bubble collapse. A jet is generated in the direction opposite to the shock front which is stronger when the shock impinges the bubble in its collapse phase.

Alles, E.J.

• Medical acoustical array expertise at Delft University of Technology
  K.W.A. van Dongen, E.J. Alles and L. Demi [I]
[A]

The history of medical ultrasound dates back as far as the mid 30s, first as a therapeutic tool (e.g. deep tissue heating) and a decade later as a diagnostic tool based on pulse-echo reflections. Nowadays, a majority of all medical imaging modalities are based on ultrasound as the technique is often non-invasive, harmless and relatively cheap and still results in high image quality.

High image quality is obtained using array technology, i.e. a large number of active elements (preferably smaller than a wavelength) are positioned either in a single line (1-D array) or in a surface (2-D array). By applying a predefined time delay to the signal of each element, electronically or via software on the measured data, it becomes feasible to focus or steer the beam to a volumetric region of interest.

At the TU Delft in the section Acoustical Imaging and Sound Control we apply array technology in various medical applications. Among the applications are breast cancer detection, hyperthermia breast cancer treatment, transesophageal echo cardiography (TEE) and intravascular ultrasound. During our presentation, we will show for each of the above applications the challenges (and solutions) of implementing array technology in clinical applications.

• Simulation of Large Spatial Domain Ultrasound Scattering Problems
  E.J. Alles and K.W.A. van Dongen [C]
[A]

Currently, many ultrasound simulations of medical imaging modalities are performed using the Field II-software. This software models a scattering body as a collection of point-scatterers and evaluates the resulting total pressure field within the Born approximation. Consequently, it assumes a constant speed of sound throughout the volume. Therefore, modelled reflections arrive at incorrect times, and errors are introduced if thickness measurements are performed. In addition, multiple scattering is neglected in this approximation, which means that for strong contrasts results obtained with Field II are incomplete.

To circumvent these issues, we have developed software which efficiently computes the total pressure field within a volume due to scattering off contrasts of arbitrary shape and magnitude. The software uses a conjugate gradient scheme to solve the scatter integral equation for the total field, given known contrasts and incident field. Spatial domain decomposition is used to allow for large domains of which the memory load would otherwise exceed the amount of memory present in current machines.

During our presentation, we will compare the simulated received signals, using both programs, for an acoustic probe measuring plaque thickness in an array.

Alphei, H.

• About Speech Intelligibility depending on different Sound Insulations (part 1)
  T. Hils and H. Alphei [I]
[A]

In the course of the development of the DAGA recommendation ''Schallschutz im Wohungsbau - Schallschutzausweis'' (sound protection in residential construction - certificate of noise protection), in a former study theoretical considerations were made on the basis of the calculated loudness of transferring noises that served, among other things, as a basis for the predefinition of the degree levels in the different classes of noise protection.

The recommendation also includes a list of verbal descriptions of the subjective perception of (typical) sounds in housing space and speech from neighbouring living areas. In another list the perception descriptions are linked with the levels of noise protection for giving an orienting evaluation of sound transfer in the seperate classes of noise protection. However, the application for the planning in practice is usually difficult, as these experimental data are only correct under special conditions, which have not been sufficiently investigated so far.

The research in hand starts here. The perception of different situations of sound insulation shall be experimentally investigated under controlled conditions. Thereby, the speech intelligibility is first investigated with sample tests for different sound insulations.

• About Speech Intelligibility depending on different Sound Insulations
  H. Alphei and T. Hils [I]
[A]

In the first part (see previous abstract Hils/Alphei, 16:00), the performed hearing tests were described, and first results were shown. The aim is the verification of the expressions of the list based on the measured intelligibility for as real as possible situations.

In this second part it is tried to create a correlation between intelligibility and sound insulation at a specific initial level in the emitting room.

It is looked at the loudness more closely that, on the other hand, verifies the relation beween loudness and the levels of sound protection.

Altinsoy, E.

• BRTF – Body Related Transfer Functions for Whole-Body Vibration Reproduction Systems
  E. Altinsoy and S. Merchel [C]
[A]

If binaural recorded signals are played back via headphones, the transfer characteristic of the reproduction system has to be compensated for. Unfortunately the transfer characteristic depends not only on the transducer itself, but also on mounting conditions and individual properties of the respective ear. This is similar with reproduction systems for whole body vibration (e.g. electro-dynamic or hydraulic exciters). The transfer characteristic depends to a great extend on the individual body properties, e.g. weight, body mass index, adipose. For fundamental scientific studies or comfort studies it is required to present the same stimulus (amplitude and frequency) to all subjects. This is only possible by taking into account individual equalization filters which are based on the individual body properties of each subject and the specifications of the reproduction system. In this study body related transfer functions of 40 subjects are measured using an electro-dynamic excitation system. In addition anthropometric data of the subjects are collected. The importance of the individual transfer functions for whole-body vibration perception experiments is discussed. One aspect is the definition and creation of an average body transfer function representing a ''best matched body''. Another topic is the influence of the body properties on the transfer function characteristic.

• Audiotactile Feedback for Touch Screens
  S. Merchel and E. Altinsoy [C]
[A]

Touch sensitive displays and touch surfaces are more and more replacing physical buttons. Ticket vending machines or telecommunication devices, such as cellular phones (e.g. the Apple iPhone), use this space and cost saving technology. If a physical button is pressed, audio and tactile feedback confirms the successful operation. The loss of audiotactile feedback in touch sensitive interfaces might create higher input error rates and user dissatisfaction. Audiotactile feedback is especially important if there is no optical confirmation (e.g. the finger hides the display). Therefore the design and evaluation of suitable signals is necessary. In this paper a touch sensitive system is presented that reproduces event triggered audio tactile feedback. The tactile component is generated using an electro-dynamic exciter. The audio signal is played back via headphones. In a pilot study with six subjects the usability of such a system is investigated. Different synthetic signals are developed and compared with each other. In a dialing task the objective performance of the subjects (effectiveness, error rate) is measured. The perceptual quality of the designed signals is evaluated using a questionnaire.

Alzugaray, R.A.

• Structure-Borne Sound Excitation by Two-Phase Flow in Drainage Pipes
  R.A. Alzugaray and B.A.T. Petersson [C]
[A]

Cylindrical structures as drainage pipes are present in many domestic and industrial applications. These applications can be related to energy and fluid transfer.
The coupling between the conveying fluid and the pipe walls generate dynamic excitations and potentially represent a structure-borne sound source, transforming part of the vibrational energy into radiated noise. The noise due to pipe works in apartment housing is of significant concern and can reduce the environmental comfort for occupants.
This work deals with the analysis of velocity and force distribution on the surface of a vertical pipe in presence of a two-phase flow. This constitutes a first stage in flow and source characterization of drainage pipes. Some experimental results are presented.

Ammler, M.

• Psychoacoustic Evaluation of Music Reproduction in Passenger Cars
  H. Fastl, M. Ammler and A. Sulowski [C]
[A]

A method to assess the quality of music reproduction in passenger cars is developed and evaluated. In essence, the subject sits in a standing car with the motor off, and is presented from a CD music of different genres via the car's audio system. The subject has to rate different aspects of the performance on a 10 step scale using a PDA. From the ratings, a net like a spider net, with different attributes like natural, squeeking, dull, reverberant, acceptable, good etc. on its rays is constructed. The goal is a net with a large area, i.e. as many as possible of the attributes should attain the maximum rating.

Examples from a pilot study are given for car audio systems of different sophistication, and aspects like gender differences or possible biases due to product relationship will be discussed.

Anciant, M.

• Modeling the Vibro-Acoustic Effect of Trim on Full Vehicle and Component Level Analysis
  D. Blanchet, M. Anciant and L. Mebarek [C]
[A]

In the automotive industry, the influence of poro-elastic components on acoustic comfort has been mostly investigated for air borne noise at mid- and high frequency ranges; however, due to the lack of adequate theoretical formulations, the influence of poro-elastic in numerical vibro-acoustic simulation at lower frequency range has often been ignored or simplified by the use of distributed spring/mass on the BIW structure and impedance on the acoustic medium. Recent theoretical developments enabled to overcome this limitation by providing an efficient FEM formulation for poro-elastic material modelling. This FEM approach, implemented in VTM (Vehicle Trim Modeller) software developed by ESI-Group, enables the computation of the coupled response of a fully trimmed vehicle by taking into account the BIW structure, the acoustic cavity and the poro-elastic components (seats, carpet, dash insulator...).

This paper presents comparison between measurement and VTM simulation results for fully trimmed vehicle. It also presents the recently developed transmission loss (TL) module using FEM/BEM approach and comparisons between test and VTM prediction for academic and industrial cases.

Anderssohn, R.

• Fluid structure interaction and non-local admittance boundary conditions: Setup of an analytical example
  S. Marburg and R. Anderssohn [I]
[A]

In systems with fluid and structure coupled at their interfaces, the system equations are formulated in terms of variables for the fluid, e.g. sound pressure, and variables for the structure, e.g. displacement. The variables for the structure can be substituted by evaluation of the Schur complement. In that case, part of the Schur complement can be understood as a non-local boundary admittance. Herein, the authors try to provide the reader with a simple example. For that, the one dimensional problem of a duct is chosen. Single degree of freedom systems are introducted at both ends of the duct. For the setup of an example with non-local boundary conditions, both ends are connected by two springs and a rigid block. For this system, the admittance matrix will be shown analytically and we will discuss certain cases, e.g. the case of a very heavy connecting block which makes the admittance condition a local one.

Andrä, H.

• Optimization of Multilayered Porous Acoustic Absorbers
  H. Andrä, M. Kabel, M. Spies and H. Rieder [I]
[A]

A new method for the optimal material and layer design of acoustic trims without the usage of prototypes is presented. As an example the acoustic behaviour of a two-layered ceiling panel is optimized without increasing its weight or its thickness.

The optimization method consists of two steps: (1) The possible varieties of porous materials are analysed by computer simulation on the microscale, i.e. a certain set of boundary value problems is solved on the fully resolved microstructure. In a subsequent homogenization procedure a characteristic set of effective acoustic material parameters is computed together with their dependency on microstructural parameters like porosity. (2) A data base of these pre-computed characteristic material parameters is used in the actual optimization algorithm, which determines both the optimal thickness of the layers and their optimal microstructure with respect to a desired acoustic absorption behaviour.

A wide variety of materials can be taken into consideration: fibrous absorbers, open and closed cell foams, viscoelastic solids, porous or viscoelastic composites.

Since the main difficulty in this procedure is the computation of the characteristic acoustic parameters, we present the verification of the numerical scheme used for the computation of these parameters for the case of cylindrical pores.

Andringa, T.

• Realistic Prospects for Sound Source Description and Recognition for Complex Soundscapes
  T. Andringa, D. Krijnders, M. Niessen and M. van Grootel [I]
[A]

Automatic analyzes and descriptions of soundscapes, which are qualitatively similar to human descriptions, are very important to objectify soundscape research and to develop convincing tools for soundscape analysis. One possible way to develop such an automatic soundscape analysis system is to start with an annotation system for unrestricted real-world sounds and enrich this system with more and more automated functions. We will demonstrate an annotation system that automatically selects coherent time-frequency regions of similar spectro-temporal properties and similar acoustic ''textures''. The artificial system will learn through the interaction with a user whose task it is to assign names to regions. According to prototype theory of categorization, naming would structure the acoustic signals into meaningful relevant categories integrating textures and properties along family resemblance. After a low number of examples the system is able to associate an annotation with a time-frequency region. In cognitive terminology, the system will be able to use the emerging categories for evaluating new signals through top-down processing and membership decisions. We propose that systems like this can be used to bootstrap a more general, and eventually automatic, soundscape analysis system that complies with the aim stated above.

• DARES-G1: Database of Annotated Real-world Everyday Sounds
  M. van Grootel, T. Andringa and D. Krijnders [I]
[A]

This paper describes the design process and content of a publicly available database for everyday listening. ''While databases for musical sounds and studio recordings are readily available, databases for everyday sounds are lacking. Such a database is essential as more and more research is aimed at environmental sounds and everyday listening. The current lack of a standardized collection of properly recorded and annotated sounds in realistic acoustic environments limits the comparability of results between research groups, and hampers interdisciplinary cooperation. The collection described in this study must to help solve these issues. The collection currently consists of approximately 120 fragments of 60-second recordings of real-life situations. The use of head-mounted binaural microphones ensures a realistic capture of the environment for headphone playback. The recordings differ in location (e.g. indoor, rural, urban) and activity (e.g. taking the bus, dish washing, typing). Detailed annotations, provided for all recordings, describe the environment of the recording as well as the sources and onset-offset information of each of the sounds present. We aim to expand and improve the collection with more recordings and more specific information. We hope that many researchers in the auditory field will benefit from this database.

• Robust Vowel Detection
  B. Valkenier, D. Krijnders, R. van Elburg and T. Andringa [I]
[A]

Inspired by the robustness of human speech processing we aim to overcome the limitations of most modern automatic speech recognition methods by applying general knowledge on human speech processing. As a first step we present a vowel detection method that uses features of sound believed to be important for human auditory processing, such as fundamental frequency range, possible formant positions and minimal vowel duration. To achieve this we took the following steps: We identify high energy cochleogram regions of suitable shape and sufficient size, extract possible harmonic complexes, complement them with less reliable signal components, determine local formant positions through interpolating between peaks in the harmonic complex, and finally we keep formants of sufficient duration. We show the effectiveness of our method of formant detection by applying it to the Hillenbrand dataset both in clean conditions, as in noisy and reverberant conditions. In these three conditions the extracted formant positions agree well with Hillenbrand's findings. Thereby, we showed that, contrary to many modern automatic speech recognition methods, our results are robust to considerable levels of noise and reverberation.

• Signal Component Estimation in Background Noise
  R. Violanda, H. van de Vooren, R. van Elburg and T. Andringa [C]
[A]

The selection of coherent regions in the time-frequency representation of sound signals is an essential step in computational auditory scene analysis. These regions, referred to here as signal components (SC), ideally represent all information about the sound sources present in the scene. However, due to the limitation imposed by the uncertainty relation between frequency and time, the corresponding energy distribution in the time-frequency representation is often smeared. This poses a difficulty in extracting SCs belonging to a specific sound source especially in the case where multiple sources are present. In this paper we present a method of estimating the SCs of a source using the instantaneous frequency (IF) and grouped-delay corrected time (GDCT) representation. The spread of the IF and GDCT in every time-frequency bin is used as a measure of the local signal-to-noise ratio, this ratio is high where a possible single sound source is present. We show examples of the applicability of this technique in the analysis of tonal sounds ranging from speech, music to tonal environmental sounds. With proper grouping of the SCs, we show that our method is an important step towards speech extraction from noisy backgrounds.

• Fingerprinting rotation machinery
  H. van de Vooren, R. Violanda, R. van Elburg and T. Andringa [C]
[A]

Malfunctioning of a complex rotation machine is often preceded by (subtle) changes in its vibrational behavior on top of the non-stationarities induced by the variation in rotation frequency. Although the analysis of these vibrations is fairly complex, these changes can be exploited to detect possible defects. An actual implementation in an early-warning system can be used to prevent machinery from severe damage. All the expected defect related vibrations , especially bearing damage, show the same dependency on rotation frequency as the vibrations associated with normal operation. For that reason, the defect-related-vibration frequency component can be found at a fixed position after normalizing the vibration signal. Our normalization technique, based on instantaneous frequency analysis and harmonic grouping, extracts the rotation frequency from the vibration signal itself rather than using a separate measurement . Nederlandse Gasunie - a gas transport company - provided us with 25 vibration recordings gathered from the main axes of 5 operating gascompressors. Each recording was normalized providing us with a characteristic fingerprint for every machine. First results show that these fingerprints show a remarkable robustness under changes in operating conditions and strong stability over longer periods of time (1 year) making them suitable for condition monitoring purposes.

Anemüller, J.

• Blind and Non-Blind Spatial Signal Processing Using Head-Related Impulse Responses
  H. Kayser, B. Kollmeier and J. Anemüller [C]
[A]

Spatial signal processing for hearing aids is of particular importance in acoustically challenging environments. The development and testing of algorithms performing digital signal processing, e.g., blind source separation (BSS), beamforming and direction-of-arrival estimation requires realistic test conditions. For the evaluation of such blind and non-blind algorithms, virtual sound fields are generated by convolution of sound signals with impulse responses measured for different spatial source positions. By the use of head-related impulse responses measured in real scenarios, a simulation is achieved that approximates real-world recordings. For this purpose, multi-channel hearing aids were mounted on a head and torso simulator. Hence, the number and type of superimposed sound sources can be chosen arbitrarily. Background noise recorded in the underlying scene can be added with a selectable SNR. Due to the knowledge of the entire transmission characteristics from each source to each microphone, optimal linear filtering techniques are applicable. The results achieved by multi-channel Wiener filtering serve as benchmark for practical filtering algorithms having less knowledge about the underlying situation. A BSS approach is applied to multiple mixtures of speech signals in anechoic conditions and an office room. The quality of separation is evaluated subject to the spatial arrangement of the sources.

• Measurement, modelling and compensation of nonlinearities in hearing aid receivers
  T. Schaer, S.D. Ewert, J. Anemüller and B. Kollmeier [C]
[A]

Hearing aid receivers operate over a large dynamic range and reach output levels of up to 130 dB. Electromagnetic receivers are commonly used in hearing aids since they offer a very high efficiency. Receivers are typically characterized by their linear transfer function. These linear distortions are easily measured with standard methods and if compensation is applied it is often limited to the magnitude transfer function. The current study addresses a more complete characterization of the hearing aid receiver by measurement of the nonlinear transfer function. Nonlinear distortions of the receiver are output level dependent and might negatively influence measures like speech intelligibility of an aided hearing impaired person, particularly at high levels. Moreover, nonlinear distortions are problematic in the context of future ''high-fidelity'' hearing aids and compensation methods are of interest. Here, a fast and efficient sinesweep method proposed by Farina was employed to estimate the linear and nonlinear transfer function. The measurement was repeated at different output levels. From the data, nonlinear input-output functions can be constructed or alternatively, the diagonal elements of a Volterra-series based description of the nonlinear transfer function can be derived. A nonlinear receiver model and compensation methods are suggested and tested.

• Audio Event Detection for In-Home Care
  P. van Hengel and J. Anemüller [C]
[A]

One in eight persons attending hospital following an accident in the home, are aged 65 and over. This number will increase with the demographic change and the desire of the elderly to remain in their own homes. Hence, the demand for unobtrusive monitoring systems based on autonomous sensor technologies will increase. Such systems can limit observation by a human operator to cases when there is evidence of an incident, by responding to events in a 'human-like' fashion. To do so, the system must pick up situations that also would attract human attention, including situations it has not encountered before. Hearing is an important sense in terms of steering the attention of a human observer. For this reason the detection of sound events that could indicate an incident is a key part of such a monitoring system. Formerly, an acoustic system was designed for the detection of verbal aggression. This technology is now expanded, combined with acoustic localization and integrated with video analysis and artificial intelligence. The resulting methods will be tested in the context of in-home care under real-life conditions. For this purpose a realistic test-home has been constructed.

• Acoustic Object Detection in Adverse Conditions
  J.-H. Bach and J. Anemüller [C]
[A]

In the current work, the problem of acoustic object detection in adverse and realistic background noise is tackled based on a recently proposed method for speech detection in noise [1]. The method is based on the decomposition of the signals into amplitude modulation spectrograms [2]. These perceptually motivated features provide a signal description in terms of power density at combinations of acoustic frequencies and modulation frequencies within meso-scale temporal windows of about 1s length. The target object sounds -ranging from speech to animal voices and characteristic elements of office activities- are embedded in recordings from noisy scenes at various SNRs. These acoustic background scenes include, e.g., recordings from heavy traffic streets, pedestrian zones, landscape. Numerous combinations of objects and environments have been analysed to provide some typical real-world acoustic scenes. These rather adverse conditions are compared to the performance in white noise as a reference. The results from our perceptually motivated features are compared to those obtained from standard of mel frequency cepstral coefficients (MFCC). Classification is based on the discriminative model of support vector machines (SVM) that are particularly suited for high-dimensional discrimination tasks.

Angelini, F.

• Vibroacoustic characterization of flexible hoses for air conditioning systems
  A. Agnani, F. Angelini, M. Martarelli and E.P. Tomasini [I]
[A]

Flexible tubes covered with rubber are the only way to damp vibrations and noise coming from compressor in air conditioning systems. Usually they have, as core, a corrugated steel pipe, for increasing the hose stiffness, this causing fluid turbulences generating vortex separation that gives a beginning of a audible and bothersome high frequency whistle (from 2500 Hz to 5000 Hz). The vibroacoustic analysis consists on the experimental and numerical (by Finite Element Methods) characterization of the hose vibrational behaviour and consequent acoustic characterization. The experimental measurement of the hose vibration has been performed in both controlled and operating conditions, the latter with the hose installed in the test bench including all the refrigeration cycle components. The acoustic characterization has been performed via propagation of the measured surface vibration in Boundary Element Codes and validated by means of acoustic intensity measurement. The acoustic measurement in operating conditions is very challenging, being the acoustic emission of the hose comparable to the total emission of the refrigeration bench (in primis the compressor), and therefore the hose must be isolated from the rest of the bench. Consequently the possibility of predicting the acoustic emission via BEM allows to avoid this kind of complicated measurement.

Angster, J.

• Reflection functions of pipes
  T. Trommer, H. Ausserlechner, J. Angster and A. Miklos [C]
[A]

The sound an organ pipe radiates is predominantly influenced by the resonator. In it sound waves travel back and forth between the lower end (consisting of upper lip, pipe mouth, languid) and the upper end being partially reflected as well as radiated from these ends. These reflections and herewith directly the pipe sound is dependent on different properties of which mainly contributing factors are the geometry of the resonator in size and shape (cylindrical, conical, diameter, length) as well as the attributes of the ends (open, closed, tuning geometries, mouth size). The first sound signal excites the stationary air column defined by the resonator. The reflected part of the sound signal from the upper end after being reflected from the lower end superimposes with the periodic sound emitted continuously by the sound source. In optimum this retroactivity leads to a steady state of the resonant circuit radiating the corresponding organ pipe sound. In order to model the resonant part of a pipe the reflection functions of both ends in dependency of the geometries are needed. These will be gathered by an especially developed method by means of time domain measurements on pipes with corresponding dimensions to in-situ employed organ pipes.

• Measurements on an adjustable pipe foot model
  H. Ausserlechner, T. Trommer, J. Angster and A. Miklos [C]
[A]

The build-up of the sound in a flue organ pipe is mainly influenced by the edge tone. A small change of the flue parameters, such as positions of the lower and upper lip, the wind pressure in the pipe foot and the width of the flue, mostly results in big differences in the build-up and the spectra. In order to understand the basic physical processes a pipe foot model of a labial organ pipe with adjustable parameters has been developed. Furthermore, different types of languids and resonators (pipe bodies) can be attached to the model. A High value was set on comparisons and measurements with real metal organ pipes in order to give an indication for advisable ranges of the parameters. Flow velocity, edge tone, mouth tone and pipe sound measurements were carried out with the help of this model. From the results of the measurements theoretical models have been developed for flow velocity, velocity distribution, edge and mouth tone frequencies and for the ratios of the frequencies of edge tone modes. These models may facilitate the design and scaling of flue organ pipes.

• Optimal scaling of the depth and width of wooden organ pipes
  J. Angster, E. Esteve and A. Miklos [C]
[A]

In case of wooden pipes the reduction of pipe width would be desirable, because the space requirement of the pipe organ could then also be reduced. On the other hand, the sound quality could be worse for too narrow pipes. The aim is to find the narrowest scaling of wooden pipes with still appropriate sound quality. Optimal scaling of the depth and width of wooden pipes has been developed in order to help the design of the instrument to the space requirements of the room. A series of differently scaled wooden pipes has been acoustically investigated in the anechoic room of the IBP. On the basis of the results a design and scaling method has been developed that can take into account the effect of wall dimensions, material and finish on the stationary spectrum of the sound. With the help of the new method the organ builder can optimize the dimensions of the wooden pipes, but maintain an optimal sound quality in the same time.

Apostoli, A.

• Reconstructing the Lituus: A Reassessment of Impedance, Harmonicity, and Playability
  A. Apostoli, S. Logie, A. Myers, J. Kemp, J. Chick and A. Braden [C]
[A]

Bach's motet `O Jesu Christ, meins Lebens Licht' (BWV no. 118), first performed in 1736-7, calls for a Lituus, believed to be a long trumpet-like instrument, with playing characteristics somewhere between a trumpet and an alphorn. Unfortunately, there are no surviving examples of Bach's instrument. Using software recently developed at the University of Edinburgh for brass instrument optimisation, a new bore profile with theoretically calculated input impedance curves has been proposed and submitted to a collaborating instrument maker with the aim of providing a natural instrument, with a distinct timbre, that responds well to the player and is in tune over a wide range of natural notes.

This paper describes the process, computational modelling, and experimental work undertaken in an attempt to ensure a satisfactory final instrument design. An initial validation of this process was achieved, in part, by investigating the playing properties of an instrument with a notionally similar bore profile, a 2.4m long Fiscorn in C (a type of Catalonian flugelhorn). This work compared the measured bore profile, input impedance, playability and intonation with those derived theoretically.

• Transient behaviour in the motion of the brass player's lips during a lip-slur
  S. Logie, S. Stevenson, A. Apostoli, J. Chick and M. Campbell [C]
[A]

When moving between notes on an orchestral brass instrument experienced players routinely change from one resonant mode of the instrument to another. Skilled players can 'lip-slur' from one note to another making the transition between resonant modes sound very smooth as perceived by the listener, and do so with little thought or understanding about what is happening to the lip/air-column interaction during this intra-note transient.

This paper offers an investigation of the transient behaviour of the player's lips when moving from a strong, fully developed coupling of the lips to one air column resonance, to the strong coupling of another resonant mode of the same air column. A high speed digital video camera is used to view the motion of the lips through a transparent mouthpiece for a range of orchestral brass instruments of different lengths. Image processing techniques allow us to measure the opening area of the lips as a function of time, and this data is synchronised with pressure data measured in the mouthpiece together with the radiated sound from the bell of the instrument to provide us with greater insight to mechanics of the lip-slur transient.

Aradilla, G.

• Improving Context Modeling for Phoneme Recognition
  D. Vasquez, G. Aradilla, R. Gruhn and W. Minker [C]
[A]

In this paper we investigate a novel technique for modeling long temporal context in phoneme recognition. Conventional techniques take a global context consisting of consecutive concatenated acoustic features as input for a phoneme classifier. Thus, the classifier must learn different variations corresponding to different parts of the global context. In contrast to conventional techniques, our scheme decomposes the temporal context of each phoneme into a set of slides. Each temporal slide is input for a non-linear classifier given by a Multilayered Perceptrons (MLP). Every MLP is trained using the central label of the global context. Thus, the assumption that the current phoneme occupies the global context is preserved. Furthermore, each MLP can robustly model different temporal context inside each phoneme. The outputs of the classifiers are combined to estimate posterior probabilities of phoneme-level classes. These posteriors are employed in a hybrid HMM/MLP framework. Experiments have shown an absolute phoneme error reduction of 3.6% compared to a baseline classifier with the same context length. A further improvement is achieved when a hierarchical approach is implemented, which introduces information across different phonemes.

Aretz, M.

• Uncertainty in Acoustic Boundary Characterisation and its Influence on the Sound Field in Room Acoustic FE Simulations
  M. Aretz and J. Knutzen [I]
[A]

The room acoustic design of the low frequency characteristics of small rooms still poses a challenging task to acoustic designers and consultants. In this field numerical methods such as the finite-element-method (FEM) are promising contenders for the realistic simulation of low frequency sound fields. While contemporary CAD and meshing tools facilitate the design of high quality geometric models with regular and sufficiently fine meshes, the Achilles' heel of room acoustic FEM simulations appears to be the realistic representation of the room boundaries. The presented study investigates the influence of different boundary representations on the simulated sound field in small rooms. Therefore measurements and FE simulations of the sound field in a scale model room with a well defined geometry and variable boundary conditions have been conducted. Two different approaches are followed to model the porous absorbers on the room boundaries. Firstly the absorber layers are modelled with their exact 3D dimensions using an 'equivalent homogeneous fluid model'. Secondly the acoustic characteristics of the absorber layers are modelled by an appropriate impedance boundary condition, which is determined by measurement or model calculation based on measured flow resistivities. The audibility of differences between simulation and measurement results is assessed by subjective testing.

Arvengas, A.

• Density Threshold for Acoustic Cavitation in Water
  F. Caupin, K. Davitt and A. Arvengas [I]
[A]

We use a focused 1 MHz acoustic burst to generate large mechanical tensions in bulk water. We describe how this technique allows us to generate cavitation and to study its statistics. We have recently built a fiber optic probe hydrophone, to measure the density threshold for cavitation. We report the temperature dependence of this quantity, and compare it with other experiments.

Aspuru, I.

• Soundscape characterization in the city of Bilbao
  G. Memoli, I. Garcia and I. Aspuru [I]
[A]

The first round of noise mapping and action plans has left a lot of unresolved questions, mainly addressing the design of perception-oriented action plans and the classification/preservation of quiet areas. While the interest in the practical use of the soundscape concept to address this problems is now growing all through Europe, psychoacoustic indicators linked to perceived noise are the key to determine the path to follow. This work presents the soundscape characterization of two different green areas in the Municipality of Bilbao (E), with in common the presence of a heavily trafficked road nearby. The acoustical experience of passers-by will be mapped using an indicator related to the time history of sound energy, related in previous studies to people's perceptions. Comparison with the same description performed by classical psychoacoustic analysis and perspectives for innovative, positive soundscape based actions and measurement techniques will be discussed.

Auerbach, M.

• DEUFRAKO 'Prediction and Propagation of Rolling Noise'
  M. Auerbach and M. Bérengier [C]
[A]

This project aims at the development of new road pavement concepts with respect to tyre-road noise in a wider context of the abatement of road traffic noise. French and German institutes headed by the ''Laboratoire Central des Ponts et Chaussées'' (LCPC) and the ''Bundesanstalt für Straßenwesen'' (BASt) cooperate throughout this project. The theoretical approach is based on the SPERoN (''Statistical Physical Explanation of Rolling Noise'') model developed by one of the partners. In a first stage, the model was validated for different kinds of French road surfaces. In a second step, the model was applied as an adapted tool to design new textures for low noise road surfaces. Finally, in a third step, the results issued from SPERoN were used as an input for outdoor sound propagation models developed in France and Germany in order to estimate the effect of this optimised model-aided low noise pavement in the far field, close to the façades. A ranking with other typical German and French pavements has lead to a common database (DEUFRABASE). The results of the project will be presented.

Ausserlechner, H.

• Reflection functions of pipes
  T. Trommer, H. Ausserlechner, J. Angster and A. Miklos [C]
[A]

The sound an organ pipe radiates is predominantly influenced by the resonator. In it sound waves travel back and forth between the lower end (consisting of upper lip, pipe mouth, languid) and the upper end being partially reflected as well as radiated from these ends. These reflections and herewith directly the pipe sound is dependent on different properties of which mainly contributing factors are the geometry of the resonator in size and shape (cylindrical, conical, diameter, length) as well as the attributes of the ends (open, closed, tuning geometries, mouth size). The first sound signal excites the stationary air column defined by the resonator. The reflected part of the sound signal from the upper end after being reflected from the lower end superimposes with the periodic sound emitted continuously by the sound source. In optimum this retroactivity leads to a steady state of the resonant circuit radiating the corresponding organ pipe sound. In order to model the resonant part of a pipe the reflection functions of both ends in dependency of the geometries are needed. These will be gathered by an especially developed method by means of time domain measurements on pipes with corresponding dimensions to in-situ employed organ pipes.

• Measurements on an adjustable pipe foot model
  H. Ausserlechner, T. Trommer, J. Angster and A. Miklos [C]
[A]

The build-up of the sound in a flue organ pipe is mainly influenced by the edge tone. A small change of the flue parameters, such as positions of the lower and upper lip, the wind pressure in the pipe foot and the width of the flue, mostly results in big differences in the build-up and the spectra. In order to understand the basic physical processes a pipe foot model of a labial organ pipe with adjustable parameters has been developed. Furthermore, different types of languids and resonators (pipe bodies) can be attached to the model. A High value was set on comparisons and measurements with real metal organ pipes in order to give an indication for advisable ranges of the parameters. Flow velocity, edge tone, mouth tone and pipe sound measurements were carried out with the help of this model. From the results of the measurements theoretical models have been developed for flow velocity, velocity distribution, edge and mouth tone frequencies and for the ratios of the frequencies of edge tone modes. These models may facilitate the design and scaling of flue organ pipes.

Bade, K.

• Everything in its right place? - Learning a User's View of a Music Collection
  K. Bade, A. Nürnberger and S. Stober [I]
[A]

Keeping one's personal music collections well organized can be a very tedious task. Fortunately, today, many popular music players (such as AmaroK or iTunes) have an integrated library function that can automatically rename and tag music files and sort them into subdirectories. However, their common approach to stick with some hierarchy of genre, artist name, and album title barely represents the way a user would structure his collection manually. When it comes to organizing a music collection according to a user-specific hierarchy, three things are required: First, the music files have to be described by appropriate features beyond simple meta-tags. This includes content-based analysis but also incorporation of external information sources such as the web. Second, knowledge about the user's structuring preferences must be available. And third, and most importantly, methods for learning personalized hierarchies that can integrate this knowledge are needed. We propose for this task a hierarchical constraint based clustering approach that can weight the importance of different features according to the user perceived similarity. A hierarchy based on this similarity measure reflects a user's view on the collection.

Bäuml, T.

• Modelling Sound Level and Pressure Loss in HVAC Duct Networks using a Framework in Modelica/Dymola
  H. Kühnelt, A. Haumer, T. Bäuml, U. Reisenbichler, C. Reichl and G. Karlowatz [C]
[A]

In the early design phase of large HVAC duct networks the joint calculation of the sound level and air flow rate and pressure loss as well as the prediction of the noise level in the passenger compartment gain more and more in importance. A unified framework for the one-dimensional flow acoustic simulation using Dymola/Modelica is introduced. The acoustic part is described by plane wave two-port theory, the fluid mechanic part by pressure loss models. Various approaches including analytic and (semi)empirical models based on experimental and simulation data can be implemented. Each component either represents the acoustic or the flow part of a duct element or can, itself, be composed of both types allowing the modelling of non-standard components. The object oriented approach provided by Modelica assures the extensibility of the models e.g. by transport and exchange of heat or the (de)humidification of the air. It also allows the integration into a complete system simulation of the HVAC network including the cooling circuit and the electronic control loop. The simulation is easily set up using a GUI; an external parameterization ensures persistent data management and a smooth workflow. The resulting system of equations is automatically pre-processed and solved by Dymola.

Bailly, C.

• A hybrid method for Computational Aeroacoustic applied to internal flows
  M. Piellard and C. Bailly [I]
[A]

A hybrid method of aeroacoustic noise computation based on Lighthill's Acoustic Analogy is first assessed, and then applied to investigate the noise radiated by a low Mach number flow through a ducted diaphragm [1]. The simulation method is a two-step hybrid approach relying on Lighthill's acoustic analogy, assuming the decoupling of noise generation and propagation. The first step consists in an incompressible Large Eddy Simulation of the turbulent flow field, during which the Lighthill's source term is recorded. In the second step, a variational formulation of Lighthill's Acoustic Analogy using a finite element discretization is solved in the Fourier space. The assessment of this method consists in validating the formal linking from the flow to the acoustic computations; the expression of the source term in Lighthill's Acoustic Analogy is also discussed and numerically demonstrated. The method is applied to a three-dimensional ducted diaphragm with low Mach number flow; a good agreement is found regarding aerodynamic results with Direct Noise Computation performed by Gloerfelt & Lafon [2]. However, the acoustic results prove the difficulties of applying a volume approach to practical flows, mainly due to the large computing resources required.

[1] Piellard & Bailly. 2008. AIAA Paper 2008-2873.

[2] Gloerfelt & Lafon. 2008. Computers & Fluids 37(4), 388-401.

Bakker, J.F.

• Design of an Ultrasound Cylindrical Phased-Array for Hyperthermia Breast Cancer Treatment
  K.W.A. van Dongen, J.F. Bakker, M.M. Paulides, I.-M. Obdeijn and G.C. van Rhoon [C]
[A]

The objective of this theoretical study is to design an ultrasound (US) cylindrical phased-array that can be used for fever-range hyperthermia (40oC to 44oC) treatment of tumors in intact breasts. Hereto, design parameters such as frequency, number of transducers per ring, ring distance and number of rings are optimized to obtain an interference pattern with a small primary focus, while suppressing secondary foci.

In addition, we have investigated the influence of acoustic and thermal heterogeneities on the specific absorption rate (SAR) and temperature patterns in order to determine the necessity of using heterogeneous models for US applicator design and treatment planning.

From the simulations it is shown that an US cylindrical phased-array can be used to obtain 44oC in the center of tumors located anywhere in the intact breast. In addition, it is demonstrated that due to the low frequency content of the signal (100kHz) acoustic heterogeneities do not disturb the SAR and temperature patterns.

Balazs, P.

• On the Structure of the Phase around the Zeros of the Short-Time Fourier Transform
  F. Jaillet, P. Balazs, M. Dörfler and N. Engelputzeder [C]
[A]

The short-time Fourier transform is a time-frequency representation widely used in audio signal processing. At first sight, the phase of the transform may appear to be difficult to interpret, a fact that led to the development of many processing schemes in audio processing relying on the modulus of the transform only. However, it has been shown, that the phase information can be successfully exploited for improved analysis and processing. This is illustrated for example by the reassignment method or the phase vocoder algorithm.

In order to better exploit the phase information in audio processing, a better understanding of its structure is desirable. This study describes a particular behavior of the phase which can be observed when looking at the phase derivative over time or frequency around the points where the transform equals zero. This recurrent behavior is illustrated using numerical simulations on synthesized and real audio signals. The performed simulations show the influence of the parameters of the transform on the observed structures. Some consequences of this phase structure for audio processing are also presented.

• Removing Components from a Time-Frequency Representation
  M. Dörfler, P. Balazs and F. Jaillet [C]
[A]

Time-frequency representations such as the spectrogram or short-time Fourier transform seem to be well suited to the task of removing certain components with approximately disjoint support in the time-frequency domain. For example, one might be interested in suppressing a certain instrument's contribution from a music signal. Such approaches are used in Computational Auditory Scene Analysis by the name of Time-Frequency masks. However, the trivial approach of just deleting the corresponding component in the time-frequency representation leads to artefacts such as ''ghost-tones'', i.e. a modified version of the deleted component is still audible, or to the perception of unnatural sound coloration comparable to a ''phasing effect''. We compare different approaches to tackle this problem: a soft-thresholding procedure is applied and compared to a method using statistical models. On the other hand, we draw conclusions from considering the optimal separating mask for known signal components. In this situation, a model promoting sparsity in the representation can lead to favourable results by yielding a feasible approximation of the optimal mask. We perform informal listening experiments to evaluate the resulting removal of certain components in a complex audio-signal.

Bardeli, R.

• A Study of Throat Microphone Performance in Automatic Speech Recognition on Motorcycles
  T. Winkler, S. Pronkine, R. Bardeli and J. Köhler [C]
[A]

Barney, A.

• A model of jet modulation in voiced fricatives
  A. Barney and P. Jackson [I]
[A]

The radiated sound during the production of voiced fricatives has harmonic and random components. The harmonic components are associated with the vibration of the larynx and the noise components are associated with the time-varying flow through a constricted region of the oral cavity. The two sources interact non-linearly to produce a noise component with an amplitude envelope modulated at the fundamental frequency of voicing. This paper describes pressure measurements made in a dynamic physical model of the larynx and vocal tract, which were used to study the aerodynamic and acoustic behaviour in a flow regime corresponding to that found during voiced fricative production. An orifice plate and a sharp-edged obstacle downstream model the geometry of the oral cavity, representing the tongue constriction and teeth respectively. The depth of modulation of the frication noise was found to be linearly related to the pressure upstream of the orifice plate. The phase of the noise bursts relative to the upstream pressure was found to vary systematically with the constriction-obstacle distance. These findings form the basis of a parametric model of voiced fricative production and give some insight into the likely aeroacoustic mechanism responsible for sound production.

Barrera Figueroa, S.

• Numerical design and testing of a sound source for secondary calibration of microphones using the Boundary Element Method
  V. Cutanda Henriquez, P. Juhl and S. Barrera Figueroa [I]
[A]

Secondary calibration of microphones in free field is performed by placing the microphone under calibration in an anechoic chamber with a sound source, and exposing it to a controlled sound field. A calibrated microphone is also measured as a reference. While the two measurements are usually made consecutively, a variation of this procedure, where the microphones are measured simultaneously, is considered more advantageous from the metrological point of view. However, it must be guaranteed that the two microphones receive the same excitation from the source, although their positions are some distance apart to avoid acoustic interaction.

As a part of the project Euromet-792, aiming to investigate and improve methods for secondary free-field calibration of microphones, a sound source suitable for simultaneous secondary free-field calibration has been designed using the Boundary Element Method. The source has a central plug with the effect of reducing the spatial variations of the sound pressure at the microphone positions within the frequency range of the calibration. After some preliminary measurements, the source has been tested thoroughly in a calibration setup at the acoustic laboratory of the Danish Fundamental Metrology Institute (DFM). The design and verification of the source are presented in this communication.

Barron, M.

• Then and now - how concert hall design of the 1960s and '70s compares with the present
  M. Barron [I]
[A]

The 1960s and '70s were an exciting time for concert hall design. Several important new halls were built to replace war damage, but arts provision also needed improving in many cities round the world. Concert hall design was in an experimental phase with some major successes and disappointments; the Philharmonic halls in Berlin (1963) and New York (1962) illustrate the ''extremes''. The De Doelen large hall of 1966 is rated as an acoustic success but is interesting as an example of a particular design philosophy: a large hall with highly scattering walls and ceiling. This approach has not been followed much elsewhere.

Current approaches to concert hall design are different in many ways to those of around 40 years ago. The tools available now include much better understanding of the important dimensions for concert hall listening, objective quantities corresponding to these subjective dimensions, modelling by scale or computer simulation modelling. Some designers also use auralisation. But as well as technical developments, the role of the acoustician is now generally more prominent in the design team. With greater responsibility, current designers often take fewer risks and adhere to one of two precedents: the parallel-sided hall and the vineyard terrace hall.

Barth, M.

• Estimation of spatially distributed temperature and flow fields in air using acoustic travel-time tomography
  M. Barth, M. Wilsdorf and A. Raabe [C]
[A]

Sound propagation in air mainly depends on temperature and flow properties along the propagation path of acoustic signals. Thus, measuring travel-time between a sound source and a receiver knowing the sound ray path length, averages of these meteorological quantities along the propagation path can be deduced. Using tomographic techniques as known from medical applications, spatially averaged distributions of temperature and flow can be calculated from a combined treatment of single-line measurements along different paths through the investigation area.

In order to estimate three-dimensional fields of temperature and wind, hard- and software of an existing two-dimensional system had to be adapted. To enable a three-dimensional arrangement of sensors around the area under investigation loudspeakers were constructed that ensure a homogeneous sound emission into one half-space. Furthermore, the software was adapted to fit the requirements of a three-dimensional reconstruction of fields.

In this contribution, sensitivity studies with different temperature and flow fields are presented to demonstrate the ability of the system to estimate three-dimensional distributions of these quantities. Furthermore first results from wind tunnel experiments are shown which were carried out to estimate the sensitivity of the new system for the reconstruction of three-dimensional flow fields on meter scale.

• Inverse reconstruction techniques and their applicability for acoustic travel-time tomography
  G. Fischer, M. Barth and A. Ziemann [C]
[A]

Remote sensing by acoustic travel-time tomography enables one to determine area-averaged sound speed fields in an area of interest, e.g. the atmospheric boundary layer. Due to the dependency of the sound speed on temperature, the distribution of this meteorological quantity can be reconstructed by inverting the given travel times of sound signals along defined sound ray paths through a scanned medium. Over the past years, different inverse approaches were used for acoustic tomography. In this study, the results achieved with matrix inversion, simultaneous iterative reconstruction technique, and stochastic inversion are compared. To identify the algorithm which yields the optimal results by reconstructing two-dimensional temperature fields, both synthetically generated and experimental travel times were considered. On the one hand, the temperature fields were compared concerning their reconstruction accuracy and spatial resolution. The results show that the stochastic approach yields spatially highly resolved and accurate temperature fields. Furthermore, to estimate the applicability of the different methods during operational studies, the computational effort was analysed. Here, the stochastic inversion offers some deficits because of the relatively long computational time. Oppositely, the matrix inversion and the iterative reconstruction technique are suitable for online interpretations by providing temperature fields with acceptable accuracy.

Bartolomaeus, W.

• Silent Road Trafic 2
  W. Bartolomaeus [C]
[A]

After finishing the research project ''Silent Road Traffic 1'' in 2003 the follow-up project ''Silent Road Traffic 2'' was started in 2005. Under the leadership of the Bundesanstalt für Straßenwesen - BASt a budget of 4.5 Mio. e was shared with ten other partners from Industry (Continental AG, RW Sollinger Hütte GmbH and Maurer Söhne GmbH & Co KG) , Research Institutes (Müller BBM GmbH, Forschungsinstitut für Pigmente und Lacke, Bundesanstalt für die Materialforschung und -prüfung - BAM) and Universities (Universität Stuttgart - Institut für Straßen- und Verkehrswesen, Technische Universität München - Fachgebiet Hydromechanik, Universität Hannover - Institut für Baumechanik und Numerische Mechanik and Universität Hamburg-Harburg - Institut für Modellierung und Berechnung). The research project consists of three parts: ''Quiet Tires'', ''Quiet Roads'' and ''Result Checking''. The end of the project will be at the end of 2009. But nevertheless there are already a lot of results available, specially from the part ''Quiet Tires''.

Bartosch, T.

• Automated modelling procedure for acoustic Wave Based Technique models
  G. Müller and T. Bartosch [I]
[A]

In modern vehicle acoustic development, fast and easygoing modelling and simulation procedures are required to meet the today's demands in automotive industry. To close the gap in the mid frequency range, several years ago, the Wave Based Technique, a deterministic vibro-acoustic simulation technique for the mid-frequency range, was established. Instead of meshing the fluid domain of interest into a vast number of small volumes (as required for FEM), the model has to be partitioned into convex polyhedron shaped sub-domains. This paper intends to present an algorithm for the automated creation of acoustic WBT models of passenger compartments out of hull models. In addition, relations between geometric model parameters and modelling/simulation time and quality are investigated. From these investigations, conclusions are drawn and automated modelling strategies are established. Furthermore, alternative modelling approaches are discussed and proposed.

• SEA vehicle model for rolling- and engine noise
  F. Nentwich, T. Bartosch and G. Müller [C]
[A]

The task is to optimize the acoustic trim in balance with cost and weight during the medium project phase. The SEA model should predict how much a trim measure will reduce the interior noise at mid and high frequencies. The body structure has a predictive SEA description based on the latest FEM model. The trim and its acoustic behaviour are either measured or estimated via sub-model. The real excitations are modelled by defining the energy levels of their adjoining subsystems as boundary conditions. These levels are measured on a forerunner model with the dynamometer test rig and subsequently imposed on the model's matching subsystems. Excitation and structure together yield a simulation of the coast down rolling noise and run up engine noise. This enables a speedy simulation of trim variations and their acoustic impact.

• Vehicle noise reduction of multiple load cases using vibro-acoustic potential analysis
  F. Nentwich, T. Bartosch and G. Müller [C]
[A]

The engine noise and rolling noise are to be reduced with trim measures derived from a SEA model. After a SEA model with 370 subsystems has been set up, there are some 25400 non-zero coupling and internal loss factors to choose from - too much for manual labour. Therefore MAGNA STEYR developed the VAPA algorithm. It investigates how a loss factor variation affects the cabin's sound pressure level. The algorithm calculates the theoretically possible reduction presuming an unlimited loss factor und thus sets the upper limits for the noise reduction expectable from practical measures with a limited loss factor. With this paper we show an easy extension of the VAPA in order to consider a multi-load case scenario. The VAPA identifies the most over all promising vibro-acoustic potentials. The engineer translates these into construction measures and finally reruns the model. Using this procedure a trim package is designed for a small car. It weighs 4kg and successfully reduces the mid and high frequency engine noise and rolling noise up to 3.5 dB(A).

Basten, T.

• Acoustic Monitoring of a Failing Dike
  T. Basten, A. Volker and P. Toet [C]
[A]

The 'ijkdijk' in Groningen, The Netherlands, is an international fieldlab for new inspection and monitoring techniques for water barriers. The goal is to learn more of their behavior and to increase confidence in dikes. The dike of the future is filled with a lot of sensors which are used for monitoring and to predict failure of the dike. During the initial experiments a lot of parameters are measured, with all kind of sensors. The current paper deals with the acoustic experiments in the dike during the macro-stability experiment in September 2008 while a complete collapse of the dike was forced. The water barrier, 100 meters long and 6 meters high was equipped with two tubes in which microphones and hydrophones were placed. Besides, an array of hydrophones was used, listening to an array of loudspeakers on top of the dike. The goal of this array was image processing based on tomographic inversion. Furthermore a fibre optic sensor was applied as a prototype hydrophone for future applications. A lot of data was generated during the experiments up to the moment of complete collapse of the dike. The paper describes the experiments and some initial results of the analysis.

• An acoustic vector based approach to locate low frequency noise sources in 3D
  H.-E. de Bree, C. Ostendorf and T. Basten [C]
[A]

Although low frequency noise is an issue of huge societal importance, traditional acoustic testing methods have limitations in finding the low frequency source. It is hard to determine the direction of the noise using traditional microphones.

Three dimensional sound probes capturing the particle velocity vector and sound pressure offer a novel possibility to locate the low frequency noise sources in 3D. An overview on the benefits of acoustic vector sensors will be given and a low frequency case study will be presented and discussed.

The method has been tested during low frequency noise measurements in an urban environment. In this case a very specific frequency of the source was known. This frequency was used to tune the measurement system. The paper will describe the results of the test.

Batke, J.-M.

• Quality Evaluation of Ambisonics Recordings
  J.-M. Batke [I]
[A]

Higher Order Ambisonics (HOA) is a sound field description techniques that is capable to represent a sound field with all three spatial dimensions. The coefficients for such a HOA representation are usually acquired by a spherical microphone array when natural sounds should be captured.

The mechanical size of such an array and the used number of microphone capsules limit the accuracy of the acquired HOA coefficients. Furthermore, the spatial extension of the microphone diaphragm can be taken into account. In this contribution, these influences are discussed with regard to the spatial quality of the sound field representation.

The theoretically calculated representation of a plane wave is compared to the output of a simulated microphone array under ideal conditions. As the major result it should be pointed out that the quality of HOA coefficients acquired by a spherical microphone array shows a strong dependency from the angle of incidence. Variations of the array size, the number of capsules and the microphone capsules' size exhibit possibilities for improvements. The simulations made are confirmed by measurements using a commercially available sound field microphone.

Bauer, P.

• Spectral Restoration of Narrowband Speech Recordings Supported by Phonetic Transcriptions
  P. Bauer and T. Fingscheidt [C]
[A]

Due to limitations of the acoustic bandwidth, historic and telephone recordings suffer from poor speech quality and intelligibility. Nevertheless they are widely found in radio, television, internet, as well as in storage media in the context of newscast reports, documentations, or simply in large databases with archived speech material. Spectral restoration of such narrowband speech recordings improves the auditory impression as well as intelligibility. As the recordings are available offline, phonetic transcriptions already exist or can be extracted from the speech data manually or automatically and support the spectral restoration process. Hence, subjective quality and intelligibility much closer to wideband speech can be expected. In this paper, an artificial bandwidth extension (ABWE) is presented to spectrally restore 8 kHz sampled speech signals by upsampling to 16 kHz and estimating further frequency regions of interest. It makes use of the phonetic transcriptions in order to improve the partially insufficient ABWE performance of critical phonemes, i.e., mainly fricatives /s/, /z/ and /f/. We found the spectrally restored speech significantly enhanced, particularly the typical lisping effect disappeared in many instances.

Baumann, U.

• Comparison of Speech Intelligibility by EAS, Bimodal, Uni- and Bilateral Cochlear Implant Patients in a ’Multi-Source Noise Field’ (MSNF)
  T. Rader, U. Baumann and H. Fastl [C]
[A]

Speech perception in noise is one of the most difficult tasks for people suffering from hearing impairment. The Oldenburg Sentence Test (OLSA) is a useful tool to investigate speech intelligibility threshold in noise. In the present study, a multi-source noise field introduced by Rader et al. (DAGA2008) consisting of a four loudspeaker array with independent noise sources was combined with the OLSA. The MSNF allows to present a more realistic noise environment and shows a higher effect of binaural interaction regarding the separation of signal and noise from different directions.

Three different noise characteristics were applied in two conditions (S0N0: Signal and noise from the front; MSNF) to investigate speech reception threshold: Oldenburg noise (generated by superposed sentences of OLSA), CCITT noise, and speech simulating noise according to Fastl (1987).

Four different groups of cochlear implant patients separated into listening conditions unilateral, bimodal, bilateral and EAS (electric-acoustic stimulation) served as subjects in the present study.

Results showed a clear discrepancy between the examined patient groups for Fastl noise in MSNF ( SNR=11.0dB). For OLSA noise in S0N0-condition, the results differ less ( SNR=4.7dB). The MSNF with Fastl-noise is particularly suitable to measure the performance of various aided CI-patients in realistic sound environments.

• Hearing Impairment: Benefit of Optimized Beamforming obtained with Hearing Glasses in a Multi-Source Noise Field
  U. Baumann and T. Rader [I]
[A]

One of the main obstacles complained by persons suffering from hearing impairment is their highly degraded speech perception under adverse listing conditions. Even high end hearing aids with multi channel beamforming do not restore this ability sufficiently in complex noise. A recently introduced hearing aid embodied in a pair of glasses frame offers advanced directivity by means of four microphones mounted in line in each temple in order to obtain sophisticated beamforming (DI 9 dB). We were interested in the performance of this device in a complex noise field compared to behind the ear hearing aids equipped with custom directional microphones. Ten experienced users of different hearing aids with directional microphone technique were fitted with this type of hearing glasses. Speech perception performance was compared in quiet and in a multi-source noise field consisting of 4 independent sources with speech simulating noise according to Fastl (1987). Results showed for the hearing glasses significantly higher signal to noise ratios compared to the own BTE device condition. With the high end hearing glasses, some subjects were able to discriminate speech even at an SNR of -15 dB in a complex noise listening situation.

Baumgart, J.

• A Three Dimensional Model of the Organ of Corti of the Guinea Pig
  J. Baumgart, M. Fleischer, R. Gärtner and A. Voigt [C]
[A]

For the function of the inner ear the mechanical linkage of basilar membrane displacement to the deflection of the inner hair cell stereocilia is essential and investigated here. Besides the passive excitation by a pressure on the basilar membrane also the excitation by electromotile outer hair cells as well as by the hair bundles of the outer hair cells is studied.

Based on geometrical data of the guinea pig and experimental velocity measurements the organ of Corti is modelled by means of the finite element method. The fluid and the structure are modelled as linear material with small displacements. The interaction of the fluid and the structure is implemented as a matrix coupling. The geometry is extracted from laser scanning microscope stacks and material properties are taken from the literature and adjusted to match experimental results.

This model provides a helpful tool to understand experimental data more in detail. It is possible to look for the influence of geometrical variations and material properties. Different configurations and boundary conditions are modelled and the relation of excitation to inner hair cell stereocilia displacement due to bending is analysed.

• Effect of the structural dynamics of the bocal on the sound spectrum of a bassoon
  T. Grothe, J. Baumgart and R. Grundmann [I]
[A]

The bocal of a bassoon is a thin curved metal tube to which a double-reed mouthpiece is attached. When the instrument is played, the reed acts as a pressure controlled valve and due to the pulsating flow the bocal starts to vibrate. The outer contour of the bocal has a major influence on the modal shapes of the structural modes and the relation between the eigenfrequencies. The matter of interest in this investigation is the interaction of the structural dynamics of the bocal and the reed. We assume that modifications of the bocals geometry and mass distribution will lead to measurable changes in the sound spectrum. Therefore the modes of a bocal are computed by FEM and the frequencies are checked experimentally with a LDV. In a playing experiment mass was added locally on the bocal. Sound pressure measurements were performed, while a professional bassoonist played sustained notes on a modifiable bocal. Very small changes, less than 4 dB SPL, occur in the overtones above 1 kHz and can be shown using principal components analysis of the sound spectra.

Baumgartner, H.

• Interactive Fitting-Wizard in a Home Environment
  H. Baumgartner, A. Schulz, A. Hein, T. Herzke and I. Holube [C]
[A]

The European Hearing at Home (HaH) project aims at supporting hearing-impaired persons in their home environment. TV headsets with linear volume control allow people with hearing deficiencies good speech perception even in bad acoustic conditions. This approach attacks sensorineural hearing losses which imply a changed loudness perception combined with reduced perceivable dynamic range.

To help sensorineural users adapting the dynamic of their head phones according to individual hearing loss, three different interactive fitting procedures - based on broadcast audio material - have been evolved to fit parameters of a three band dynamic compressor performed in the master hearing aid (MHA). The MHA is the real-time audio-signal-processing framework in the background.

The fitting schemes can be used without any expertise and are intuitive and easy to use. A full walkthrough of the fitting wizard takes about 10 to 15 minutes. First results of user studies with 11 subjects (54-73 years old, 5 women, 6 men, slight to moderate hearing loss) show effectiveness of the approaches with respect to the improvement of the speech intelligibility and acceptance of sound quality. Further user studies will be performed.

Bechtold, M.

• Evaluation of Spatial Approaches for Improvement of Patient Speech Acquisition at Magnetic Resonance Tomography Systems
  M. Friedrich, K. Porzelt, M. Bechtold and P.-C. Eccardt [C]
[A]

During operation of magnetic resonance (MR) tomography systems, switching noise is produced with high sound pressure levels. This noise is a disturbance for patient-to-operator communication, as well as for examinations that require synchronous recording of patient speech (fMRI). Signal processing algorithms can be applied for the removal of MR and other noise sources from patient speech. However, the disturbed speech signal should be obtained with the highest possible input SNR, which is complicated by limitations in practical application. On a typical MR installation, the room acoustics and sound field characteristics were examined, with the objective of optimising the microphone solution with regard to patient speech pick-up. Applying coherence based measures, the noise field is characterised on and around the MR system. General conclusions are drawn about possibilities to employ spatial information for optimising the speech level with regard to interfering noise sources, such as using microphones with directional characteristics, or multiple microphones and source separation algorithms. The choice of suitable microphones and their positioning on a MR installation are discussed. In addition, experiment results are presented for SNR improvement on a real MR installation, using microphones with different directional characteristics on suitable position.

Beck, S.

• Flow-induced Sound Radiation from Air-ducting Structures
  S. Beck and S. Langer [C]
[A]

In product development prediction of flow-induced noise becomes an increasing important issue. Examples include both liquid and gas flow such as heating and sanitary ducts, ventilation channels and air intakes.

Here, a numerical approach is presented to determine this acoustic radiation. Once validated, it can reduce expensive prototyping and laboratory testing.

The model consists of the three interacting parts air flow, structure and surrounding space. Air flow within the structure is derived using the volume of fluid approach used by commercial software. The structure is represented by a shell model discretized by the finite element method (FEM). Acoustic radiation into the surrounding space is modelled accounting for the wave formulation and Sommerfeld's radiation condition employing the boundary element method (BEM).

Becker, S.

• Aeroacoustics taking Fluid-Structure Interaction into Account
  M. Kaltenbacher, S. Becker, J. Grabinger and F. Schäfer [I]
[A]

In the last years, manufacturers have started to consider the aerodynamic noise level in many industrial applications as a relevant design parameter (e.g. airplanes, cars, air conditioning systems, etc.). Because of the growing demand for reducing noise levels and for fulfilling noise regulations, there is a great motivation to investigate the basic aeroacoustic phenomena and mechanisms of sound generation and propagation. In many cases, the occurring noise is not only generated by the turbulent flow but also due to vibrations resulting from the interaction between the flow and the mechanical structure.

In the present work, we investigate the three field coupling for a flow past a thin flexible structure. Our developed numerical algorithms are capable to simulate the complex interaction between fluid, structure and acoustics. The scheme is based on a partitioned approach employing a finite-volume flow solver and a finite-element structural mechanics and acoustics solver. An important feature of the computational scheme is that it allows for a separate prediction of flow-induced and vibration sound. We will present detailed numerical studies of the interaction of the fluid flow with the flexible structure and the generated noise. Moreover, a comparison of computational results to our experimental investigations will be provided.

• Separation of Aero- and Vibroacoustic Noise from a Flexible Structure
  J. Grabinger, S. Müller and S. Becker [C]
[A]

The reduction of noise due to turbulent flow over a car's underbody is an important--yet complicated--issue in automotive engineering. Our objective is to study the mechanisms of noise generation at a simplified model of a car's underbody. For this purpose we mount a flexible rectangular plate into a rigid surrounding plate, which is then placed in a wind tunnel in parallel to the main flow direction. A square cylinder is mounted upstream of the flexible plate in order to model a strut in an underbody. This assembly generates noise by means of two mechanisms, namely the vibrations of the flexible plate and the turbulent flow.

These two mechanisms require different counter measures for noise reduction. Therefore, one has to separate the aeroacoustic and vibroacoustic contributions in the radiated sound. We measure the sound pressure with a microphone and, simultaneously, the normal deflections of the plate with a laser scanning vibrometer. On the basis of these data, we conduct FEM simulations to compute the vibrational sound at the position of the microphone. Subsequently, we estimate the contribution of the vibrational sound by comparing it to measured data. We find that the vibration generates dominant tonal components at the plate's resonance frequencies, while the flow-induced noise gives a broadband contribution.

• Sound Generation in a Flow-Induced Vibrating Human Vocal Folds Model
  S. Becker, S. Kniesburges, S. Müller and A. Delgado [I]
[A]

For investigating the fluid-structure-acoustic coupled (FSA) process of human phonation, synthetic models of the vocal folds (VF) were designed which showed flow-induced oscillations and produced a tonal sound. Two different kinds of models were developed: one for observing the flow field vibrating at constant frequency and a more realistic which could be stretched in longitudinal direction to simulate the prestress within human VFs. Besides the investigation of the flow field using PIV the major aim was to observe the FSA process to determine the acoustic sources. Therefore two approaches were applied: Firstly the vibration of the VFs was recorded using a high-speed camera triggered by the subglottal pressure to determine phase shift between the two signals. Secondly synchronous measurements of the subglottal pressure, the flow velocity behind the VFs and the acoustic pressure were done to determine their coherence. The results support the existence of the Coanda-effect. Furthermore the vibration frequency and the phase shift between the subglottal pressure and the glottal width depended on the stiffness and the tension in the VFs. The synchronous measurements determined the pulsating flow rate as main acoustic source.

Becker, T.

• Speaker Verification Based on Formants Using Gaussian Mixture Models
  T. Becker, M. Jessen and C. Grigoras [C]
[A]

In a new method, the well known UBM-GMM framework for automatic speaker verification is applied to formant parameters replacing the common cepstral parameters (e.g. Mel frequency cepstral coefficients (MFCC)). By using full covariance matrices, both the static and variable vocal tract configurations are reflected in the speakers' models. Speaker comparisons are achieved by relating speaker similarity to speaker typicality, expressed as the likelihood ratio. The method reveals a low equal error rate while providing easy interpretability and low dimensionality.

Becker-Schweitzer, J.

• Just-Noticeable Roughness Differences of Technical Sounds
  A. Leutheuser, R. Sottek and J. Becker-Schweitzer [C]
[A]

Roughness has become a central focus for product sound design. The roughness-induced perception, varying from a sporty character to a very unpleasant impression, consistently proposes new questions and challenges in sound engineering. This paper is concerned with one of the important aspects of subjective roughness perception: the investigation of just-noticeable roughness differences. Listening tests using synthetic sounds (modulated sinusoids with additional noise) and technical sounds (based on engine sounds) have been performed, both dependent on parameters such as degree of modulation, sound pressure level, and signal-to-noise ratio. The influence of a single parameter and of a combination of parameters was studied using adaptive test procedures providing more reliable and precise results within a limited time for the subjective tests. The results of the studies will be discussed. Further investigations relate to using the results for optimizing the roughness calculation based on the Hearing Model according to Sottek with respect to technical sounds.

Bederna, C.

• Contributions to a Better Understanding of Tire Cavity Noise
  C. Bederna and E.-U. Saemann [I]
[A]

Beentjes, W.

• Certified sound insulation of different building elements
  W. Beentjes [I]
[A]

In order to develop new building products one needs to know sometimes the sound insulation in situ. Since 1980 special procedures are developed for different types of building products in the Netherlands. The two key-issues of these procedures are :

1. The statistics of predicting the sound insulation of a large series on basis of a random test, to get reliable results. The measurements were made according to ISO 140 part 4 and 7.

2. The contribution of the building element to the total sound insulation must be determined and/or the test situation must be described in order to get comparable results.

The procedures for inner walls, complete building systems and roof elements are presented and discussed. Ground floors and other floor elements and facade-elements are acoustically classified according to Dutch codes of practice and/or measurements according to ISO 140 part 3.

Beer, D.

• Electro Active Materials for Flat Panel Transducers
  L. Ehrig, D. Beer and S. Brix [C]
[A]

Electroactive materials are a class of materials that change their geometry due to the presence of an electrical field. A well known effect that occurs in crystals, polarized ceramics and plastic materials is piezoelectricity. Since there is no center of symmetry and a preferred direction in such materials respectively an electrical voltage applied on the materials surface yields a deformation. The same macroscopical behaviour can be observed in relatively new materials called ferroelectrets. Voids in these cellular polymer foams are internally charged and form giant dipoles. The dipoles are deformed by an electrical field likewise in piezoelectric materials. Electroacoustic transducers based on electroactive materials may have a unique feature: driver and membrane are one part. Thus it is possible to build extremely simple, compact and flat transducers. Using the material by means of a thickness oscillator even the rearward air spring, causing a major part of problems for thin transducer constructions, can be avoided. In this work several approaches for transducers using electroactive materials are presented.

Beerends, J.

• PESQ Based Speech Intelligibility Measurement
  J. Beerends, R. van Buuren, J. van Vugt and J.A. Verhave [I]
[A]

Several measurement techniques exist to quantify the intelligibility of a speech transmission chain. In the subjective domain the CVC score is mostly used. In the objective domain the Articulation Index and the Speech Transmission Index (STI) are standardized for predicting intelligibility. The STI uses a signal that contains spectro-temporal characteristics similar to natural speech. Comparing intensity fluctuations of degraded and reference signals, the modulation transfer function is derived from which the STI can be calculated. In speech transmission, coding is used that may behave differently for speech then for the STI test signal, implying a possible incorrect intelligibility prediction. A more fundamental approach in assessing intelligibility is to take natural speech signals, and derive internal representations of the reference and degraded from which the difference can then be used to estimate the intelligibility. In this paper we investigate the idea to use a testsignal composed of concatenated CVC words, that are also used in the subjective CVC test, in combination with PESQ to measure the intelligibility. Normally testing a speech link requires long subjective listening sessions with CVC words in carrier sentences, while a compact testsignal approach in combination with PESQ allows a fast measurement.

Behr, W.

• The project 'Low Noise Train on a Real Track (LZarG)'
  W. Behr [C]
[A]

Deutsche Bahn has the goal of halving the rail noise exposition of lineside residents by the year 2020 compared to 2000. Therefore the project ''LZarG'' investigates in technical improvements of boogies, wheelsets, wheel-absorbers, rail-absorbers, rail fastening systems and special under sleeper pads. Besides the German noise mitigation programme for existing lines and retrofitting the freight wagons with composite brake blocks, LZarG is the third column in the strategy of Deutsche Bahn to reduce railway noise by at least 10 dB(A) by 2020 compared to 2000. Deutsche Bahn will assure that each development within LZarG meets the relevant regulations and the special requirements of Deutsche Bahn. Also the potential impact on future maintenance procedures will be taken into account. At the end of the project in 2010, several field test measurements will evaluate the potential of the improvements developed or improved within LZarG, using a noise-optimized trainset on the various track systems.

Behrendt, R.

• Advances in Ultrasonic Measurements Using Laser Doppler Vibrometer
  R. Behrendt and H. Steger [C]
[A]

Laser Doppler Vibrometry is a widely used technique to exermine ultrasonic waves in solids , liquids and gases. The main advantages using the Laser Doppler Vibrometry in the field of ultrasonic measurements are the constant sensitivity over a very wide frequency range, the non contact measurement and the high spatial resolution. Different application examples prove the high perfomance of this measuring technique. The sensitivity and the resolution of different ultrasonic Laser Doppler Vibrometer will be discussed among thmeselves and in comparison with standard ultrasonic detectors. Suggestions for the improvement of the signal-to-noise ratio are given and verificated experimentally.. Finally the properties of new high-performance Ultrasonic-Laser-Doppler-Vibrometer systems will be discussed and new appülications will be pronounced.

Bein, T.

• LOEWE-Zentrum AdRIA – An important step towards the commercialization of adaptive systems
  J. Bös, T. Bein and H. Hanselka [C]
[A]

The LOEWE-Zentrum AdRIA (Adaptronik - Research, Innovation, Application) is a large interdisciplinary research project funded by the government of the German federal state Hessen with a duration of 3 years and a budget of 17.7 million euros for the time being. A second phase of three years with a tentative budget of 20 million euros is planned. It involves the Fraunhofer Institute for Structural Durability and System Reliability LBF, 22 institutes from 3 different departments of the Technische Universität Darmstadt as well as one department of the University of Applied Sciences, all located in Darmstadt, Germany. The goal of this project is the creation and sustainable implementation of an internationally leading adaptronics research institution in Darmstadt. Eventually, it will lead to the foundation of a new Fraunhofer Institute for Adaptronics and of four new university institutes. The project is organized in ten technology areas (e.g., materials, sensors and actuators, systems and control, embedded systems, simulation tools, reliability), which supply the necessary technologies for three application scenarios, i.e., adaptive car, quiet office, and adaptive tuned vibration absorber. An important step towards the commercialization of adaptive systems will be the close cooperation with companies and suppliers during the second phase of the project.

Beissner, K.

• Metrology of HITU fields
  K.-V. Jenderka, K. Beissner and J. Haller [C]
[A]

High Intensity Therapeutic Ultrasound (HITU) is a non-invasive technique for the treatment of various types of cancer, as well as non-malignant pathologies by inducing localized necrosis of the tissue. Even though HITU is already in clinical use, this technique is still an emerging therapeutic procedure, with a lack of knowledge on reliable and stable treatment planning and monitoring procedures. A crucial step for the prediction of the desired temperature rise in a HITU-treated region is the measurement of the ultrasound output power. The evaluation of tissue thermometry methods, using MR or ultrasound itself, is important for the treatment monitoring, as well as for choosing the appropriate ultrasound power for a desired local intensity and temperature rise respectively. While ultrasound fields are well-investigated for output powers smaller than 50 W, this paper describes first results of measurements in much stronger, highly focused fields ( 300 W ultrasound power at 1.5 MHz) using an absorbing target arranged close to the transducer. The measurement uncertainties are discussed.

Bekke, D.

• The development of a CPX measurement set-up capable of measuring tire-road noise effectively
  D. Bekke [I]
[A]

In the Directive 2001/43 limits are put on the maximum noise level of a pass-by noise measurement on an ISO 10844 surface. This prescribed measurement set-up only gives an LAmax level of the total passage of a whole vehicle equipped with four tires and does not provide enough information for a problem analysis or model validation. Measuring on one tire with (closed or open) CPX-trailers or -cars gives a better insight into tire-road noise. The difficulties with these measurement set-ups can be the noise generation of the set-up itself, the acoustics of the trailer-interior or vibrations directly disturbing the microphone signals. A CPX measurement set-up has been developed keeping these considerations in mind in order to be able to do proper problem analysis and model validation.

• A source path contribution analysis on tire noise using particle velocity sensors
  H.-E. de Bree, D. Bekke and J.H. de Vries [C]
[A]

Road tire noise is an important topic of research where acoustic particle velocity based testing techniques can be expected to bring new insights. Modal analysis can be carried out using non contact particle velocity sensors, and PU sound probes can be used to measure the radiated sound without a need to use anechoic testing conditions. A further breakdown of the overall sound pressure levels measured in to its various sources can be made by applying a source path contribution analysis, using the PU probes to measure velocity, sound intensity and, for determining the reciprocal transfer path, the sound pressure. The concept of using this type of transfer path analysis will be outlined and illustrated by a tire noise case.

Bellomini, R.

• Harmonized Strategies for the Development of Noise Action Plans
  S. Luzzi and R. Bellomini [C]
[A]

In this paper an approach for developing Noise Action Plans, with special regard to hotspots and quiet areas detection and delimitation, is presented. Authors have been involved in the preparation of Noise Action Plan for the agglomeration of Florence and in the management of an Urban Matrix of targeted knowledge exchange for the Eurocities Network. A specific procedure for hotspot detection in areas annoyed by concurrent sources has been developed and criteria for selection of quiet areas in noisy city centres have been found.starting for their experiences the approach considers area characterisations and filters helping to achieve priorisation, such as noise level filter, minimum area filter, etc. The main actions of the Florence case study plan are then shown, together with a basic method for their harmonization with those of other noise reduction and noise control plans, required by National laws and to be applied to the same areas.

Berckmans, D.

• Quantification of the Equivalent Sources of Different Tire Noise Models by means of the ASQ Technique
  D. Berckmans, P. Sas and W. Desmet [I]
[A]

Tire noise is a predominant noise contributor in many traffic situations nowadays. Accurate models are therefore needed to predict the spectrum that results at receiver locations due to the tire rolling under different operating conditions. Predicting the spectrum relies on a source - transmission path - receiver model which involves the modeling of all noise sources and the determination of all relevant transfer paths.

Locating and quantifying the different noise sources on an operating tire remains an inherent difficult task. The sources are thought to be numerous and complex, with their relative magnitudes changing with tire specifications, road specifications and operating conditions.

A rolling tire was experimentally characterized on a roller bench by means of the substitution monopole technique: the running tire is substituted by the non-operating tire covered by monopoles. The monopoles are quantified by means of the airborne source quantification (ASQ) technique, based on operating indicator pressure measurements in the neighborhood of the rolling tire. The role of regularization techniques, important in the quantification process, is highlighted and results for different tire models are presented. Moreover, the contributions of the different tire segments to the resulting spectrum at the receiver position are evaluated.

Bérengier, M.

• DEUFRAKO 'Prediction and Propagation of Rolling Noise'
  M. Auerbach and M. Bérengier [C]
[A]

This project aims at the development of new road pavement concepts with respect to tyre-road noise in a wider context of the abatement of road traffic noise. French and German institutes headed by the ''Laboratoire Central des Ponts et Chaussées'' (LCPC) and the ''Bundesanstalt für Straßenwesen'' (BASt) cooperate throughout this project. The theoretical approach is based on the SPERoN (''Statistical Physical Explanation of Rolling Noise'') model developed by one of the partners. In a first stage, the model was validated for different kinds of French road surfaces. In a second step, the model was applied as an adapted tool to design new textures for low noise road surfaces. Finally, in a third step, the results issued from SPERoN were used as an input for outdoor sound propagation models developed in France and Germany in order to estimate the effect of this optimised model-aided low noise pavement in the far field, close to the façades. A ranking with other typical German and French pavements has lead to a common database (DEUFRABASE). The results of the project will be presented.

Bergmans, D.

• Background noise: an increasing environmental problem
  J. Jabben, D. Bergmans, E. Schreurs and T. Koeman [C]
[A]

Noise quality in most studies is related to the average noise level (Lden) only, although this indicator merely expresses the average intensity of the soundscape. A complementary feature of increasing importance is the background noise level, which is the noise level that is exceeded during 95 % of the observation time. Inhabitants of areas with a high background noise level may be more vulnerable to annoyance or health effects, as recuperation times are narrowed. Due to strong growth of road, railway and air traffic, background noise levels have steadily increased in the Netherlands. In order to assess the impact of environmental noise levels from transport sources, RIVM in cooperation with the Dutch National Aerospace Laboratory have started a research project in 2007 with the aim of setting up background noise maps for road-, railway- and air traffic. This paper discusses the mapping and characteristic behavior of background noise. The implications of the background noise levels on health and annoyance will be discussed. The results indicate that a further increase of background noise levels in the near future is to be expected that may affect the Dutch population. Hence an intensified reduction policy of noise emission is most urgent.

Bethke, C.

• Technical aspects in the qualification of free-field environments
  C. Bethke and V. Wittstock [C]
[A]

Free-field environments are usually qualified according to annex A of ISO 3745. The sound pressure decrease with an increasing distance from a point source is compared to the ideal free-field behaviour. The tolerated deviations of 1,0 dB to 3,0 dB depend on the frequency but also on the type of the room to be qualified. In the last years, PTB has developed and optimised a measurement setup to perform this test. In view of the small tolerances, measurement uncertainties must be small. This requires special technical solutions which will be overviewed in the presentation. One focus is on the different sources and the achieved directivities and stabilities. Other important points are the development of the multi-sine test signals, the automatic microphone traversing system, the background noise handling and the data acquisition and analysis. Finally, some selected results of room qualification measurements will be presented.

Betke, K.

• Measurements of Construction Noise During Pile Driving of Offshore Research Platforms and Wind Farms
  R. Matuschek and K. Betke [I]
[A]

Beutelmann, R.

• Modelling binaural speech recognition
  T. Brand, R. Beutelmann and B. Kollmeier [I]
[A]

In many cases speech has to be recognized in adverse listening conditions including noise sources from different directions and adverse room acoustics. Furthermore, hearing-impairment compromises intelligibility in such conditions. Binaural processing of the auditory system can cause a substantial improvement of speech recognition. A binaural speech intelligibility model (Beutelmann and Brand, JASA, 2006) yields accurate predictions of speech reception thresholds (SRTs) in presence of a stationary noise source at arbitrary azimuths and in different rooms. The model combines a multi-frequency channel equalization-cancellation (EC) process and the speech intelligibility index (SII). The effect of hearing loss is simulated by adding uncorrelated masking noises (according to the pure-tone audiogram) to the ear channels. A recent revision of this model is based on an analytical expression of binaural unmasking for arbitrary input signals and is computationally relatively efficient. An extension for non-stationary interferers was realized by applying the model to short-time frames of the input signals and averaging over the predicted SRT results. Model predictions are compared to binaural speech intelligibility data from normal-hearing and hearing-impaired listeners, incorporating different combinations of rooms, sound source setups and noise types.

Bietz, H.

• Experiences with characterising Simple Sources of Structure-borne Sound by the Two Plate Method
  H. Bietz and V. Wittstock [I]
[A]

The assessment of the ability of a vibratory source to inject sound power into a receiver is a major task in different fields of acoustics. Recently, a proposal has been made to characterise structure-borne sound sources by two properties, an activity and a mobility quantity. They are determined by connecting the source under test to two different receiving plates with very different point mobilities. From the different sound powers injected into the different receiving plates, source quantities can be calculated. At PTB, it was checked whether this method is capable to determine the source characteristics of the very easy case of an electrodynamic shaker. This source has the advantage that it can be modelled by electromechanic analogies and thus, source properties can be determined by a second method. Additionally, the power input into the receiving plates can be measured by an impedance head and the electric input power is also known. Furthermore, impulse signals can be used with shakers as well as stationary ones which gave the opportunity to investigate whether the two plate method works in all cases. Presented results will include comparisons between sound powers and source quantities determined by different methods for stationary and impulse signals.

• Side Effects of the ISO Tapping Machine as a Walking Noise Source
  W. Scholl, H. Bietz and V. Wittstock [I]
[A]

Characterising noise from different floors, when people are walking on them, has become important in some countries, as there exist legal requirements. Finding an adequate measurement method was made a work item of an adhoc-group of CEN TC 126 / WG 1. Based on former research, it is proposed to use the standard tapping machine according to ISO 140 as a walking noise source. As this kind of noise has to be measured in the same room where the the walker acts, the tapping machine might unwantedly influence the sound from the floor in two ways: On the one hand its body might interfere with the sound radiation from the floor covering by shielding or resonance effects, or on the other hand it might produce noise by itself ('self-noise'). A particular problem is, that the self-noise of the tapping machine depends on the surface it hits on. So it has to be determined on a very silent floor mockup. Investigations were carried out at Physikalisch-Technische Bundesanstalt Braunschweig (PTB) and some results as well as a proposal for such a mockup are presented.

Bijsterbosch, K.

• Room Acoustic Scale Model Measurements Using a “Spark Train”
  C. Hak and K. Bijsterbosch [C]
[A]

The spark gap, used as a sound source in scale model investigations, has advantages and disadvantages. One of the advantages is its limited size, which is particularly important for investigations at small scales. Also, the spark gap is omnidirectional and can be constructed such that the shape has a negligible influence on the sound field to be measured, for any scale factor. A disadvantage of an excessive discharge is the appearance of a shockwave. This leads to non-linear effects that make it impossible to acquire reliable impulse responses at small distances from the spark gap. A second disadvantage is the spread in successive separate impulse levels as a result of the micro-climate near the spark gap. Experiments were performed with a low energy spark gap (maximum electrode separation of 8 mm). In these experiments the energy was spread over a longer time interval through the use of spark trains. This allows for high INR values and reliable measurements of sound pressure levels. The period stability (jitter) of the spark train and spread in strength of the individual discharges were studied.

Bisitz, T.

• Noise reduction for media streams
  T. Bisitz, T. Herzke, M. Zokoll, A.-M. Öster, S. Al Moubayed, B. Granström, E. Ormel, N. van Son and R. Tanke [C]
[A]

In the Hearing at Home project the central Home Information and Communication platform is developed to support hearing-impaired persons with easy to configure ``Supportive Audio Signal Processing'' and visual support on a TV screen in home environments. Amongst other, individually configurable algorithms, several algorithms can be used for noise reduction of the media streams without individual fitting to enhance the accessibility of the presented audio materials using the Master Hearing Aid as a real-time audio-signal-processing framework: (1) Single-channel noise reduction according to Ephraim-Malah together with a speech pause detector, (2) stereo noise reduction additionally removing the central signal in a control path, (3) frequency shaping noise reduction by emphasising a frequency range important for speech intelligibility, (4) time-varying attenuation and (5) noise reduction by selection of channels. It will be possible to classify the acoustic situation in the media stream automatically in real-time using six classes allowing to activate the appropriate noise reduction strategies. In user tests the performance of algorithm 1-3 is evaluated: Results of sound quality assessment, listening effort estimation and speech intelligibility measurements with hearing impaired subjects from groups with different types of hearing impairment will be presented.

Bitzer, J.

• Multi-channel noise reduction for binaural hearing aids by using short-time spectral attenuation combined with noise estimators for non stationary noise
  S. Franz, J. Bitzer and U. Simmer [C]
[A]

Multi-channel noise reduction (NR) with binaural output is an ongoing research topic for future hearing aids. Especially the preservation of interaural level and time differences is important in order to keep the spatial impression intact. In this contribution we present an extended noise reduction algorithm based on the two-channel NR scheme from Kim et al. The algorithm has a structure similar to the generalized sidelobe canceler (GSC) structure, however, instead of reducing the noise components by cancellation, short-time spectral amplitude filtering is used. We extended and tested this algorithm with different compensation filters to reduce the estimation error as well as musical noise. As a new approach a compensation filter based on minimum statistics will be presented. This method has the advantage that no voice activity detector is necessary. The evaluation will compare different filter estimation techniques on a realtime developing platform for hearing aid algorithms. Finally, we will discuss the advantages and drawbacks of the compared methods.

Blanchet, D.

• Building SEA Predictive Models to Support Vibro-Acoustic Ship Design
  D. Blanchet and A.S. Matla [C]
[A]

In the past, the shipping industry has used empirical models extensively to predict the vibro-acoutic response aboard ships. This method has proven effective for reasonable ship size, shapes and typical materials. However, empirical approaches offer little flexibility and are limited to known construction types and materials. Nowadays, ships are getter larger and use more sophisticated material such as composite. Regulation and ship owners impose more stringent vibro-acoustic performance of their vessels and the luxury yacht segment is no exception. In addition, the time available to design the vibro-acouctics package aboard a ship is shrinking forcing designers to seek industrially efficient methods to build ship models and perform design analysis.

This paper presents an efficient method for quickly building a detailed predictive vibro-acoustic model from 2D or 3D data. This model includes all SEA structural subsystems of the hull, superstructure, interior bulkhead, etc. It also includes SEA acoustic subsystems of all cabins and other living areas. Structureborne and airborne sources are defined for all types of excitation encountered in a ship such as engine, generator, HVAC, bow thrusters, propeller.... Vibration, SPL and power inputs are available at any locations and serve as basis for vibro-acoustic design analysis.

• Modeling the Vibro-Acoustic Effect of Trim on Full Vehicle and Component Level Analysis
  D. Blanchet, M. Anciant and L. Mebarek [C]
[A]

In the automotive industry, the influence of poro-elastic components on acoustic comfort has been mostly investigated for air borne noise at mid- and high frequency ranges; however, due to the lack of adequate theoretical formulations, the influence of poro-elastic in numerical vibro-acoustic simulation at lower frequency range has often been ignored or simplified by the use of distributed spring/mass on the BIW structure and impedance on the acoustic medium. Recent theoretical developments enabled to overcome this limitation by providing an efficient FEM formulation for poro-elastic material modelling. This FEM approach, implemented in VTM (Vehicle Trim Modeller) software developed by ESI-Group, enables the computation of the coupled response of a fully trimmed vehicle by taking into account the BIW structure, the acoustic cavity and the poro-elastic components (seats, carpet, dash insulator...).

This paper presents comparison between measurement and VTM simulation results for fully trimmed vehicle. It also presents the recently developed transmission loss (TL) module using FEM/BEM approach and comparisons between test and VTM prediction for academic and industrial cases.

Blanco, J.L.

• Reworking spoken dialogue systems with context awareness and information prioritisation to reduce driver workload
  J.L. Blanco, Á. Sigüenza, D. Díaz, M. Sendra and L. Hernández [C]
[A]

An entire host of new information, communication and entertainment applications and security systems are assaulting our cars. Interaction with onboard systems is moving beyond the classic switches, knobs, sticks, pedals and dashboard lights and sounds, to incorporate speech, touch-screens and a variety of other displays. While some of these systems are intended to help with the task of driving (e.g., navigation aids, lane departure and vehicle-in-blind-spot warnings), a constellation of displays, messages and requests threaten to overload the driver with information that is too rich and may distract him. Speech being eyes- and hands-free, it is increasingly being considered as a safe communication channel. However, interactive dialogue can increase the driver's mental workload, especially when robustness problems occur (e.g., recognition errors). Here we present our research results on dialogue as a secondary task, to gain understanding on how it may affect driving, on what dialogue strategies foster safer driving and on which applications, messages and interaction modalities should be prioritised over which others in each driving context. The proposed contribution is related to our current research activities within MARTA (Mobility for Advanced Transport Networks; www.cenitmarta.org), a Spanish, publicly-funded project. Experimental results from tests with a simulated driving platform are presented.

Blau, M.

• Effect of airborne sound on installation noise - Part 1: Basic investigations
  M. Ebersold, L. Weber, S. Öhler and M. Blau [C]
[A]

The noise excitation of buildings by service equipment takes place by both structure-borne and airborne sound. Mostly sound transmission is determined by the structure-borne portion. For some sources, however, the contribution of airborne sound can't be neglected. A typical example is the EMPA-hammer, defined in the swiss standard SIA 181 as a standardized source for the simulation of user noise. If airborne sound contributes substantially, conventional measures for noise reduction based on elastic isolation between source and building won't work properly. In order to develop efficient measures for that kind of excitation, structure-borne and airborne sound must be separated and investigated in context with each other.

The separation of structure-borne and airborne sound was performed by means of a simplified experimental set-up consisting of two parallel plates acoustically connected by an adjustable structure- borne sound bridge and by the air gap between the plates. Since the receiving plate was bedded on an elastic support the transmitted sound power could be determined directly from the mean velocity level on the surface of the plate. For investigation of sound transport the experimental conditions such as distance of the plates or damping and division of the air gap were varied systematically.

• Methods to characterize the acoustic properties of periodic surfaces
  K. Hengst, H. Drotleff, R. Wack and M. Blau [C]
[A]

Sound striking a periodic surface is absorbed and scattered depending on the surface properties. In comparison to non-periodic surfaces, periodic surfaces cause a higher sound absorbtion, due to their nearfield: If the nearfield is large, the scattered soundfield will be small and the amount of the absorbed sound high. If the scattered sound field is large, the nearfield is becomming smaller and the periodic surface is less absorbing because more sound energy is reflected. The mechanisms of sound absorption and acoustic scattering by periodic surfaces consisting of bands of alternating impedance are demonstrated first by means of a numerical model. Subsequently, measuring methods are presented to determine the acoustic properties of periodic surfaces at normal, diffuse and directed sound incidence. These measuring methods are assessed with regard to accuracy, informational value and practicability and compared to the theoretically calculated values. It is shown that it is possible (with varying accuracy) to determine the structural absorption at normal, diffuse and directional incidence. The geometrical reflection can be separated from the scattered field at diffuse and directional sound incidence. None of the measuring methods presented allows the separate examination of the near and farfield. The measurements essentially confirm the calculated values.

• Predicting the acoustics of individual ears for hearing aid and audio applications - model framework and future work
  M. Blau, T. Sankowsky, P. Roeske and S. Fischer [C]
[A]

Applications involving sound delivery to the ear, such as hearing aids, mobile phones, ear- and headphones, will always have to cope with the problem of varying acoustic loads, due to interindividual differences in shape of the outer ear and mobility of the ear drum. It has been known for a long time that this variability can lead to considerably different sound pressure levels at the ear drum for frequencies above a few kHz. An additional challenge is the more and more frequent use of ear molds which do not acoustically seal the ear canal (''open fittings'' in hearing aid terminology). In this paper, a simple 1-dimensional model framework is presented, which may be used to predict individually different acoustics (including the pressure at the ear drum and acoustical feedback), or for real-time simulations of hearing instruments. The discussion will include known model sub-components, as well as open questions concerning the needed ''individuality'' parameters, or resulting for instance from violations of the 1D assumption in some parts of the model. Some of the open questions will be dealt with in the companion papers by Sankowsky et al. and by Roeske et al.

• Estimation of the Sound Pressure at the Ear Drum for Hearing Aid Applications
  T. Sankowsky, M. Blau, P. Roeske, H. Mojallal, J. Schroeder, M. Teschner and M. Bokemeyer [C]
[A]

The sound pressure at the ear drum is a reference quantity not only in hearing aids but in virtually all applications involving sound delivery to the ear, such as mobile phones, ear- and headphones. As the direct measurement of the ear drum pressure is too elaborate, it must be predicted in some way. One approach to do so is to establish a 1D model of the residual ear canal, based on the measured acoustic impedance at the interface between the ear shell and the residual ear canal. In a study on 10 human temporal bones with closed ear shells, 1D models established this way were used to predict the drum pressure in the 100Hz..10kHz frequency range. The models were all based on the same measured impedances, but several variants of algorithms to derive the ear canal model were employed. The model predictions were then compared to direct measurements of the drum pressure. It turned out that while the robustness of the different algorithms varies, their mean predictions do not differ substantially. The agreement with the direct measurement will be discussed.

• Modeling the sound field in front of the human ear drum for hearing aid applications
  P. Roeske, M. Blau, T. Sankowsky, H. Mojallal, J. Schroeder, M. Teschner and M. Bokemeyer [C]
[A]

The sound pressure at the ear drum is a reference quantity not only in hearing aids but in virtually all applications involving sound delivery to the ear, such as mobile phones, ear- and headphones. As the direct measurement of the ear drum pressure is too elaborate, it must be predicted in some way. Most attempts to do so exclude the influence of the vibrating drum from their analyses, although it is known that the ear drum vibrates in a complicated manner. In this paper, a finite element (FE) model comprising a detailed representation of the vibrating drum while preserving enough flexibility to be applied to a number of dead and live human specimen is presented and validated. The proposed method involves the representation of the vibrating drum by velocity sources whose source quantities are in turn based on laser doppler vibrometry (LDV) data measured on the respective specimen. The sensitivity of this method to calibration errors, to the effect of missing drum velocity data (towards the edges) and to deviations between the direction of measurement and the predominant direction of drum vibration is analyzed via comparison to a complete FE model (comprising the middle ear and the cochlea load).

Blauert, J.

• Critique of Pure Psychoacoustics
  J. Blauert and R. Guski [I]
[A]

Since product-sound quality has become a major engineering issue, psychoacoustics is widely applied in industrial R&D processes. It can. however, be observed that engineers which apply psychoacoustics sometimes have a rather naive understanding of the foundations and, thus, the potency of psychoacoustic methods. A common belief is, e.g., that psychoacoustic quantities can reliably be assessed with instrumental methods. The notion behind this belief is that, in laboratory experiments, ''unbiased'' versions of psychoacoustic quantities can be identified and assessed. It is the task of cognitive psychology to point to the risks that such an over-simplifying notion of psychoacoustics implies, and to provide knowledge and means to accordingly select and refine the psychoacoustic methods applied in engineering. It will be discussed that psychometric methods which are suitable for obtaining meaningful and valid information on product-sound quality vary substantially with the degree of abstraction of the references.

• Psychoacoustics Without Psychology?
  R. Guski and J. Blauert [C]
[A]

Bless, M.

• Prediction of structure-borne sound and secondary airborne sound by three dimensional finite element models
  M. Bless and F. Koopmans [C]
[A]

Nowadays sensitive utilisations as residential, cultural or office buildings often are located near traffic infrastructure. This situation leads to a rising noise and vibration exposure in urban environment. Typical vibration sources in urban environments are road and railway traffic, industrial sources or construction works. The prediction of structure-borne sound and secondary airborne sound in buildings can be difficult and uncertain due to the numerous characteristics of the source, transmission path and receiving structure. The two and three-dimensional finite element method (FEM) is a prediction procedure which is able to simulate high complex transfer paths and/or receiving structures. Models and methods based on the FEM will be introduced to predict structure-borne and secondary airborne sound in complex buildings.

Blome, M.

• Creating Temporary Venues for High Demanding Classical Concerts using a Room Enhancement System
  G. Engel and M. Blome [I]
[A]

In the field of classical concerts an increasing tendency of arranging concerts in unusual and sometimes unsuitable venues can be observed. There are various possibilities to acoustically improve these venues, but most of them turn out to be impracticable due to the limited preparation time and an unacceptable appearance of the acoustical measures. In many cases the only solution can be provided by the use of room enhancement systems, which counteract the acoustic deficiencies with electro acoustical means. The lecture presents case studies of two festivals in St. Moritz and in Abu Dhabi, where ball rooms, dining halls or auditoria of first class hotels are used as venues for top level orchestras and soloists. Beside the benefit for the musicians and concert audience the use of enhancement systems helps acousticians to improve the knowledge about the correlation between measurement results and the audible impression.

Bocquillet, A.

• Validation of a measurement based Area Contribution Analysis
  H. Jakob, A. Bocquillet, S. Marburg and H.-J. Hardtke [C]
[A]

The Automotive Industry is in need of effective NVH measurement methods to optimize the design of isolation and damping concerning weight and cost constrains. In order to gain better knowledge on the contribution of each panel from the interior it is imperative to find a new and simple method of analysis in the early prototype stages of a vehicle. Using the technology of the Microflown-element it became possible to grasp the particle velocity close to any surface based on measurements. This information and the quantification of the transfer path allow for a prediction on contribution from subsystems in the cabin. The method is based on a combination of the Transfer Path Analysis principle and the quantification of the Emission of the surface. A simple example below shows the validation of such measurements as well as the comparison between a direct measurement and a simulation. This example depicts the critical range of the mid frequencies (200-3500Hz). The precision and usability of these simulated results will be discussed moving forward based on this specifically designed setup.

Boeckx, L.

• Surface and guided waves in porous materials
  J. Descheemaeker, L. Boeckx and W. Lauriks [C]
[A]

The last decades, a lot of research has been done about the propagation of sound in heterogeneous and porous materials. Especially the two-phase nature of the poro-elastic material leads to interesting physical phenomena. During recent years, several measuring methods have been developed to determine the material parameters. However, a lot of problems aren't solved yet. Because of the visco-elastic behavior of a lot of these materials, the elastic moduli will become frequency dependent (rubberlike behavior at low frequencies and glasslike behavior at high frequencies). Experimental data about the frequency- and temperature dependence of the elastic moduli are scarce because of experimental difficulties. Some experimental and numerical results about this subject will be presented . Porous materials still have to be studied in a lot of configurations. One of them is a porous material in an elastic cylinder. Some numerical results about the dispersion curves of the wave propagating in the system will be presented.

Böhme, S.

• Practical Approach for Application of Impedance Control for Noise Attenuation
  M. Sobon, S. Böhme and D. Sachau [C]
[A]

Active noise control (ANC) becomes more and more an important issue, because of increased restrictive regulations regarding noise emission, e.g. in employment protection. Instead of using passive noise reduction methods, which cannot always be adopted, active control applications can provide efficient sound pressure attenuation in the lower frequency range. The cancellation of a disturbing primary noise field by a destructively interfering secondary sound pressure at the area of interest is widely known and has been the subject of many publications. In the early 90's several paper described the theoretical work on other control mechanisms like power absorption and power output minimization. This paper deals with a practical approach for an ANC-system that uses the radiation characteristics of a loudspeaker, which is applied as secondary sound source. The method is based on measuring active sound intensity at the centre axis of the loudspeaker membrane. For the practical implementation the signals obtained from a phase-corrected pair of sensors consisting of an accelerometer and a microphone serve as input for an adaptive control algorithm.

• Concepts of Global Noise Control in Cabins
  J. Greßkowski, D. Sachau and S. Böhme [C]
[A]

Nowadays people have to be protected for high sound pressure levels at low frequencies. An example is a propeller driven aircraft, where the propellers induce periodic noise inside the cabin. Therefore active noise reduction systems are applied to reduce these unwanted disturbances. To reach an acceptable noise level in the whole cabin actuators and sensors have to be allocated global in the distributed area. Also large numbers of actuators and sensors are required for attenuation of unwanted sound in a large volume with a high density of acoustic modes. This results in a high computational load for the controller. In order to reduce the number of components and consequently the effort of the controller unwanted sound transmission in cabin can be reduced in the propeller plane by controlling the radiated sound power. This innovative concept is investigated using an A380 structure as primary source which radiated noise due to excitation. The control sources are allocated in front of the panel. Also the microphones were placed in the vicinity of the structure. The possibility and first experimental results is shown in this research.

Böhnke, F.

• Acoustic Streaming in a Viscous Fluid-Structure System
  F. Böhnke and C. Gerstenberger [C]
[A]

The numerical evaluation of fluid-structure coupled systems in complex geometries, e.g. the acoustical wave propagation in the inner ear of mammals, requires elaborate mathematical formulations. One of these is the consideration of the stimulating boundary condition by an acoustic radiator (JASA, Bradley C.E. 1996), which leads to linear behaviour in Lagrangian but is nonlinear in an Eulerian description. In coupled systems the interaction between the structure guided Rayleigh wave and the fluid causes an internal acoustic streaming (AS). AS effects are a result of the nonlinearity of the Navier-Stokes equations in combination with the viscosity. For a numerical evaluation different discretizations are used. A complete Finite Element (FEM) approach is computationally too extensive at least when three-dimensional formulations are used. The fast BEM-FEM mortar coupling for acoustic-structure interaction was recently formulated but the acoustic field is governed by the Helmholtz equation and therefore viscosity was not considered. Another approach which is mainly used for the simulation of AS effects in compressible, viscous gas filled two-dimensional enclosures is the Flux-Corrected Transport (FCT) algorithm to solve nonlinear, time-dependent continuity equations applying a Finite Volume Method (FVM). Special care has to be taken with the formulation of the density boundary condition in this case.

Bös, J.

• Active control of the sound transmission through a double-glazing window
  J. Bös, S. Herold, O. Heuss, M. Kauba and D. Mayer [C]
[A]

Within the framework of the European Integrated Project InMAR (Intelligent Materials for Active noise Reduction) a system for the active control of the sound transmission through a double-glazing window was developed. The window has a size of 1.2 m x 0.6 m and consists of two glass panes (6 mm and 4 mm thick) with a gas-filled gap of 20 mm in between. The vibrational behavior of the window panes was analyzed by means of finite element simulations and experimental modal analyses using a scanning laser vibrometer. The results of these analyses were used to determine suitable locations for the application of piezoelectric patch transducers. If these patches are used as sensors they can serve as error sensors or reference sensors for an active control system, if they are used as actuators they can suppress the vibrations of the window panes and, therefore, reduce the sound transmission through the window. Several transducer configurations and control strategies were tested and compared with each other. The sound radiation of a broad-band shaker excitation was reduced by up to 20 dB at various frequencies, and the sound transmission through the window was reduced by up to 12 dB at various frequencies.

• LOEWE-Zentrum AdRIA – An important step towards the commercialization of adaptive systems
  J. Bös, T. Bein and H. Hanselka [C]
[A]

The LOEWE-Zentrum AdRIA (Adaptronik - Research, Innovation, Application) is a large interdisciplinary research project funded by the government of the German federal state Hessen with a duration of 3 years and a budget of 17.7 million euros for the time being. A second phase of three years with a tentative budget of 20 million euros is planned. It involves the Fraunhofer Institute for Structural Durability and System Reliability LBF, 22 institutes from 3 different departments of the Technische Universität Darmstadt as well as one department of the University of Applied Sciences, all located in Darmstadt, Germany. The goal of this project is the creation and sustainable implementation of an internationally leading adaptronics research institution in Darmstadt. Eventually, it will lead to the foundation of a new Fraunhofer Institute for Adaptronics and of four new university institutes. The project is organized in ten technology areas (e.g., materials, sensors and actuators, systems and control, embedded systems, simulation tools, reliability), which supply the necessary technologies for three application scenarios, i.e., adaptive car, quiet office, and adaptive tuned vibration absorber. An important step towards the commercialization of adaptive systems will be the close cooperation with companies and suppliers during the second phase of the project.

Bokemeyer, M.

• Estimation of the Sound Pressure at the Ear Drum for Hearing Aid Applications
  T. Sankowsky, M. Blau, P. Roeske, H. Mojallal, J. Schroeder, M. Teschner and M. Bokemeyer [C]
[A]

The sound pressure at the ear drum is a reference quantity not only in hearing aids but in virtually all applications involving sound delivery to the ear, such as mobile phones, ear- and headphones. As the direct measurement of the ear drum pressure is too elaborate, it must be predicted in some way. One approach to do so is to establish a 1D model of the residual ear canal, based on the measured acoustic impedance at the interface between the ear shell and the residual ear canal. In a study on 10 human temporal bones with closed ear shells, 1D models established this way were used to predict the drum pressure in the 100Hz..10kHz frequency range. The models were all based on the same measured impedances, but several variants of algorithms to derive the ear canal model were employed. The model predictions were then compared to direct measurements of the drum pressure. It turned out that while the robustness of the different algorithms varies, their mean predictions do not differ substantially. The agreement with the direct measurement will be discussed.

• Modeling the sound field in front of the human ear drum for hearing aid applications
  P. Roeske, M. Blau, T. Sankowsky, H. Mojallal, J. Schroeder, M. Teschner and M. Bokemeyer [C]
[A]

The sound pressure at the ear drum is a reference quantity not only in hearing aids but in virtually all applications involving sound delivery to the ear, such as mobile phones, ear- and headphones. As the direct measurement of the ear drum pressure is too elaborate, it must be predicted in some way. Most attempts to do so exclude the influence of the vibrating drum from their analyses, although it is known that the ear drum vibrates in a complicated manner. In this paper, a finite element (FE) model comprising a detailed representation of the vibrating drum while preserving enough flexibility to be applied to a number of dead and live human specimen is presented and validated. The proposed method involves the representation of the vibrating drum by velocity sources whose source quantities are in turn based on laser doppler vibrometry (LDV) data measured on the respective specimen. The sensitivity of this method to calibration errors, to the effect of missing drum velocity data (towards the edges) and to deviations between the direction of measurement and the predominant direction of drum vibration is analyzed via comparison to a complete FE model (comprising the middle ear and the cochlea load).

Bonin, G.

• Loudness perception with headphone presentation compared to loudspeaker presentation in the diffuse field
  S. Goossens, G. Bonin and R. Stumpner [C]
[A]

The same sound pressure level in the ear-canal can lead to different loudness perception for headphone and loudspeaker presentation. This effect, which is known for a long time, was examined once more for different headphone types and signal presentations. During these tests the subjects were sitting in the reverberation chamber and had to adjust a third-octave band noise to the same loudness caused by loudspeaker and headphone reproduction. The headphone reproduction was carried out in a monotic, diotic and dichotic way. Significant differences of sound pressure levels had been measured in the ear-canal for the monotic and diotic case although the same loudness had been perceived. In contrary the level differences disappear largely for a dichotic stimulation. This is valid both for a binaural noise corresponding to the loudspeaker signal with spatial information, which was taken by artificial head recordings in the reverberation room and for a simple incoherent noise without spatial information at all. In the dichotic case equal levels in the ear-canal had been measured for the same loudness.

Booij, N.

• Road Traffic Noise of Expansion Joints – Get a Grip on it
  W.-J. van Vliet, J. Hooghwerff, N. Booij and R. van Loon [C]
[A]

Expansion joints used in road traffic bridges can be a serious source of noise annoyance. This is caused by the impulsiveness of the sound of vehicles passing over the joints and sometimes the presence of low frequencies in the sound. In the Netherlands large-scale application of low noise road surfaces creates a need for low noise expansion joints. Therefore Rijkswaterstaat has created a guideline to deal with noise of expansion joints. The goal is to avoid excessive noise problems by setting noise limits in contracts. The approach in the guideline is based on existing calculation and measurement methods with which there is already considerable experience in the field of low-noise road surfaces. The application of these methods to the noise caused by expansion joints is new.

Boone, M.M.

• Optimized microphone arrays for the hearing glasses
  M.M. Boone [I]
[A]

In many cases normal hearing aids do not give sufficient speech intelligibility. This is mainly caused by bad speech to noise ratio's such as during parties and under reverberant conditions. To overcome these problems the best way is to use highly directive microphones. We developed a solution by designing a special hearing aid in which the arms of a pair of spectacles are used to house microphones arrays. Optimized beamforming is used to obtain a high directivity index over the speech range of frequencies. Normally this method is applied separately for both ears. In this paper we will give an overview of the theoretical aspects of the optimized beamforming and results are presented where theoretical directivity patterns of the hearing glasses are compared with physical measurements under anechoic and reverberant conditions. The influence of the human head has also been investigated. Our results show that a high average directivity index of 8 dB is reached in combination with a low noise sensitivity. This result is obtained under undisturbed free field conditions. The influence of the human head reduces the directivity index to 7 dB, which is still very high as compared to normal directive hearing aids.

• On the application of auditory scene analysis in hearing aids
  A. Schlesinger and M.M. Boone [I]
[A]

A big challenge in hearing aid (HA) design is the improvement of speech intelligibility (SI) in adverse acoustical situations. Improving SI in noise is a main demand for the majority of hearing impaired people who are afflicted by a sensorineural hearing loss. Approaches to compensate for this dysfunction do not pose a general solution to the problem. There are encouraging approaches that rather replace auditory functions, in particular the function of spatial filtering. Today's HA's are thus often equipped with directional microphones or microphone arrays, which successfully enhance the signal-to-noise ratio (SNR). As the hearing apparatus is a superior speech processor, a further developmental leap in HA design is expected from the emulation of models of auditory scene analysis (ASA). Common to ASA-models is the categorization of sound sources in a feature-space. Using this amended representation, the target-speaker and interfering sound sources are separably displayed. In analogy with the psychological attention to one perceptual stream, it is possible to facilitate these ASA-models to enhance a given target. Several of these functional ASA-models showed a considerable improvement of SNR in a broad range of acoustical situations. In this contribution we discuss representative functional ASA-models and sketch their implementation in HA's.

Boonen, R.

• Iterative learning control for an active exhaust noise attenuation valve for internal combustion engines
  R. Boonen, G. Pinte, W. Desmet and P. Sas [I]
[A]

An active silencer has been developed to attenuate combustion engine exhaust noise. The silencer consists of an electrically controlled valve connected to a buffer volume. The pulsating flow from the engine is buffered in the volume. The valve connected to it is controlled such that only the mean flow passes to the atmosphere. This flow is free from fluctuations and consequently free of sound. An iterative learning controller has been developed to control the active valve. In the control algorithm, non-causal filters are implemented. These filters allow to control systems with high dynamics over a broad frequency range. The active silencer has been experimentally validated on a cold engine simulator, which generates realistic exhaust noise and gas flow using compressed air. The exhaust noise has been reduced in a frequency range starting from 5 Hz until 800 Hz. Depending on the rotational speed of the engine, the typical reductions ranges from 13 dB to 24 dB.

Bos, L.

• An interactive and real-time based auralization system for room acoustics, implementing directional impulse responses and multiple audio reproduction modules for spatialization (the AURALIAS project)
  L. Bos and J.-J. Embrechts [I]
[A]

The AURALIAS project aims at developing an interactive and real-time auralization system for room acoustics. The user is immersed in a virtual sound field, looking at the virtual room and interacting with it through an appropriate interface. The real-time auralization system first imports pre-computed Directional Room Impulse Responses (DRIRs) according to the current situation in the virtual world. Those DRIRs, previously obtained by a ray tracing software, are then convolved with the anechoic signal using a frequency block segmented convolution. Filtered signals are then distributed either to a VBAP or an HRTF sound field reproduction module. The first prototype allows the auralization of multiple isotropic sound sources and the displacement of the virtual listener including head's orientation. Some other features like changing ''on the fly'' the anechoic signals and the current room are also implemented. Further developments will offer the support for moving sources, mobile sweet spot, user's own speech auralization and both real-time acoustical and geometrical modifications of the virtual room. Besides, the software's architecture should be compatible with several room acoustics algorithms, convolution modules or sound field reproduction systems.

Bosschers, J.

• Model Scale Measurements of Surface Ship Radiated Flow Noise
  C. de Jong, J. Bosschers and H. Hasenpflug [C]
[A]

Advances in weapon and sensor capabilities are driving an increased interest in the control of underwater signatures of naval platforms. The control of machinery and propeller noise is well understood but there is a shortfall of knowledge of the mechanisms that govern noise due to the flow around the hull of a surface ship. The subject has been investigated by the collaborative US-NL project Mechanisms and Prediction of Surface Ship Radiated Flow Noise. The project aimed to determine and quantify the sources of flow noise generated by surface ships in terms of ship speed and hull shape. The focus was on the underwater noise generated by turbulence excited hull plating vibration and noise resulting from breaking bow waves. A combination of full scale, large scale, model scale tests and computational fluid dynamics was used. The paper will discuss the test set-up and results of the model scale experiments which were made in a (highly reverberant) towing tank. A special silent towing carriage was developed to reduce the background noise, while an acoustic antenna was used to measure the noise due to the breaking bow waves. The pressure fluctuations due to the turbulent boundary layer were measured at several hull locations.

• Investigation of the resonance frequency of a cavitating vortex
  J. Bosschers [C]
[A]

The cavitation pattern of a modern ship propeller is often characterized by a cavitating vortex. This cavitating vortex may generate (low frequency) hull pressure fluctuations which cause noise and vibration hindrance on board the ship. Existing measurement data of the pressure fluctuations suggest the presence of a resonance frequency, but detailed investigation is hindered by the highly intermittent character of the cavity. A distinct resonance frequency has been observed in the sound generated by a cavitating vortex trailing from a wing of elliptical planform in a cavitation tunnel by Maines & Arndt [JFE199-1997]. The paper will present a theoretical formulation for such a resonance frequency. The agreement between theory and experiment is good for low cavitation numbers but becomes less with increasing cavitation number. The theorically derived dispersion relation is based on the convected Helmholtz equation for a disturbance velocity potential and assumes an inviscid cavitating vortex which shows small deformations for prescribed oscillation modes. The dispersion relation is investigated with respect to phase speed and group speed. The first mode, a breathing mode, may have a zero group speed which is used to derive a unique relation between the non-dimensional (resonance) frequency and the cavitation number.

Bost, W.

• Contrast Enhanced Optoacoustics for Molecular Imaging
  M. Fournelle, W. Bost, S. Wagner, F. Stracke and R. Lemor [I]
[A]

Optoacoustic imaging can provide valuable information about tissue structures by combining the high intrinsic contrast of optical modalities with the good resolution and the high penetration depth of ultrasound. When biological tissue is irradiated with ultrashort laser pulses of durations of a few nanoseconds, the light is absorbed according to the local optical properties of the tissue, and is converted into pressure by means of the thermoelastic effect. The resulting pressure waves can be acquired and analyzed with techniques similar to conventional ultrasound and provide an image of the distribution of absorbed light in tissue. Although endogenous chromophores such as melanin or haemoglobin result in remarkably high optoacoustic contrast, the administration of additional targeted nanoparticular contrast agents becomes necessary if a more selective molecular imaging is needed. Different types of gold nanoparticles are particularly suitable in this context since their strong plasmon resonance results in a high absorption coefficient. An overview on results regarding the usability of the different particle types as optoacoustic contrast agents will be given and the technology and its applications will be discussed. Further, first results on contrast-enhanced in-vivo imaging with a system for real-time optoacoustics will be presented.

Bothe, H.-H.

• Precise stimulation in auditory neuroimplants
  S.W. Kristiansen and H.-H. Bothe [C]
[A]

The work concerns auditory neural implants with specific application to cochlear implants. A usual cochlear implant consists of an external and internal part, with the first being the acoustic receiver, processor and transcutaneous transmitter to the internal part consisting of receiver electronics and an electrode array inserted into the inner ear or cochlea. Electrical stimulation of the auditory nerve fiber array is then conducted by pulsed stimulation with a relatively high frequency through the tissue of the inner hair cells. Thereby, some problems arise with the most obvious ones being a signal distortion by the reduction in available frequency channels and relatively strong channel aliasing or crosstalk effects due to electrode proximity in the electrically conducting tissue.

The aim of this work is to reduce the channel crosstalk by employing a new electrode stimulation method known from high frequency engineering as Phased Array Channeling.

Botteldooren, D.

• Informational masking and attention focussing on environmental sound
  D. Botteldooren and B. de Coensel [I]
[A]

Appreciation of a soundscape depends on the cognitive and emotional evaluation of the sonic environment with particular emphasis on the meaning associated to the perceived sound. Before meaning can be attributed, a sound has to be detected and recognised. Energetic masking due to the working of the inner ear has been mentioned as a reason for not hearing an environmental sound. The central nervous system however also plays role. Informational masking may prohibit recognising or understanding the sound. In this context attention focussing is always around the corner. As attention gets focussed on an auditory stream, informational masking could be reduced even if the signal to noise ratio drops considerably after initial detection. It is for this reason that we include attention shifting in our computational model for environmental noise perception. The model starts from a separation of auditory streams based on spectro-temporal characteristics. A robust mechanism using activation and inhibition models the ability of auditory streams to attract attention. Top down attention is related to activities. The outcome of the model includes the time that particular environmental sounds are heard, their level, saliency, ... This could allow to validate designs based on masking unwanted components of the sonic environment.

Brackhane, F.

• Perception tests with a replica of von Kempelen's 'speaking machine'
  F. Brackhane [C]
[A]

We present experiments with a replica of the historic speaking machine constructed by von Kempelen around 1780.

Contemporaries reported that the machine sounded very realistic, like the voice of a child aged three to six. Despite the mentioned authenticity of the auditory impression it is impossible to generate most speech sounds in a human-like manner and quality. Notable exceptions are bilabial stops, bilabial nasals as well as open vowels which allow words like [mama] and [papa].

In a perception test 22 subjects verbally described various sounds (including giggle, breathing, scratching etc.). The results show that the words [mama] and [papa] produced by our replica were recognized as the voice of a human child by 17 subjects.

In a further speech quality test we plan to explore possibilities to vary different prosodic features (word stress and intonation patterns) by modifying air pressure and segmental timing.

In addition, we report on several experiments generating different voice qualities by using different materials (ivory, brass, reed) and different measures of breadth and length for the reed pipe simulating the vocal chords.

This research helps to explain the fascinating possibilities of the probably oldest working speech synthesis to be presented with a live demonstration.

Braden, A.

• Reconstructing the Lituus: A Reassessment of Impedance, Harmonicity, and Playability
  A. Apostoli, S. Logie, A. Myers, J. Kemp, J. Chick and A. Braden [C]
[A]

Bach's motet `O Jesu Christ, meins Lebens Licht' (BWV no. 118), first performed in 1736-7, calls for a Lituus, believed to be a long trumpet-like instrument, with playing characteristics somewhere between a trumpet and an alphorn. Unfortunately, there are no surviving examples of Bach's instrument. Using software recently developed at the University of Edinburgh for brass instrument optimisation, a new bore profile with theoretically calculated input impedance curves has been proposed and submitted to a collaborating instrument maker with the aim of providing a natural instrument, with a distinct timbre, that responds well to the player and is in tune over a wide range of natural notes.

This paper describes the process, computational modelling, and experimental work undertaken in an attempt to ensure a satisfactory final instrument design. An initial validation of this process was achieved, in part, by investigating the playing properties of an instrument with a notionally similar bore profile, a 2.4m long Fiscorn in C (a type of Catalonian flugelhorn). This work compared the measured bore profile, input impedance, playability and intonation with those derived theoretically.

Brand, T.

• Crosstalk Cancellation in audiology
  A. Winkler, T. Brand and B. Kollmeier [C]
[A]

Sound reproduction via loudspeakers is influenced by the crosstalk that occurs between loudspeakers and listener. Crosstalk cancellation can be used to reproduce the acoustical room impression of a recording room in a different playback room using two loudspeakers. In order to achieve this, the contralateral pathways need to be cancelled and the ipsilateral pathways need to be equalized. In this study different algorithms for crosstalk cancellation were evaluated by assessing the localization performance in the horizontal plane and by speech intelligibility tests in spatial noise with 4 normal-hearing listeners. The experiments took place in two different playback rooms (acoustically damped and with reflections). The results showed that crosstalk cancellation yielded an improvement up to 9dB of speech intelligibility in noise when the speech was spatial separated from the noise even in reverberant rooms. Using crosstalk cancellation, listeners were able to localize stimuli outside the triangle between them and the loudspeakers (20) which was not possible without using crosstalk cancellation. The use of crosstalk cancellation makes it possible to perform speech intelligibility tests and localization measurements in reflecting rooms in a similar way as in anechoic rooms or like using head phones.

• Modelling binaural speech recognition
  T. Brand, R. Beutelmann and B. Kollmeier [I]
[A]

In many cases speech has to be recognized in adverse listening conditions including noise sources from different directions and adverse room acoustics. Furthermore, hearing-impairment compromises intelligibility in such conditions. Binaural processing of the auditory system can cause a substantial improvement of speech recognition. A binaural speech intelligibility model (Beutelmann and Brand, JASA, 2006) yields accurate predictions of speech reception thresholds (SRTs) in presence of a stationary noise source at arbitrary azimuths and in different rooms. The model combines a multi-frequency channel equalization-cancellation (EC) process and the speech intelligibility index (SII). The effect of hearing loss is simulated by adding uncorrelated masking noises (according to the pure-tone audiogram) to the ear channels. A recent revision of this model is based on an analytical expression of binaural unmasking for arbitrary input signals and is computationally relatively efficient. An extension for non-stationary interferers was realized by applying the model to short-time frames of the input signals and averaging over the predicted SRT results. Model predictions are compared to binaural speech intelligibility data from normal-hearing and hearing-impaired listeners, incorporating different combinations of rooms, sound source setups and noise types.

• Consonant Recognition of Listeners with Hearing Impairment and Comparison to Predictions Using an Auditory Model
  T. Jürgens, T. Brand and B. Kollmeier [C]
[A]

Contrary to listeners with normal hearing, persons with high-frequency sensorineural hearing loss suffer especially from a poor consonant recognition. This is mainly attributed to the loss of audibility within high-frequency regions. But still it is not clear how suprathreshold factors like the loss of dynamic range and thus the modification of dynamic compression in the impaired auditory system contribute to the poor consonant recognition. In a former study we showed that consonant recognition for normal-hearing listeners can be predicted well using an auditory model combined to a speech recognizer. In this study we address the following question by modifying the auditory model to model the signal processing of listeners with hearing-impairment: Does a detailed modelling of hearing-impairment by including suprathreshold factors like a modified dynamic compression enhance the prediction of consonant recognition compared to a model using audibility only? Therefore, observed consonant recognition rates are compared to model predictions using different model approaches to involve hearing-impairment.

• Prediction of Speech Intelligibility in Fluctuating Noise for Listeners with Normal and Impaired Hearing
  R. Meyer, T. Brand and B. Kollmeier [C]
[A]

Speech intelligibility prediction based exclusively on the pure-tone audiogram (e.g. SII-based models) shows some limitations. For listeners with similar pure-tone audiograms the same speech reception thresholds (SRT) are predicted even though the observed SRTs may differ. For SRTs measured in fluctuating noises, these differences are even larger than in stationary noises. One hypothesis is that cognitive parameters significantly influence the observed SRTs in fluctuating noise. In this study the speech reception thresholds for stationary and different fluctuating noises are measured for 12 listeners (4 with normal hearing and 8 with impaired hearing). For the listeners with impaired hearing three SRTs for different noises are measured using the master hearing aid (a computer program simulating an hearing aid), to evaluate the capabilities of the prediction models in predicting aided thresholds. Additionally, a pure-tone audiogram and a categorial loudness scaling are measured. For the measurement of cognitive parameters, the (visual) text reception threshold test and the lexical decision test are used. The results of the cognitive tests are used to explain parts of the remaining deviations between the audiogram-based predictions and the measured values. Preliminary results show significant correlations between the deviations and some of the cognitive tests.

Brandao, E.

• Large scale in situ acoustic reflection measurements in a theatre
  E. Tijs, H.-E. de Bree and E. Brandao [C]
[A]

The acoustic absorption properties of materials used in a theatre are a major concern, both in the initial design stage and during renovation. The PU surface impedance method allows the simultaneous measurement of both sound pressure and acoustic particle velocity, providing the opportunity to measure materials as they are installed on seats, floors, walls and ceilings, assessing their sensitivity to both normal and oblique angles of sound wave incidence. Using a portable piece of equipment, actual testing time is short, allowing large scale testing for comparisons. The method was applied at the Musis Sacrum theatre in Arnhem, the Netherlands. Results will be presented and findings will be discussed.

Brandstetter, K.-D.

• Measurement of sound insulation in laboratory - comparison of different methods
  L. Weber, H. Schreier and K.-D. Brandstetter [C]
[A]

For measuring the sound insulation of building elements in laboratory three standardized methods based on different acoustic principles are available. According to ISO 140-3 the element under test is installed between two reverberant rooms. The measurements acc. to ISO 140-5 combine free field excitation with a reverberant receiving room and ISO 15186-1 uses sound intensity for determination of transmission loss. All methods assume idealized acoustic conditions that differ more or less from the situation found in practice. In addition to the inherent statistic variation of the measuring results this gives rise to systematic deviations between the different measuring methods.

Since the deviations are so far only partly understood, they were investigated by means of an extensive series of tests. The tests comprised three different types of building elements (a chipboard, a lightweight double-leaf construction and a membrane partition) and were performed under well-defined conditions in the test facilities of the Fraunhofer-Institute for Building Physics. Apart from comparing the different measuring methods additional investigations on the influence of the most important measuring parameters (reverberation time, number and position of microphones, etc.) are presented. The results of the investigations contribute to a better understanding of accuracy and reliability in measuring sound insulation.

Breebaart, J.

• Parametric Binaural Synthesis: Background, Applications and Standardization
  J. Breebaart, F. Nater and A. Kohlrausch [I]
[A]

The synthesis process of binaural sounds has been subject to research for several decades. The employment of head-related transfer functions (HRTFs) has greatly progressed both binaural research as well as binaural applications. At the same time, the amount of information present in HRTFs and the required processing capabilities for real-time binaural rendering have long been a challenge for many applications in the consumer domain. More recently, parametric methods to capture the perceptually-relevant information from HRTFs have been developed. By means of extracting perceptually-relevant attributes from HRTF pairs, binaural rendering can be performed at lower complexity compared to the employment of HRTF convolution. Moreover, parameter-based binaural rendering can be efficiently integrated with parametric audio coders, it can provide a more convincing spatial image when converting conventional stereo to binaural signals, and can be used to provide means to independently adjust the spatial position of individual sound sources present in a down mix. This paper provides an overview of the perceptual consequences and limitations of HRTF parameterization, its applications, and relevant standardization efforts.

• Tone in noise detection: a review of spectral integration
  N. Le Goff, A. Kohlrausch and J. Breebaart [I]
[A]

An experiment is conducted to measure binaural (NoS ) and monaural (NoSo) tone in noise detection thresholds for masker bandwidths that varied from 10 to 1000 Hz while the tone to detect had a frequency of 500 Hz. Thresholds are expressed as signal to masker spectrum level ratio (S/No). For bandwidths less than 1 ERB, thresholds in all measured conditions increase with increasing masker bandwidth. For bandwidths wider than 100 Hz three different behaviors are observed. Thresholds for NoSo conditions with a running noise masker remain essentially constant which is in line with the critical band paradigm. In contrast, thresholds for NoSo conditions with frozen noise masker and NoS conditions both show a slight increase with increasing masker bandwidth beyond 100 Hz. A third effect of bandwidth is observed in our data for NoS conditions with an overall ILD of 30 dB. In this case, thresholds show a decrease of about 5 dB when the masker bandwidth is increased from 100 to 1000 Hz. This unusual bandwidth dependence is also found for binaural detection with a reduced masker correlation and for monaural detection of a 6-kHz tone in noise. An overview of the spectral integration for these various conditions is given and potential explanations are discussed.

Breuer, F.

• Sound Insulation for practice rooms used by rock bands
  F. Breuer and J. Paprotny [C]
[A]

During our work as acoustic consultant we had to judge the sound transmission of a cluster of more than 20 practice rooms in the neighbourhood. One of the basic question was to determine ''typical'' sound levels in practice rooms. To answer this question measurements were made in other practice rooms with different kinds of bands and music. In addition to handmade music we analyzed recorded pop-music with very low-frequent parts. During the measuring at the location the sound insulation of all 20 practice rooms to the nearest flat was measured and calculated. Based on these data, in a further step, the sound pressure level in the flat could be calculated as if the bands actually would play in these rooms, or the pop music would be played there. The results of the 20 calculated sound insulations and the corresponding sound levels in the flat are presented. These results are compared to german standards and an unexpected rating of the different kinds of music is detected.

Brick, H.

• Point-source-scattering from tyre-like structures above an impedance plane
  H. Brick, M. Ochmann and W. Kropp [I]
[A]

In preceding publications (DAGA'08, Acoustics'08), we discussed the possibility to incorporate an infinite impedance plane into the direct Boundary-Element-Method (BEM) by means of an appropriate Green's function. This approach does not require a discretization of the impedance plane, since the Green's function fully takes into account the effects of reflection and absorption of the sound field caused by the impedance plane. The implementation was verified by analytical test cases for radiation problems. Now we focus on the scattering problem in presence of an impedance plane. The incident sound field is included in the Helmholtz-Integral-Equation by an additional summand, which is also influenced by the boundary conditions of the half space problem. This extension of the BEM-half space approach makes it possible to investigate the influence of the ground impedance on the well known horn effect of a tyre/road interface by means of BEM-simulations. The horn effect, an amplification mechanism of the horn-like geometry of the tyre/road contact area, was experimentally investigated in previous papers. These measurements serve as validation data for the BEM solutions.

Brix, S.

• Spherical Array Systems - On the Effect of Measurement Errors in Terms of Perceived Auralization Quality
  F. Melchior, Z. Kuang, D. de Vries and S. Brix [I]
[A]

The analysis of acoustic fields using an open cardioid spherical microphone array enables the extrapolation of the measured field to arbitrary positions inside the reconstruction area of the array. An application of such a system is the generation of virtual stereo main microphone setups, which can be used for auralization purposes. Practical limitations of a array system will lead to several types of errors. In this work the effects of measurements errors on the perceptual characteristics of diffuse fields is investigated. In a listening experiment two channel ideal microphone setups are compared with corresponding virtual sensor setups based on spherical microphone array measurements.

• Electro Active Materials for Flat Panel Transducers
  L. Ehrig, D. Beer and S. Brix [C]
[A]

Electroactive materials are a class of materials that change their geometry due to the presence of an electrical field. A well known effect that occurs in crystals, polarized ceramics and plastic materials is piezoelectricity. Since there is no center of symmetry and a preferred direction in such materials respectively an electrical voltage applied on the materials surface yields a deformation. The same macroscopical behaviour can be observed in relatively new materials called ferroelectrets. Voids in these cellular polymer foams are internally charged and form giant dipoles. The dipoles are deformed by an electrical field likewise in piezoelectric materials. Electroacoustic transducers based on electroactive materials may have a unique feature: driver and membrane are one part. Thus it is possible to build extremely simple, compact and flat transducers. Using the material by means of a thickness oscillator even the rearward air spring, causing a major part of problems for thin transducer constructions, can be avoided. In this work several approaches for transducers using electroactive materials are presented.

Brockt, G.

• Safe and Sound - Exposure Control for Workers in Music and Entertainment
  G. Brockt [I]
[A]

The EU-Directive 2003/10/EC on occupational noise refers to all workers expressly including those in music and entertainment industry. In this sector sound is intended and indispensable but may be harmful at the same time. Nevertheless, the fundamental principles of noise control e.g. the general obligation for noise reduction at source or the priority of collective protection measures over individual protection measures are implemented in the directive. Where noise exposure exceeds action levels further measures have to be applied: use of hearing protection, health surveillance, marking of noisy work places and the implementation of noise reduction programmes. Moreover a limit of 87 dB(A) for the noise exposure level, taking into account the attenuation of hearing protection, shall be complied. The general way of noise control is noise reduction at source, on the transmission path, by organizational measures and the application of hearing protection. But this procedure appears to be a challenge in music and entertainment. This contribution covers the exposure in different sectors of the entertainment industry like orchestras, clubs or discotheques. Principles and examples for exposure limitation are described. Obviously, there exists no general solution, but only the combination of several individually chosen measures can yield a practicable exposure control.

Bron- van der Jagt, S.

• Sound and vibration transmission through lightweight junctions: Numerical investigation and validation by laboratory tests
  S. Lentzen, E. Gerretsen, C. van Bentum and S. Bron- van der Jagt [C]
[A]

In this work the sound and vibration characteristics of several lightweight junctions are numerically investigated with FEM (for the low- and the mid-frequency range) and SEA (for the mid- and the high-frequency range). In order to validate these analyses, the junctions are tested in a laboratory setting.
For sound insulation the investigations are focused on obtaining the vibration reduction indices . These are not only investigated for the floor-floor flanking transmission path, but also for all the other possible paths (e.g. floor-wall, floor-ceiling and wall-ceiling). Although applied in different frequency ranges, the results of the FEM and the SEA analyses agree fairly well with those obtained in the laboratory tests.
Looking at the vibration behaviour and transmission, the main focus was put on obtaining the One-Step-RMS velocities of the floors. In these investigations the point- and transfermobilities are acquired either by numerical analyses or by measurements. Since these spectra are only required for frequencies up to 40Hz, the numerical study consists merely of FEM analysis. Here, both results agree fairly well.
From the investigations in both areas, the appropriate conclusions concerning good building practices in lightweight buildings concerning sound and vibration can be drawn.

Bronkhorst, A.

• Auditory capture in an auditory and visual spatial cueing task
  T. Koelewijn, M. van der Hoeven and A. Bronkhorst [I]
[A]

Previous cueing studies show that a sound coming from a particular location in space can capture auditory and visual attention. We conducted two cueing studies where a baseline condition in the form of an auditory spatially diffuse cue was included, using out-of-phase presentation through two loudspeakers. The first study examined how an auditory non-informative spatial cue influences detection and localization of auditory targets. Results show that, compared to the baseline, cues shortened reaction times at small cue-target angles (up to 7) and increased them at larger angles. The reaction time difference was up to 80 ms. Furthermore, cueing seems to have no effect on sound localization: false alarms were normally distributed around the hits. These findings demonstrate that strong auditory capture effects occur that depend on cue-target distance. The second study shows how an auditory non-informative spatial cue influences detection of visual targets. Here the results show performance benefits when the target was preceded by a cue presented at target location and performance costs when a cue was presented elsewhere. Additionally, this study shows that even when attention is highly visually focused by means of a 100% valid endogenous visual cue, auditory stimuli still capture visual attention.

Bronsvoort, A.

• Contribution of stiffeners to the sound transmission of aircraft skin panels
  J. Wind, A. Bronsvoort and A. de Boer [I]
[A]

Traditionally, the sound radiation of aircraft skin panels is calculated using a structural model with stiffeners and an acoustical model where the stiffeners are omitted. We present a numerical study on the impact of this omission.

The numerical model consists of three parts. Firstly, the pressures in a diffuse sound field are modeled. Secondly, a Finite Element (FE) model calculates the structural response caused by the sound field. Thirdly, a Boundary Element (BE) model calculates the sound intensity. The impact of stiffeners is determined by comparing the results of a BE model with and without stiffeners.

The emphasis lies on the results of this model. Several stiffener configurations and frequencies are compared and a number of new analysis techniques are introduced to pinpoint the cause of these differences. Differences of upto 50% have been found for the geometries that have been tested.

Brücker, C.

• Human phonation analysis by 3d aero-acoustic computation
  S. Zörner, M. Kaltenbacher, W. Mattheus and C. Brücker [I]
[A]

A numerical model of the phonation process, which takes place inside the larynx, is presented. The air flow through the glottis has been replaced by water and the physical dimensions have been scaled accordingly. Therewith, the characteristic frequencies and velocities are reduced. The movement of the vocal folds are prescribed by geometric changes of the computational domain.

In utilising numerical simulations, the goal is to study the complex interaction of the modulated flow, due to the prescribed geometric variations. The fluid flow is simulated by solving the incompressible Navier- Stokes equation with the help of a direct numerical simulation (DNS) based on the software OpenFOAM. The results are scaled to comply with a simulation in air. This data is then taken to calculate the acoustic sources inside the model by using Lighthill's analogy and the acoustic sound propagation by solving the inhomogeneous wave equation. Currently we prescribe a periodical flow on the inflow boundary and compute the flow field for fixed vocal fold geometries. The computed acoustic field inside the glottis shows the same dominant frequency as the prescribed fluid inflow.

• Three-Dimensional Unsteady Flow Nature in the Vocal Tract during Human Phonation
  M. Triep, W. Mattheus, M. Stingl and C. Brücker [I]
[A]

The quality of the human voice during phonation is strongly related to the three-dimensional (3D) unsteady flow field in the vocal tract downstream of the vocal folds. A slight pathological change of the vocal fold's geometry and/or their movement may have already a strong effect on the flow field and thereby on the flow-induced noise (sources) which is a parameter that characterizes the voice quality. In this study the 3D unsteady flow field in a dynamical model of the vocal folds is simulated numerically and visualized with the same boundary conditions in the experiment. In addition to the 3:1 up-scaling of the model water is used as the working medium so that the time scale is best accessible in the experimental approach for the measurement of the flow field structures by time-resolved Particle-Image Velocimetry and of the time-dependent volume flow rate. A realistic pre-phonatory transglottal pressure difference and time-varying motion and profile of the glottis during the phonatory cycle are prescribed in our model, and the resulting characteristic volume waveform is used as input for the numerical simulation with the flow solver code OpenFOAM. In particular the comparison between the flow dynamics with and without model ventricular folds, which build a second constriction in the test channel, is analyzed.

Bryan, A.G.

• The Aubrapan: Revival of a lost invention
  S. Maloney, C.A.C. Imbert and A.G. Bryan [C]
[A]

The Aubrapan is a new design of the melody (or soprano) pan developed in 1978 by Guyanese pan maker and panist Aubrey Bryan in the United Kingdom. In the Aubrapan the higher frequency notes are placed nearest the rim of the pan in contrast to the center of the pan where most high notes are generally located. The lower octaves are placed directly opposite each other in order to facilitate the pendulum-like action of the arms - these ascend in whole tones. This concept facilitates easier and rapid playing of the chromatic scale in single and double note form. In this paper we describe the development of the Aubrapan and an attempt is made to discover why the Aubrapan has not yet found a place among the steelpan family. To answer this and other questions, the experiences of the Aubrapan's inventor are noted and a comparison of the Aubrapan's playability is made with that of the conventional Trinidad and Tobago 4th's and 5th's soprano pan.

Buchholz, J.M.

• A loudspeaker-based room auralisation system for auditory perception research
  J.M. Buchholz and S. Favrot [I]
[A]

Most research on basic auditory function has been conducted in anechoic or almost anechoic environments. The knowledge derived from these experiments cannot directly be transferred to reverberant environments. In order to investigate the auditory signal processing of reverberant sounds, a loudspeaker-based room auralisation (LoRA) system is proposed here. The LoRA system efficiently combines modern room acoustic modelling techniques with higher-order Ambisonic auralisation. Thereby, aspects of the auditory precedence effect are utilized to realise highly authentic room reverberation. This system provides a flexible research platform for conducting auditory experiments with normal-hearing, hearing-impaired, and aided hearing-impaired listeners in a fully controlled and realistic environment. This includes measures of basic auditory function (e.g., signal detection, distance perception) and measures of speech intelligibility. A battery of objective tests (e.g., reverberation time, clarity, interaural correlation coefficient) and subjective tests (e.g., speech reception thresholds) is presented that demonstrates the applicability of the LoRA system.

Buchholz, U.

• Investigation of the Sound Transmission of a Partial Discharge through a Solid Dielectric Multilayered Device
  U. Buchholz and B.A.T. Petersson [C]
[A]

Acoustic emission (AE) sources caused by partial discharges in solid dielectric devices like transition joints or outdoor terminations are investigated. Together with a prepared high voltage cable end this device forms a multilayer structure. -The hypothesis is that due to the high damping of polymeric insulation materials no AE signal can be detected on the surface of the prescribed structures. -An analytical technique is used to determine the transmission for a typical multilayer insulation. For quantitative computations it is necessary to know the material properties such as complex Young's modulus, density and the Poisson's ratio. Since the properties found in the literature were not fully satisfactory, measurements had to be done. The experimental setup and the results are shortly presented. -Finally it will be shown that the transmission path plays not the significant role as was assumed in the beginning.

Buchschmid, M.

• A semi-analytical model for rooms with absorptive boundary conditions
  M. Buchschmid, G. Müller and M. Pospiech [C]
[A]

The prediction of sound-fields in closed volumes with vibrating delimiting surfaces typically is carried out by the use of energy methods. Those methods are robust for subsystems with high modal density. However their performance is limited if the spatial resolution of the response has to be described and if boundary conditions are investigated in detail. In order to gain spatial information about the sound pressure in such a volume, the coupled system of fluid (acoustic volume) and structure (boundary conditions) has to be considered in the frequency range of interest. First the components are calculated decoupled. A modal approach is used for the fluid, where analytical results are available for simple geometries. For an efficient numerical implementation, especially for analyzing more complex geometries, spectral approaches are considered in addition to FEM. The boundary conditions are modeled with the help of impedances. The angle of inclination of the incoming waves is considered by a wavenumber dependent description. Plate resonators and arbitrary plate-like compound absorbers are treated using the Theory of Porous Media for modeling porous foam structures and the Lamé Equation for homogeneous materials. The coupling of fluid and structure is carried out with the help of the component mode synthesis.

Buckert, S.

• ADP: Practical experience in acoustics education
  S. Buckert, A. Skowronek and K. Wolf [C]
[A]

Conveying interdisciplinary competences is an important aspect of the master degree program at the Technische Universität Darmstadt. A so called ''Advanced Design Project'' (ADP) largely contributes to the communication of these skills. The ADP of the Department of System Reliability and Machine Acoustics (SzM) preferably takes place at industrial sites (e.g. manufacturing plants, test and research facilities), tasking real problems. Thus the students get prepared for the challenges of their prospective profession. The problems have to be solved by the students within a given time frame working in teams of three to four people. Successfully accomplishing a project like this requires not only professional knowledge but also a high degree of teamwork, communication and presentation skills as well as a good time management. Besides these soft skills, professional knowledge in the field of measurement techniques, use of data acquisition and analysis software, noise control engineering, silencer design, etc. is attained. This project complements the courses in the field of machine acoustics offered by the department of System Reliability and Machine Acoustics.

• ADP: Baseline analysis of the acoustic emission of induction furnaces
  A. Skowronek and S. Buckert [C]
[A]

An ADP is an industry-oriented course, in which students are to solve a described task as independently as possible, enhancing their ability to work as a team. In collaboration with several industrial partners the Department of System Reliabiltiy and Machine Acoustics SzM of the Technische Universität Darmstadt has accomplished various ADPs in recent years.

As an example for such an ADP, this paper describes the baseline analysis of the acoustic emission of induction furnaces at SMS Elotherm in Remscheid, Germany. Alternating magnetic fields and Lorentz forces cause oscillatory impulses in the induction furnaces as well as in the metal components which are to be heated up, partly creating sound pressure levels far above 100 dB.

After a week-long introduction in the handling of the available measuring instruments, the students were to perform acoustic measurements and analyses on these furnaces. For the most part independently, the students hence measured several air-borne and structure-borne sounds.

The necessary preparation, measurement procedures, and some selected results are presented.

Budde, C.

• Fast distortion measurements in relation to frequency and level
  C. Budde and S. Müller [C]
[A]

Distortion measurements vs. frequency and/or signal level are normally performed with a series of pure tones whose frequency or level is incremented in discrete steps, turning this type of test rather tedious. This work presents two different methods in which one of the two parameters is changed continuously to speed up the overall process and to deliver data for a 3D plot easily interpretable. The pros and cons of the two methods (log sweeps and level sweeps) are discussed and some practical examples shown.

Bütikofer, R.

• Aircraft noise calculations for relevant periods of day using a complete set of radar data
  B. Schäffer, S. Plüss, G. Thomann and R. Bütikofer [C]
[A]

Aircraft noise contours are usually determined with model calculations. As the model results have important consequences for land use planning and payments of compensation, the model calculations need to be highly accurate. The hypothesis of this study was that modelling the daytime-specific changes in flight geometries determined from one year's radar dataset enhances the accuracy of the model results. To date, the model FLULA2 developed at Empa randomly chooses up to 100 flight paths per aircraft type and route in simulations to calculate 'footprints', i.e. daytime-unspecific averaged sound exposure levels scaled to one flight, which are then weighted with operational data to obtain daytime-specific noise contours. In this study, we extended the model to perform separate simulations for relevant daytime periods, using all available flight geometries of one year's flight operation to obtain daytime-specific footprints. Exemplarily for the year 2006, we determine the noise contours around Zurich Airport, Switzerland, with the common ( 27'000 simulated flights) and the extended simulation ( 235'000 flights). We compare the resulting daytime-specific noise contours with monitoring data to determine how much the extended simulation may improve the results and what possible consequences are, such as the changes in the number of persons above threshold values.

Bula, V.

• PIV Measurement of Flow-Patterns in a Human Vocal Tract Model
  J. Horacek, P. Sidlof, V. Uruba, J. Vesely, V. Radolf and V. Bula [I]
[A]

The contribution describes developed complex physical model of the voice production that consists of the trachea, the self-oscillating vocal folds and the vocal tract with acoustical spaces corresponding to the vowel /a:/. The vocal tract model was developed from the 3D FE models designed from MR images The vocal folds were made of a latex thin cover layer filled by a polyurethane rubber, joined to the subglottal spaces. The time-resolved PIV method was used for instantaneous velocity field evaluation. Preliminary results will be presented for measurements performed within a physiologically range of mean air flow rates and fundamental frequencies. The vibrating vocal folds during one oscillation period were recorded by the high-speed camera in the same instants as the airflow velocity patterns. Flow structures resembling large vortices with dimensions comparable with the channel cross-section is possible to see above the ventricular folds. The vortices disappear in the narrowest epilaryngeal part of the vocal tract where the flow is uniform. Large eddies of a size comparable with the channel height were observed in the model of the mouth cavity where the airflow was attached to one wall of the channel resembling a Coanda effect.

Burgschweiger, R.

• Benchmarking for sound transmission and scattering from thin elastic structures using analytical, BE and FE coupling methods
  R. Burgschweiger, I. Schäfer, R. Piscoya and M. Ochmann [I]
[A]

The calculation of the pressure scattered from elastic structures composed of thin elastic materials is one of the main purposes for the detection of underwater objects. For this reason, the sound pressure scattered from spherically structures (''shells'') placed in and filled with fluid will be calculated using different numerical and analytical coupling methods. Also, the sound transmitted through the shell into the interior of the structure will be investigated. We will compare and benchmark analytical solutions based on spherical wave functions with results of an in-house developed BEM-package and commercial BEM/FEM applications. Especially, the minimum limit of the possible ''wall thickness'' depending on frequency and material parameters will be figured out by comparing the fully elastic model with a thin-walled shell.

Burkhart, C.

• DEGA Sound insulation certificate - a concept for more transparency
  C. Burkhart [I]
[A]

The German society for acoustics has published the DEGA recommendation 103 ''sound insulation in the housebuilding - sound insulation identity card''. The two main targets of the DEGA recommendation 103 are:

(1) creating a multi-level system for the differentiated designing and marking of the structural sound insulation between room-situations regardless of the kind of the building, (2) developing a classification system on this basis for the simple marking of the sound insulation performance of whole housing units or buildings.

The introduction of a multi-level requirement- and assessment-system is meaningful and necessary for a clear differentiation and evaluation of the sound-related quality of buildings. The system is co-ordinated with the today usual building methods and with the today's build-up-obviously introduced minimum requirements according to DIN 4109. A differentiated and practical classification is possible by the organization into altogether 7 levels both for new buildings and for the old building existence. Unfortunately, the identity values of the sound insulation are very badly clear for planners and users. With the ''DEGA sound-insulation certificate'' a simple system is given to mark the sound insulation. The user and the consumer can carry out high-class comparisons without deeper professional knowledge and make purchase decisions.

• DEGA Sound Insulation Certificate – Application of DEGA Recommendation with Practical Examples
  F. Schnelle, R. Kurz and C. Burkhart [I]
[A]

According to the current German jurisdiction newly constructed dwellings with standard comfort usually have to meet higher sound insulation requirements than those demanded in the valid building regulations of DIN 4109: 1989-11. Consumer without acoustical experience faces difficulties in comprehending the acoustic requirements for improved sound insulation, e.g. according to Supplement 2 to DIN 4109 or VDI 4100.

The method of classifying dwelling units into sound insulation classes A* - F according to the DEGA sound insulation certificate can provide the consumer with a comprehensible acoustic rating. The criteria list of the DEGA sound insulation certificate can be used either for the planning of new buildings or the acoustic valuation of existing dwelling units in houses. The sound insulation certificate can be issued based on results of estimations and measurements.

We present the practical use of the criteria list by showing examples for dwelling units in different buildings (apartment houses, semi-detached and terrace houses). The effects of the individual criteria on the overall rating of the sound insulation classes A* - F will be explained.

Buss, S.

• Transmission of noise through a pipe as structure-borne sound and fluid sound
  S. Buss, M. Wildemann and F. Hartmann [C]
[A]

Busscher, G.

• Finding the source of Low Frequency Noise by measuring ground born vibrations
  G. Busscher and C. Ostendorf [C]
[A]

The residents of an old farmhouse in the countryside were suffering from nuisance by low frequency noise. Although the residents could hear the noise all the time, the noise was heard better during specific moments in the evening and night. The combination of vibration measurements on a wall in the farmhouse and sound measurements outside the house, did not reveal the direction of the source of low frequency noise. During a second measurement session, the vibration measurements in the farmhouse were performed in combination with ground-born vibration measurements on several positions outside the farmhouse. Based on the phase-shift of the vibrations on the different measurement positions, it was possible to determine the direction of the source of low frequency noise.

Busse, S.

• A method based on the ray structure of acoustic modes for predicting the liner performance in annular ducts with flow
  A. Moreau, S. Guerin and S. Busse [C]
[A]

The objective is to develop a fast method for predicting the sound attenuation by acoustic liners mounted in the inlet and bypass duct of turbofan engines. The method will be integrated into a prediction tool dedicated to design-to-noise. It works as follows: 1) The sound field generated by the fan stage is decomposed into acoustic modes (m,n) for a hard-walled annular duct with flow. 2) The ray structure of each acoustic mode is determined using an extension of an existing theory. 3) Finally the sound attenuation is calculated for each mode in function of the polar and azimuthal angles of propagation (angles at which the rays impinge on the wall) and the number of bounces; it is assumed that liners do not significantly modify the propagation angles and generate no mode scattering. The wall impedance is predicted using a model based on the geometrical parameters of practical liners. The results presented include the impact of acoustic liners on the modal distribution, the sound spectra, and the overall sound power. Both broadband and tonal components of fan noise are considered for a realistic configuration.

• Experimental and Numerical Characterisation of a Non-Locally Reacting Liner
  S. Busse, C. Richter, C. Kückens, U. Müller, L. Enghardt and F. Thiele [C]
[A]

A new type of aero-engine liner was investigated. It is manufactured by combining a cell structure of assembled metal strips with a perforated face sheet and a solid backing plate by using a point welding process. The structure is completed with an additional face sheet layer of meshed metal baffle. Since neighbouring cells are not air tight the damper is non-locally reacting and inter-cellular communication is possible. The acoustic properties in terms of the insertion loss were experimentally analysed by integrating a liner sample in a side wall of a grazing flow acoustic test facility. Measurements with microphones and plane wave excitation were conducted in order to calculate the reflection and transmission coefficients. The influence of parameters like the flow Mach number and the liner length on the damping behaviour were investigated. By using a numerical method which incorporates the Extended Helmholtz Resonator model (EHR) the impedance was educed from the measurement values utilising an optimisation algorithm. It is shown that the method provides reasonable results even for this special type of liner.

Butenweg, C.

• An approach for numerical calculation of glottal flow during glottal closure
  A. Gömmel, M. Kob, T. Niendorf and C. Butenweg [I]
[A]

Limitations of classical low-degree-of-freedom approaches (multiple-mass models) of modeling vocal fold (VF) oscillation have indicated the need for more advanced models. A couple of finite-element (FE) models have been presented in the last years, which include a number of numerically difficult solver algorithms. The presented model represents a coupled two-dimensional model of VF movement including fluid-structure interaction effects. A structural model of vocal fold tissues and a fluid model of the airflow are combined by the Arbitrary Lagrangian-Eulerian formulation. As solvers ANSYS was chosen for the structure and CFX for the fluid. A special problem is handling the stop of the airflow in the closed phase of the glottal cycle. Since the fluid elements in the glottal gap channel would have to be zero-volume elements, a special modeling strategy is necessary. In this model a loss coefficient is introduced when the glottal gap falls below a certain control volume. As exemplary area of use, results of comparative simulations are shown, which differ in driving pressure values and VF shape. It is shown that typical convergent/divergent shapes are only obtained when a minimum length of a channel between the VF remains.

Calderón, E.

• Simulation of differences between male and female vocal fold configuration during phonation
  M. Kob, P. Dejonckere, E. Calderón and S. Kaynar [I]
[A]

Statistically, some vocal fold lesions such as vocal fold nodules - considered as a localized tissue reaction to voice loading - are much more frequent in female ( 95%) than in male ( 5%) adult humans. This can only partially be explained by the about two times higher oscillation frequency during voice production in females. In previous studies, gender-related differences in the shape of the vocal fold edge have been observed that can be related to different geometries of the laryngeal framework. With a time-domain model of vocal fold function the oscillation patterns and resulting forces between and inside the vocal folds have been simulated, and the conditions for eliciting a hourglass-shaped vibration pattern (necessary for determining the localized tissue reaction of the nodules) could be clearly defined. Our simulations and measurements support the assumption that collision and tearing forces on the vocal fold edges during phonation are higher and more specifically located in females compared to males.

Campbell, M.

• Embouchure Control of Brassiness at Constant Pitch and Dynamic Level in Orchestral Horn Playing
  L. Norman, J. Chick, M. Campbell and A. Myers [C]
[A]

It is generally acknowledged that an increase in the dynamic level of a brass instrument brings about a change in timbre, often described as 'brassiness'. This phenomenon can be attributed to the generation of shockwaves at higher dynamic levels. However, the point at which this change occurs is not always clearly defined and some brass players have found that, by employing slight changes in their embouchure, they have a certain degree of control over the 'brassiness' at a constant dynamic level.

This paper focuses on how embouchure control influences wave propagation in the air column of a brass instrument during the onset of brassiness. A number of horn players were asked to play a long note at a loud dynamic level at the threshold of the onset of brassiness, and to move between a non-brassy and brassy timbre keeping dynamic level and pitch constant. Different horns were used to explore whether small changes in bore profile affected the ease of manipulation. The sound pressure in the mouthpiece and that of the radiated sound from the bell of the instrument were analysed and compared to gain a better understanding of the acoustic processes involved in player control of brassiness.

• Transient behaviour in the motion of the brass player's lips during a lip-slur
  S. Logie, S. Stevenson, A. Apostoli, J. Chick and M. Campbell [C]
[A]

When moving between notes on an orchestral brass instrument experienced players routinely change from one resonant mode of the instrument to another. Skilled players can 'lip-slur' from one note to another making the transition between resonant modes sound very smooth as perceived by the listener, and do so with little thought or understanding about what is happening to the lip/air-column interaction during this intra-note transient.

This paper offers an investigation of the transient behaviour of the player's lips when moving from a strong, fully developed coupling of the lips to one air column resonance, to the strong coupling of another resonant mode of the same air column. A high speed digital video camera is used to view the motion of the lips through a transparent mouthpiece for a range of orchestral brass instruments of different lengths. Image processing techniques allow us to measure the opening area of the lips as a function of time, and this data is synchronised with pressure data measured in the mouthpiece together with the radiated sound from the bell of the instrument to provide us with greater insight to mechanics of the lip-slur transient.

• Brass wind instruments: can the three dimensional motion of the lips account for the `brassy' sound?
  S. Stevenson, M. Campbell and J. Chick [C]
[A]

One distinctive feature of orchestral brass instruments is that there is a marked change in timbre with dynamic level. It is generally accepted that the explanation for this behaviour is caused by shock-wave generation within the body of the instrument. However, it has also been suggested that changes in the nature of the players' lip vibration with increasing amplitude could play at least a small role in generating this distinctive `brassy' sound. Previous studies have shown that the brassy sound cannot be accounted for by a dramatic change in the behaviour of the lip opening area during extremely loud playing. The experiments described here are designed to test the hypothesis that variations in the motion of the lip parallel to the airflow into the instrument may help produce the `brassy' effect. A high speed digital camera and transparent mouthpieces are used to capture the motion of the lips in the plane perpendicular to the face of the player for different dynamic levels. Measurements have been taken over a variety of instruments and musicians and over a dynamic range from mezzoforte to fortissimo.

Campmans, T.

• Source Identification in Practice, Two Cases with Tonal Noise
  T. Campmans [I]
[A]

Caraeni, M.

• Acoustic Modal Analysis using CFD
  M. Caraeni and M. Oswald [C]
[A]

One of the most important issues in aeroacoustics design and development is the optimization of noise and vibration in both interior and exterior design process. Especially the subjective comfort highly depends on the interior noise (e.g. sunroof buffeting). Besides the experimental investigations of this aeroacoustic phenomenon, numerical simulations become more and more an effective tool to shorten the development period. This paper proposes a new approach on finite volume method for determining the resonance frequencies for interior enclosures. The current formulation uses as input non-uniform mean flow obtained from a steady state solution together with prescribed boundary conditions. The method involves coupling the Linearized Navier-Stokes Equations with the Arnoldi algorithm. The Acoustic Modal Analysis, based on the iterative Implicitly - Restarted Arnoldi method, offers a fast and efficient solver for large industrial problems with inlets and outlets. Using the data from this model, the natural acoustic modes of the system and the according frequencies can be identified. These modes can resonate with interior acoustic source processes. The potential of the new method will be presented by means of several academic and industrial test cases ranging from thermo-acoustics to typical automotive and aerospace applications.

Carette, E.

• Sound localization with and without hearing aids
  T. van den Bogaert, E. Carette and J. Wouters [I]
[A]

Building a realistic spatial representation of our surrounding environment is an important task of our auditory system. This allows us to localize sounds and to understand speech in adverse scenarios. An inaccurate spatial map will degrade localization and speech perception and will lead to an unnatural perception of sounds. The question whether hearing aids transmit enough cues and whether hearing aid users are able to use these cues has been gaining a lot of interest recently. This is partly due to recent technology which enabled binaural hearing aid designs aiming at preserving binaural localization cues and flexible microphone configurations which aim at preserving the monaural spectral cues.

Normal hearing and hearing impaired subjects were evaluated in different localization experiments: elevation and front-back discrimination both focusing on monaural spectral cues and left-right localization focusing on binaural cues. Hearing impaired subjects were tested without and with different types of hearing aids. Most of the hearing impaired subjects were capable of localizing sounds relatively accurate if the test procedure included a frequency specific compensation for their hearing impairment in all three dimensions. A difference between the different hearing aid designs was found. This difference was especially prominent in the front-back and elevation experiments.

Carral, S.

• Frequency domain simulation of a Scottish bagpipe chanter with different tapers using the Harmonic Balance Method
  S. Carral and C. Vergez [C]
[A]

The bore of conical bagpipe chanters, such as those found in Scottish Highland or Border bagpipes, usually consist of several cones of different tapers. Scottish bagpipe makers agree that the position where the cones meet, as well as the specific tapers that form them are crucial factors that influence the sound and behaviour of the chanter. This study is aimed at finding out what influence different tapers have on the radiated sound of the instrument. For this purpose, five simplified virtual chanters with no tone holes, two cones, and bottom tapers that are 0, 0.25, 0.5, 0.75, and 1 degrees wider than the top taper were modelled. The cones meet at approximately half way from the top of the chanter. The input impedance and transfer function of each bore configuration were calculated using the program VIAS. The playing frequency and spectrum inside the mouthpiece were simulated using the Harmonic Balance Method. The reed parameters were kept constant in all bore configurations. The radiated sound was calculated by multiplying the spectrum of the pressure inside the mouthpiece by the transfer function of each instrument. The differences in RMS, pitch and spectral centroid found between the instruments are presented and discussed.

Cauberg, H.

• The Use of Reciprocity in determining Rainfall Noise
  N. Geebelen and H. Cauberg [C]
[A]

The theorem of reciprocity exists in many forms and is therefore applied in a wide range of scientific fields. With regard to building acoustics, a relation exists between the sound radiation and the vibration response of a structure. From this, a formula can be derived that connects the air-borne and structure-borne sound insulation of the structure, being only dependent on frequency. This way it is possible to calculate the air-borne sound insulation when the structure-borne sound insulation is known and vice versa, creating lots of practical possibilities. Rainfall noise comes from the mechanical excitation of light structures by rain drops. The measurement procedure for determining rainfall noise is however quiet elaborate. The question rises if reciprocity can be used to simplify this specific set-up? Or else said, when the type of rain (number of drops, intensity, etc.) is known, is it possible for a light structure to predict rainfall noise from the air-borne sound insulation? The use of reciprocity creates many practical advantages, but the relations and formulas have a lot of preconditions. In this paper it is investigated if these preconditions are met in the case of rainfall noise or if they can be overcome be using certain techniques?

Caupin, F.

• Density Threshold for Acoustic Cavitation in Water
  F. Caupin, K. Davitt and A. Arvengas [I]
[A]

We use a focused 1 MHz acoustic burst to generate large mechanical tensions in bulk water. We describe how this technique allows us to generate cavitation and to study its statistics. We have recently built a fiber optic probe hydrophone, to measure the density threshold for cavitation. We report the temperature dependence of this quantity, and compare it with other experiments.

Chen, Z.

• Auralization of noise recordings behind a simulated noise barrier
  H. Waubke, Z. Chen and W. Kreuzer [I]
[A]

The insertion loss spectra of noise barriers are calculated using the boundary element method to test the psycho-acoustical performance of planned noise barriers. Recordings were made at the reference position 25m behind the planned noise barrier at a road and a railway in 2m height. The recordings are mixed in time to avoid audible pass by events. A boundary element method in two dimensions is used. A two parametric model for grassland developed by Keith Attenborough is implemented to simulate a partially reflecting surface. This model leads to smeared sources that do not allow for the application of the Fast Multipole Method. The insertion loss of the noise barrier is calculated using the difference of the spectrum with and without a noise barrier in 25m distance and 2m height behind the barrier. The assumption of a constant line source in the 2D simulation leads to strong interference effects in the spectra. The curvature of curved noise barriers is orientated away from the source allowing a better sight into the surrounding area. The test results of the psycho-acoustic tests for 3m height curved walls are as good as the results for a 4m height straight wall.

Chick, J.

• Reconstructing the Lituus: A Reassessment of Impedance, Harmonicity, and Playability
  A. Apostoli, S. Logie, A. Myers, J. Kemp, J. Chick and A. Braden [C]
[A]

Bach's motet `O Jesu Christ, meins Lebens Licht' (BWV no. 118), first performed in 1736-7, calls for a Lituus, believed to be a long trumpet-like instrument, with playing characteristics somewhere between a trumpet and an alphorn. Unfortunately, there are no surviving examples of Bach's instrument. Using software recently developed at the University of Edinburgh for brass instrument optimisation, a new bore profile with theoretically calculated input impedance curves has been proposed and submitted to a collaborating instrument maker with the aim of providing a natural instrument, with a distinct timbre, that responds well to the player and is in tune over a wide range of natural notes.

This paper describes the process, computational modelling, and experimental work undertaken in an attempt to ensure a satisfactory final instrument design. An initial validation of this process was achieved, in part, by investigating the playing properties of an instrument with a notionally similar bore profile, a 2.4m long Fiscorn in C (a type of Catalonian flugelhorn). This work compared the measured bore profile, input impedance, playability and intonation with those derived theoretically.

• Embouchure Control of Brassiness at Constant Pitch and Dynamic Level in Orchestral Horn Playing
  L. Norman, J. Chick, M. Campbell and A. Myers [C]
[A]

It is generally acknowledged that an increase in the dynamic level of a brass instrument brings about a change in timbre, often described as 'brassiness'. This phenomenon can be attributed to the generation of shockwaves at higher dynamic levels. However, the point at which this change occurs is not always clearly defined and some brass players have found that, by employing slight changes in their embouchure, they have a certain degree of control over the 'brassiness' at a constant dynamic level.

This paper focuses on how embouchure control influences wave propagation in the air column of a brass instrument during the onset of brassiness. A number of horn players were asked to play a long note at a loud dynamic level at the threshold of the onset of brassiness, and to move between a non-brassy and brassy timbre keeping dynamic level and pitch constant. Different horns were used to explore whether small changes in bore profile affected the ease of manipulation. The sound pressure in the mouthpiece and that of the radiated sound from the bell of the instrument were analysed and compared to gain a better understanding of the acoustic processes involved in player control of brassiness.

• Transient behaviour in the motion of the brass player's lips during a lip-slur
  S. Logie, S. Stevenson, A. Apostoli, J. Chick and M. Campbell [C]
[A]

When moving between notes on an orchestral brass instrument experienced players routinely change from one resonant mode of the instrument to another. Skilled players can 'lip-slur' from one note to another making the transition between resonant modes sound very smooth as perceived by the listener, and do so with little thought or understanding about what is happening to the lip/air-column interaction during this intra-note transient.

This paper offers an investigation of the transient behaviour of the player's lips when moving from a strong, fully developed coupling of the lips to one air column resonance, to the strong coupling of another resonant mode of the same air column. A high speed digital video camera is used to view the motion of the lips through a transparent mouthpiece for a range of orchestral brass instruments of different lengths. Image processing techniques allow us to measure the opening area of the lips as a function of time, and this data is synchronised with pressure data measured in the mouthpiece together with the radiated sound from the bell of the instrument to provide us with greater insight to mechanics of the lip-slur transient.

• Brass wind instruments: can the three dimensional motion of the lips account for the `brassy' sound?
  S. Stevenson, M. Campbell and J. Chick [C]
[A]

One distinctive feature of orchestral brass instruments is that there is a marked change in timbre with dynamic level. It is generally accepted that the explanation for this behaviour is caused by shock-wave generation within the body of the instrument. However, it has also been suggested that changes in the nature of the players' lip vibration with increasing amplitude could play at least a small role in generating this distinctive `brassy' sound. Previous studies have shown that the brassy sound cannot be accounted for by a dramatic change in the behaviour of the lip opening area during extremely loud playing. The experiments described here are designed to test the hypothesis that variations in the motion of the lip parallel to the airflow into the instrument may help produce the `brassy' effect. A high speed digital camera and transparent mouthpieces are used to capture the motion of the lips in the plane perpendicular to the face of the player for different dynamic levels. Measurements have been taken over a variety of instruments and musicians and over a dynamic range from mezzoforte to fortissimo.

Christophe, J.

• Aeroacoustic Prediction and Measurements of Jet-Airfoil Interaction Noise
  J. Christophe, C. Schram and M. Tournour [I]
[A]

The noise emitted by the landing gear the and high-lift devices is a significant component of the overall noise radiated by aircrafts. In these applications, unsteady turbulent interactions inside the flow, and interactions of the flow with solid boundaries play an important role.

These mechanisms are investigated in this work by considering the case of a turbulent jet that interacts with a large span airfoil. Our prediction method is based on the aeroacoustical analogy, which is well suited for applications with low-Mach numbers. A flow model is used to generate equivalent aeroacoustic sources, which noise may in turn be propagated towards the listener. Incompressible Large Eddy Simulation is used to provide the flow data, and two approaches are combined to yield the acoustic field over the frequency range of interest. Curle's analogy is employed at low Helmholtz numbers (based on the airfoil chord), and Amiet's theory is implemented to obtain the large-frequency spectrum including chordwise non-compactness effects. Both the flow field and the acoustic field are compared with hotwire and microphone measurements conducted in the anechoic facility of the von Karman Institute.

Cigada, A.

• Active Attenuation of Squeal Noise at Train Wheels
  H. Siebald, H. Fehren, M. Schiedewitz, A. Cigada, S. Manzoni and M. Redaelli [I]
[A]

The present work represents preliminary experimental investigation aimed at the possibility to prevent the squealing noise of a resilient train wheel under real operating conditions. The control effectiveness of two different kind of piezoelectric actuators have been tested in a non-rotating test stand designed with the aim of reproducing the wheel working conditions as closely as possible. It is not possible to fully simulate the rotating wheel behaviour, but promising results and important information have been obtained with collocated control. In the next step piezoelectric stack actuators have been adopted on a rotating wheel, on the condition that slip rings have been installed both for signals and power transmission. Tests with a prepared wheel set have been performed on a rotating test stand. A good vibration reduction at the wheel and also a noise reduction in the nearfield of the wheel have been achieved.

Cik, M.

• Traffic Noise Annoyance on Roads and Rail (TNAR) in an experimental laboratory setup
  M. Cik, K. Fallast, R.B. Raggam, E. Marth, E. Gallasch, A. Lackner and M. Fellendorf [C]
[A]

At present, according to the current standards and specifications, the impact of road and rail traffic noise is represented by the A-weighted energy-equivalent sound level (LA,eq). This quantity does not account for the subjective annoyance of sound events as it is perceived by the affected persons.

Pass-by road and rail vehicle noise samples were binaurally recorded under predefined conditions with a dummy head measurement system and synthetically composed to defined vehicle ensembles. A total of 200 persons were selected and exposed to these defined vehicle ensembles in different sessions in a hearing laboratory. Annoyance was rated on an 11-graded interval scale and by a hand power-dynamometer which measures the intensity of the applied strength. Psychoacoustic parameters of the defined vehicle ensembles were calculated. Annoyance ratings were combined with psychoacoustic parameters by multiple regression analysis. The relevant parameters were statistically combined into the index TNAR.

The analyses revealed that loudness is the parameter describing annoyance best, significantly better than the LA,eq (P0.0001). Further significant influencing parameters are sharpness and a new designed parameter representing the frequency characteristic of traffic noise. These three parameters together (R=0.807), correlate significantly better with the subjective estimation of noise-induced annoyance than the LA,eq (R=0.672) does (P0.0001).

Clausen, M.

• Handling Scanned Sheet Music and Audio Recordings in Digital Music Libraries
  C. Fremerey, D. Damm, M. Müller, F. Kurth and M. Clausen [I]
[A]

Significant digitization efforts have resulted in large music collections, which typically contain various types and modes of data. In particular, such collections may comprise audio data in form of CD recordings as well as image data in form of scanned sheet music, thus concerning both the auditory and the visual modalities. In this paper, we review various techniques for automatically processing, analyzing, and organizing such multimodal music collections. We then present a music player system along with various interfaces that allow a user to conveniently search, navigate, and browse across the two types of music data. For example, during audio playback our system synchronously highlights the corresponding musical measures within the sheet music. Furthermore, a user may formulate a query by marking a region of measures within the scanned score to trigger an audio search. The system retrieves all available audio recordings along with the temporal positions of the audio segments that correspond to the queried measures. The user may then freely navigate across the various retrieved interpretations. Our techniques and interfaces are prototypically integrated into the library service system of the Bavarian State Library as part of the PROBADO project.

• Towards timbre-invariant audio features for harmony-based music
  S. Ewert, M. Müller and M. Clausen [I]
[A]

In the context of automatic music processing, chroma features have turned out to be a powerful midlevel representation with a wide range of applications such as content-based retrieval, music synchronization, and chord transcription. An important step of the feature calculation is the grouping of spectral energy components that belong to the same pitch class or chroma of the equal tempered scale. Here, the octave identification introduces a high degree of invariance to changes in timbre and instrumentation. In this paper, we introduce a strategy to further increase this invariance by combining the concept of chroma features with the well-known concept of mel-frequency cepstral coefficients (MFCCs). The lower MFCCs are known to capture information on timbre. Therefore, to enhance robustness to timbral changes, we discard this information by only keeping the upper coefficients. Furthermore, using a pitch scale instead of a mel scale allows us to group the remaining coefficients into twelve chroma bins. The resulting audio features, as indicated by our systematic experiments, have indeed gained a significant boost towards timbre invariance without a degradation of discriminative power.

Condorachi, D.

• Urban noise mapping in Romania – preliminary results for Iasi city
  D. Condorachi [C]
[A]

Noise mapping is known to be a complex task. As a newcomer in EU, Romania has to comply the EU Directive regarding noise control especially in the urban areas. Iasi is the second largest city in Romania (340,000 inhabitants in 2002) and is in the first group of large cities that has to create noise strategic maps. Hybrid methods to collect data were used, required digital thematic layers were extracted in a GIS environment (TNTMips v.6.9 from Microimages Inc.) from cadastral maps and updated from orthorectified air imagery, numerous points for monitoring traffic at fixed hours were settled, classification of the roads regarding the traffic, database creation, points validation using a Quest sonometer, etc. Final integration in a software for noise mapping modeling (SoundPlan v.6.4) enabled us to compute the first city noise map for Iasi area using recommended EU noise calculation methods and standards.

Coninx, F.

• Application of binaural technology in an adaptive auditory speech test for children
  J. Fels, F. Coninx and W.H. Döring [I]
[A]

This contribution reports on the development of a screening procedure for the identification of hearing loss in children by using a speech test in various spatial conditions.

A study looking at the performance of the Adaptive Auditory Speech Test (AAST) under spatial locatable noise as a possible diagnostic procedure for Auditory Processing Disorders (APD) has been carried out. The AAST was developed to determine the Speech Recognition Threshold (SRT) in a fast and reliable way under quiet and noisy conditions for children of 3 to 4 years of age.

In the study the AAST is used to develop a diagnosis method with a noise made by a talk between children (Two-Talker-Noise) or talking noise (person reading a newspaper) to test the intellectual abilities, separation and localization.

The test was carried out under various spatial conditions with the signal and noise coming from different positions in the horizontal plane. The signals were generated using head-related transfer functions which are appropriate for this age group. Therefore, the test words and the talker-noise were convolved with the corresponding children HRTFs. The scenarios were presented with the help of headphones.

Significant differences were detected between the different spatial situations. Results and experiences from current screening programs will be presented.

Connelly, T.

• Using the Fast Multipole Boundary Element Method to Update Loads in System Level SEA Models
  T. Connelly, J. Cordioli, V. Cotoni, S. Müller and A. Wilson [C]
[A]

Automotive interior noise at mid and high frequencies is dominated by the airborne noise from acoustic sources spatially distributed around the vehicle. Given the vehicle size and the frequency range of interest (typically above 400 Hz), it is standard practice to use Statistical Energy Analysis (SEA) models of the entire vehicle to help designing for interior noise. Accurate description of the acoustic loads on the exterior SEA subsystems (windows, windshield, roof, etc...) is then required.

Exterior sound pressure fields can be efficiently modeled with the boundary element methods (BEM). However, because of the computational expense, standard BEM is limited to fairly low frequencies for full vehicle geometry. The Fast Multipole Method (FMM) allows to extend the BEM to reach the frequency range of interest for airborne noise applications.

This paper demonstrates and investigates the use of Fast Multipole BEM to calculate the automotive exterior acoustic loads at mid and high frequencies for inclusion in a system level SEA model. Numerical experiments are performed to define some guidelines regarding the sensitivity of the prediction to (i) geometrical details, (ii) mesh density, (iii) frequency resolution, (iv) surface impedances, and (v) details of the source description.

Conter, M.

• Noise Reducing Devices: an Austrian Experience with the new European Technical Specification for Measurements of in-Situ Sound Diffraction
  M. Conter and M. Haider [C]
[A]

Traffic noise is one of the major environmental concerns within the European Union. The most important solution to noise abatement along roads and railways is the construction of noise barriers. In order to measure the acoustic performance in real life conditions and the long-term performance of these devices, the working group CEN/ TC226/ WG6 has developed the so called Adrienne method. This measurement method was introduced in 2003 with the technical specifications CEN/TS 1793-5, dealing with in-situ sound absorption and sound insulation, and CEN/TS 1793-4, dealing with in-situ sound diffraction. Added devices for sound diffraction are becoming more relevant in the last years to increase the acoustic performance of barriers. This paper describes the first Austrian research dealing with measurements of in-situ sound diffraction.This study, carried out in 2005 by arsenal research, compares the acoustic behaviour of aluminium cassettes, alternately equipped with reference barriers, partially non flat products, added devices on the top of the barrier, transparent insertions and digital and digital prints on the surface. In detail, the authors have measured and analysed the in-situ acoustic properties sound absorbtion, sound insulation and sound diffraction of each variant. The research has been carried out for the company Forster Metallbau GmbH.

Cordioli, J.

• Using the Fast Multipole Boundary Element Method to Update Loads in System Level SEA Models
  T. Connelly, J. Cordioli, V. Cotoni, S. Müller and A. Wilson [C]
[A]

Automotive interior noise at mid and high frequencies is dominated by the airborne noise from acoustic sources spatially distributed around the vehicle. Given the vehicle size and the frequency range of interest (typically above 400 Hz), it is standard practice to use Statistical Energy Analysis (SEA) models of the entire vehicle to help designing for interior noise. Accurate description of the acoustic loads on the exterior SEA subsystems (windows, windshield, roof, etc...) is then required.

Exterior sound pressure fields can be efficiently modeled with the boundary element methods (BEM). However, because of the computational expense, standard BEM is limited to fairly low frequencies for full vehicle geometry. The Fast Multipole Method (FMM) allows to extend the BEM to reach the frequency range of interest for airborne noise applications.

This paper demonstrates and investigates the use of Fast Multipole BEM to calculate the automotive exterior acoustic loads at mid and high frequencies for inclusion in a system level SEA model. Numerical experiments are performed to define some guidelines regarding the sensitivity of the prediction to (i) geometrical details, (ii) mesh density, (iii) frequency resolution, (iv) surface impedances, and (v) details of the source description.

Corteel, E.

• Qualitative evaluation of Wave Field Synthesis with expert listeners
  A. Pras, E. Corteel and C. Guastavino [I]
[A]

A qualitative evaluation of Wave Field Synthesis sound quality was conducted with a panel of 5 expert listeners to investigate the perceptual effect of partial decorrelation at high frequencies and of reducing the number of active loudspeakers. The experiment was carried out on a 24 loudspeaker array. In each trial, participants were presented with different versions of the same auditory scene (using 24 or only 8 speakers, with and without partial decorrelation in high frequencies) and asked to freely describe the perceived differences. The verbal data was analyzed to derive semantic scale for sound quality evaluation.

Cotoni, V.

• Using the Fast Multipole Boundary Element Method to Update Loads in System Level SEA Models
  T. Connelly, J. Cordioli, V. Cotoni, S. Müller and A. Wilson [C]
[A]

Automotive interior noise at mid and high frequencies is dominated by the airborne noise from acoustic sources spatially distributed around the vehicle. Given the vehicle size and the frequency range of interest (typically above 400 Hz), it is standard practice to use Statistical Energy Analysis (SEA) models of the entire vehicle to help designing for interior noise. Accurate description of the acoustic loads on the exterior SEA subsystems (windows, windshield, roof, etc...) is then required.

Exterior sound pressure fields can be efficiently modeled with the boundary element methods (BEM). However, because of the computational expense, standard BEM is limited to fairly low frequencies for full vehicle geometry. The Fast Multipole Method (FMM) allows to extend the BEM to reach the frequency range of interest for airborne noise applications.

This paper demonstrates and investigates the use of Fast Multipole BEM to calculate the automotive exterior acoustic loads at mid and high frequencies for inclusion in a system level SEA model. Numerical experiments are performed to define some guidelines regarding the sensitivity of the prediction to (i) geometrical details, (ii) mesh density, (iii) frequency resolution, (iv) surface impedances, and (v) details of the source description.

Coutinho, A.

• The Sound Insulation of Water
  R. Wenmaekers, B. van der Aa, A. Pronk, A. Coutinho and L. van Luxemburg [C]
[A]

Liquid filled multilayer membranes can be used as a building material for roofs [Pronk et all. 2007]. Between the membranes a fluid is injected for heat or cold transfer. The sound insulation of the fluid filled membranes is mainly determined by the fluid. Since fluids are unusual building materials, the sound insulation is unknown and hard to predict by common prediction models. Sound insulation measurements have been performed in a small laboratory setup [Rodrigues and Coutinho, 2008]. To investigate the sound insulation of water further, sound insulation measurements of a 26, 44 and 83 mm thick water layer have been performed according to ISO 140-3 in the Laboratorium voor Akoestiek of Eindhoven University of Technology. A water basin was created in the 10 m floor opening, constructed on a wooden structure with steel net and foil. The resulting airborne sound insulation index Rw of the water layers are 35, 38 and 42 dB respectively. The graph of the sound insulation per frequency shows an average increase of about 3 dB/octave. The doubling of the water thickness also results in an increase of about 3 dB. It is shown that water is a useful material in terms of sound insulation properties.

Cranen, B.

• Evaluation of missing data techniques for in-car automatic speech recognition
  Y. Wang, R. Vuerinckx, J. Gemmeke, B. Cranen and H. van Hamme [I]
[A]

Background noise has a detrimental effect on the accuracy of automatic speech recognition (ASR) systems because it changes the statistical properties of the spectrally related features extracted from the speech. If the recognizer's acoustic model that describes these statistics is trained under noise conditions different from those of the test data, the likelihoods of speech hypotheses will deviate and recognition errors result. Many practical systems therefore compensate the spectral features for noise with techniques like spectral subtraction, modify the statistics by acoustic model adaptation or train the acoustic model under multiple noise conditions. In our ASR system which is based on missing data techniques (MDT), spectral regions that are dominated by noise are replaced by an estimate that depends on the speech hypothesis of the decoder. In this paper, we discuss the pro's and con's of this approach in terms of computational requirements, training data requirements and versatility. In a benchmark study using in-car recorded data, as well as other non-stationary noise types, we compare the accuracy of a MDT system with that of the commercially available Nuance VOCON 3200.

Cremezi-Charlet, C.

• The UIC-project Noise Impact of Composite Brake Blocks (Nicobb)
  N. Meunier, F. Létourneaux and C. Cremezi-Charlet [C]
[A]

Since the adoption of the TSI Noise by the European commission, all new freight wagons taken into operation have to fulfil stringent noise limits, making the use of composite brake shoes of the type K indispensable. The coefficient of friction of K-blocks is much higher than that of cast iron blocks. Therefore a retrofit of existing freight wagons would require an expansive modification of the brake system. UIC (International Union of Railways) supports the development of new composite brake blocks of type LL who combine the positive effect of K-blocks on wheel treads with the coefficient of friction of cast iron shoes. This opens the opportunity to substitute cast iron blocks by LL-blocks within regular maintenance, hence making the LL-blocks very attractive for freight operators. Acoustic assessment of brake blocks generally requires field tests, which are very time and money consuming. Thus there is a need of simplifying the acoustic validation procedure to encourage manufacturers to develop new products. The project Nicobb is a UIC project whose members are SNCF and Deutsche Bahn. It aims at developing a testing procedure for the validation of low roughness levels retained by new composite brake blocks as a low-cost alternative to field tests.

Cutanda Henriquez, V.

• Calculation of Head Related Transfer Functions of bats using the Boundary Element Method
  P. Juhl, V. Cutanda Henriquez and D. Vanderelst [I]
[A]

ChiRoPing (Chiroptera, Robots, and Sonar) is an EU-funded research project aimed at understanding how bats use their echolocation perception ability and apply this knowledge to the design of new robotic senses. Four species of bats are selected for the study and models of their heads including minute details of ears, mouth and nose are obtained through CT scans. The project involves, among other things, the use of numerical methods on such scanned models to study the role of their features in the bat sensorial performance. As the bats operate at very high frequencies and as their ears and noses are very complicated and involve thin surfaces, the numerical modelling is challenging. The paper reports preliminary results of the numerical study and investigates the possibilities of reducing the computational complexity of the models.

• Numerical design and testing of a sound source for secondary calibration of microphones using the Boundary Element Method
  V. Cutanda Henriquez, P. Juhl and S. Barrera Figueroa [I]
[A]

Secondary calibration of microphones in free field is performed by placing the microphone under calibration in an anechoic chamber with a sound source, and exposing it to a controlled sound field. A calibrated microphone is also measured as a reference. While the two measurements are usually made consecutively, a variation of this procedure, where the microphones are measured simultaneously, is considered more advantageous from the metrological point of view. However, it must be guaranteed that the two microphones receive the same excitation from the source, although their positions are some distance apart to avoid acoustic interaction.

As a part of the project Euromet-792, aiming to investigate and improve methods for secondary free-field calibration of microphones, a sound source suitable for simultaneous secondary free-field calibration has been designed using the Boundary Element Method. The source has a central plug with the effect of reducing the spatial variations of the sound pressure at the microphone positions within the frequency range of the calibration. After some preliminary measurements, the source has been tested thoroughly in a calibration setup at the acoustic laboratory of the Danish Fundamental Metrology Institute (DFM). The design and verification of the source are presented in this communication.

d'Udekem, D.

• Vibro-Acoustic Simulation of Structure-borne Induced Radiation of Ship Windows
  B. van Antwerpen, D. d'Udekem and C. Weißenborn [C]
[A]

Large ship windows seem to represent a weak point in terms of acoustic radiation within a ship cabin. There are is a emerging need to predict the behavior of those components prior to their final design, construction and assembly on board.

In this framework, numerical methods offer a good platform for predicting the acoustic radiation and structural vibration of such windows or to compare the performance of different windows system designs. For instance, it allows predicting the radiation efficiency, which represents a key point in reducing the global sound pressure level within the cabin. This paper presents a vibro-acoustic study performed with the finite element code ACTRAN. The radiation efficiency of different window configurations is estimated. The focus is placed on the determination of the most sensitive system parameters.

In this paper, the simulation method is first briefly described. The obtained results are also compared to a set of full scale experimental measurements for validation. Different window designs approaches are compared in order to guide the choice of an optimal design to select.

Damm, D.

• Handling Scanned Sheet Music and Audio Recordings in Digital Music Libraries
  C. Fremerey, D. Damm, M. Müller, F. Kurth and M. Clausen [I]
[A]

Significant digitization efforts have resulted in large music collections, which typically contain various types and modes of data. In particular, such collections may comprise audio data in form of CD recordings as well as image data in form of scanned sheet music, thus concerning both the auditory and the visual modalities. In this paper, we review various techniques for automatically processing, analyzing, and organizing such multimodal music collections. We then present a music player system along with various interfaces that allow a user to conveniently search, navigate, and browse across the two types of music data. For example, during audio playback our system synchronously highlights the corresponding musical measures within the sheet music. Furthermore, a user may formulate a query by marking a region of measures within the scanned score to trigger an audio search. The system retrieves all available audio recordings along with the temporal positions of the audio segments that correspond to the queried measures. The user may then freely navigate across the various retrieved interpretations. Our techniques and interfaces are prototypically integrated into the library service system of the Bavarian State Library as part of the PROBADO project.

Dannemann, M.

• Design and Tests of Thermoplastic Textile-reinforced Composite Trays for Vibro-acoustic Relevant Applications
  W. Hufenbach, M. Dannemann, S. Friebe, F. Kolbe and M. Krahl [C]
[A]

Driven by increasing customer demands, environmental and economical requirements, lightweight engineering gains in importance for various industrial and transport applications. Here, composite materials based on thermoplastic hybrid yarns offer new possibilities for multifunctional design due to its adjustable properties. Requirements concerning high stiffness and damping as well as low weight and short cycle times during manufacture can be met by an adapted composite lay-up and manufacturing process. In the presented work, the vibro-acoustic behaviour of selected textile-reinforced thermoplastics was characterized dependent on the kind of reinforcement, number of layers and its fibre orientation. Therefor, models based on Finite Element Method and Boundary Element Method for the calculation of modal and acoustic parameters of textile-reinforced structures were developed. For the verification of this simulation models as an example composite trays were manufactured by a novel hot pressing technology. The modal parameters and transmission loss of the produced trays were measured. Subsequently, the experimental results were used for the verification of the numerical simulations. The performed investigations illustrate that the vibro-acoustic behaviour of composite trays is influenced by the complex interaction of geometry and material parameters. Thus, an optimal sound radiation behaviour can only be achieved by a comprehensive material-adapted vibro-acoustic design.

• Identification of the Optimal Lay-up for Fibre-reinforced Composite Plates with Minimized Sound Emission
  M. Dannemann and W. Hufenbach [I]
[A]

Increasing customers requirements and environmentally demands, the shortened development times as well as the necessary weight reduction and function-integration are pressing influence factors for the design of future lightweight structures. In the course of these challenging demands, composites with fibre- or textile-reinforcements offer new possibilities for a function-integrated design due to their versatile property profile. The big amount of adjustable parameters like layerwise fibre orientation, thickness and material requires the utilization of efficient optimization methods, which take into account not only the stiffness but also the vibration and sound behaviour. For this, an optimization tool based on a semi-analytical model for the calculation of the sound radiation of composite plates was developed. The design tool enables the calculation of optimal parameters to given requirements. Within this paper the results of selected optimization scenarios are presented including the influence of frequency/stiffness limits and mass restrictions on the optimal composite lay-up with minimized sound radiation. The developed tool and the performed studies allow the derivation of design guidelines giving the engineer the possibility of fully exploit the lightweight and vibro-acoustic potential of composite materials.

Dantscher, S.

• Methods for the determination of individual sound attenuation of ear plugs
  S. Dantscher [I]
[A]

The European directive 2003/10/EC on noise at the workplace defines exposure limit values that may not be exceeded at the ear of the worker. Thereby, also the attenuation of hearing protectors has to be taken into account. For checking the compliance with this requirement usually the attenuation data specified by the manufacturer is used. It is based on the measurement for the EC type examination with a group of test subjects. But the individual values of attenuation may differ from the group average. To cope with this problem some manufacturers of hearing protectors and especially custom moulded ear plugs have developed measurement systems for the determination of individual sound attenuation. Different techniques are used: objective measurements without active participation of the wearer of the hearing protector (e.g. with two microphones inside and outside of the plug) or subjective methods (e.g. hearing thresholds with and without hearing protector). This presentation gives an overview of the different measurement systems on the market. Moreover the calculation (and correction) methods will be discussed that are necessary to determine an individual value of sound attenuation that is comparable to the laboratory data.

Dassen, A.G.M.

• New Dutch legislation for motorways, national cost-benefit analysis explained.
  A. van Beek and A.G.M. Dassen [C]
[A]

The Dutch state is in a process to modernize the national legislation for noise from major roads and railroads. The objectives are to stop increase of noise and to reduce the amount of high exposed houses and other noise sensitive buildings. An important aspect is that the costs for noise measures to achieve those goals should be restrained. The legislation consists of both environmental as well as economical aspects. The new structure that is choosen consists of ''noise emission ceilings'' to stimulate the most cost effective measures at noise sources. The Netherlands Environmental Assessment Agency has evaluated the plans for a new form of Dutch legislation by modelling the results of legislation in full perspective. The results of the study show the framework of costs, amounts of high exposed houses and expected noise annoyance due to highways within different scenarios of traffic growth. With these figures the choices in legislation are made visible and comparable with other options.

Dauenhauer, T.

• Crying colours and their influence on loudness judgments
  D. Menzel, T. Dauenhauer and H. Fastl [I]
[A]

The German term ''schreiende Farbe'' denotes a very salient colour which stands out among other colours. It can be translated as ''loud colour'' or literally as ''crying colour''. As it is known from previous studies that the colour of objects like trains or cars can influence loudness judgments, experiments were performed to determine if such crying colours lead to higher loudness ratings compared to non-crying colours. As a first step, subjects were asked to synthesise colours which they perceived as ''crying'' by adjusting values in the Lab colour-space. A representative sample of these colours was then combined with non-crying colours. The resulting group of 15 colours was rated on a five-point scale ranging from ''not-crying'' to ''very crying''. In a final listening experiment, the same 15 colours were presented simultaneously with uniform exciting noise of different sound pressure levels to asses the influence of the visual stimuli on the subjective loudness ratings. Results indicate that some crying colours where able to elicit - at same SPL of the noise - higher loudness ratings compared to a neutral colour.

Davaadorj, T.

• Properties of Bubbles at Pressure Antinodes in a Standing Sound Field
  T. Davaadorj, P. Koch, A. Pluta, U. Parlitz and W. Lauterborn [C]
[A]

The parameter space for trapping bubbles in an acoustic resonator is limited by various factors, notably dissolution of the bubble by diffusion of the gas inside the bubble into the surrounding liquid, the primary Bjerknes force driving strongly oscillating bubbles away from the pressure antinode and parametric instabilities leading to surface oscillations of large amplitude. The region of stability in the parameter plane made up of the bubble radius at rest and the driving sound pressure amplitude at a frequency of 1 MHz is determined for different static pressures. The temperatures at bubble collapse in this region are calculated. The maximum attainable temperatures appear along the upper stability boundary in bubble resonance conditions.

Davies, B.

• Speech Communication in Outdoor Soundscapes
  B. Davies, P. Mahnken and C. Plack [I]
[A]

The positive soundscape project is a large multi-disciplinary researching the perception of soundscapes. Recent findings indicate that speech communication is a principal factor in users' perceptions of urban soundscapes. The project has therefore explored how this factor might be quantified. This paper reports on an attempt to use speech intelligibility index (SII) to characterise time-varying speech intelligibility in real outdoor soundscapes. The relationships between SII, subjective intelligibility and subjective quality will be explored, as functions of signal-to-noise ratio. Possible applications in sound quality mapping of soundscapes will be discussed. Potential implications for rational planning of soundscapes will be outlined.

Davitt, K.

• Density Threshold for Acoustic Cavitation in Water
  F. Caupin, K. Davitt and A. Arvengas [I]
[A]

We use a focused 1 MHz acoustic burst to generate large mechanical tensions in bulk water. We describe how this technique allows us to generate cavitation and to study its statistics. We have recently built a fiber optic probe hydrophone, to measure the density threshold for cavitation. We report the temperature dependence of this quantity, and compare it with other experiments.

de Beer, E.

• Whistling Building Objects, Origins and Solutions
  J. Granneman, M. Trooster, E. de Beer and T. van Diepen [C]
[A]

Widespread complaints often occur due to whistling sounds caused by building objects, such as (high rise) buildings and viaducts, under specific meteorological circumstances, especially wind speed and direction. In many cases this is related to the application of specific grating. Much is already known about the relevant parameters for this phenomenon from measurements in practice and in laboratories. To obtain more understanding into this subject and to be able to prognosticate the possible occurrence and intensity of it in the design phase of building objects, systematic research has been done under laboratory conditions. The sound power level and frequency characteristics of whistling sounds dependent of wind speed, angle of incidence and grid configurations were studied. Also the sound reducing effect of different and practically feasible modifications of gratings was part of the study. The paper presents the results of this study.

de Boer, A.

• On Solving the Tyre-Road Contact Problem at High Frequencies
  Y. Wijnant, J.H. Schutte and A. de Boer [I]
[A]

The contact between tyre and road is non-linear and most accurately described in the time domain. Currently however, accurate finite element models are too large to be solved at the small time scales required to capture tyre/road noise. At the Structural Dynamics and Acoustics group of the University of Twente an alternative contact algorithm has been developed. The algorithm solves the dynamic equations and, while solving, satisfies exactly the contact condition (the condition that states that there is no penetration of the tyre into the road). As a result, there is no need for contact elements and no additional parameters are required to stabilize the solution process. The possibility to optimize and speed up the algorithm by means of multigrid is the major advantage of the new approach. In this paper the contact algorithm is explained for a finite element discretization.

• Contribution of stiffeners to the sound transmission of aircraft skin panels
  J. Wind, A. Bronsvoort and A. de Boer [I]
[A]

Traditionally, the sound radiation of aircraft skin panels is calculated using a structural model with stiffeners and an acoustical model where the stiffeners are omitted. We present a numerical study on the impact of this omission.

The numerical model consists of three parts. Firstly, the pressures in a diffuse sound field are modeled. Secondly, a Finite Element (FE) model calculates the structural response caused by the sound field. Thirdly, a Boundary Element (BE) model calculates the sound intensity. The impact of stiffeners is determined by comparing the results of a BE model with and without stiffeners.

The emphasis lies on the results of this model. Several stiffener configurations and frequencies are compared and a number of new analysis techniques are introduced to pinpoint the cause of these differences. Differences of upto 50% have been found for the geometries that have been tested.

de Bree, H.-E.

• PU nosecone intensity measurements in a wind tunnel
  H.-E. de Bree and E. Tijs [C]
[A]

Flow induced noise is known to depend on details in the geometry and surface of an object. Acoustic near field measurements in the flow might be helpful in analysing the noise sources. As accelerometers can not be used to measure flow induced noise, acoustic particle velocity sensors become a viable alternative worth while considering.

Although acoustic particle velocity sensors are susceptible to wind, they can be used in the presence of a flow if the sensors are mounted in a nosecone. Cross correlation techniques can be used to circumvent the issue of flow induced noise even further.

Experiments carried out up to 70m/s at the windtunnels at University of Twente and NLR will be presented and discussed.

• A PU sound probe for high noise levels
  E. Tijs and H.-E. de Bree [C]
[A]

For several high noise level applications, e.g. jet engines, the standard PU and USP probes can not be used. In order to meet these challenging acoustic requirements, both the particle velocity and the sound pressure transducers have been optimized. The novel design will be explained and its acoustic performance will be discussed. Attention will be paid to the calibration of such a probe.

• Large scale in situ acoustic reflection measurements in a theatre
  E. Tijs, H.-E. de Bree and E. Brandao [C]
[A]

The acoustic absorption properties of materials used in a theatre are a major concern, both in the initial design stage and during renovation. The PU surface impedance method allows the simultaneous measurement of both sound pressure and acoustic particle velocity, providing the opportunity to measure materials as they are installed on seats, floors, walls and ceilings, assessing their sensitivity to both normal and oblique angles of sound wave incidence. Using a portable piece of equipment, actual testing time is short, allowing large scale testing for comparisons. The method was applied at the Musis Sacrum theatre in Arnhem, the Netherlands. Results will be presented and findings will be discussed.

• An acoustic vector based approach to locate low frequency noise sources in 3D
  H.-E. de Bree, C. Ostendorf and T. Basten [C]
[A]

Although low frequency noise is an issue of huge societal importance, traditional acoustic testing methods have limitations in finding the low frequency source. It is hard to determine the direction of the noise using traditional microphones.

Three dimensional sound probes capturing the particle velocity vector and sound pressure offer a novel possibility to locate the low frequency noise sources in 3D. An overview on the benefits of acoustic vector sensors will be given and a low frequency case study will be presented and discussed.

The method has been tested during low frequency noise measurements in an urban environment. In this case a very specific frequency of the source was known. This frequency was used to tune the measurement system. The paper will describe the results of the test.

• An in situ method to measure the acoustic absorption of roads whilst driving
  E. Tijs and H.-E. de Bree [C]
[A]

Traditionally, the acoustic absorption properties of road surface asphalts are determined in a Kundt's tube, requiring a sample to be taken, thus destructing the surface. Furthermore, this method is time consuming and only limited to measure sound wave under a normal angle of sound wave incidence, which is not the case in practive. The PU surface impedance method, measuring both sound pressure and acoustic particle velocity,has been applied on asphalt road surfaces, measuring acoustic absorption under both normal and oblique angles of sound wave incidence. The method can also be applied while moving the probe along the test surface. Thus it becomes feasible to measure the acoustic properties of road surfaces whilst driving. As many mechanical properties are related to acoustic absorption, a road surface monitoring system based upon acoustic absorption measurements might be feasible. First results will be presented and an outlook will be discussed.

• A source path contribution analysis on tire noise using particle velocity sensors
  H.-E. de Bree, D. Bekke and J.H. de Vries [C]
[A]

Road tire noise is an important topic of research where acoustic particle velocity based testing techniques can be expected to bring new insights. Modal analysis can be carried out using non contact particle velocity sensors, and PU sound probes can be used to measure the radiated sound without a need to use anechoic testing conditions. A further breakdown of the overall sound pressure levels measured in to its various sources can be made by applying a source path contribution analysis, using the PU probes to measure velocity, sound intensity and, for determining the reciprocal transfer path, the sound pressure. The concept of using this type of transfer path analysis will be outlined and illustrated by a tire noise case.

de Coensel, B.

• Informational masking and attention focussing on environmental sound
  D. Botteldooren and B. de Coensel [I]
[A]

Appreciation of a soundscape depends on the cognitive and emotional evaluation of the sonic environment with particular emphasis on the meaning associated to the perceived sound. Before meaning can be attributed, a sound has to be detected and recognised. Energetic masking due to the working of the inner ear has been mentioned as a reason for not hearing an environmental sound. The central nervous system however also plays role. Informational masking may prohibit recognising or understanding the sound. In this context attention focussing is always around the corner. As attention gets focussed on an auditory stream, informational masking could be reduced even if the signal to noise ratio drops considerably after initial detection. It is for this reason that we include attention shifting in our computational model for environmental noise perception. The model starts from a separation of auditory streams based on spectro-temporal characteristics. A robust mechanism using activation and inhibition models the ability of auditory streams to attract attention. Top down attention is related to activities. The outcome of the model includes the time that particular environmental sounds are heard, their level, saliency, ... This could allow to validate designs based on masking unwanted components of the sonic environment.

de Jong, C.

• Characterization of Ships as Sources of Underwater Noise
  C. de Jong [I]
[A]

There is a growing interest in the possible impact of anthropogenic underwater noise on marine life. One of the concerns is the increasing contribution of shipping noise, with the growing number and size of commercial ships. Traditionally, underwater radiated noise control was only of interest for naval and research vessels. Due to the potential environmental impact, it becomes relevant for commercial shipping also. The challenge is to bring the expertise from the naval and research area to the maritime industry. One of the issues is the measurement of ship radiated underwater noise. The definition of the ship as a noise source, the measurement set-up and the environment in which the measurements are taken all have a large impact on the results of ship underwater noise measurements. The various influences are discussed. The elimination of environmental effects from the radiated noise characteristics is illustrated by means of numerical simulations and the analysis of measurement trials with a point source in different environments.

• Model Scale Measurements of Surface Ship Radiated Flow Noise
  C. de Jong, J. Bosschers and H. Hasenpflug [C]
[A]

Advances in weapon and sensor capabilities are driving an increased interest in the control of underwater signatures of naval platforms. The control of machinery and propeller noise is well understood but there is a shortfall of knowledge of the mechanisms that govern noise due to the flow around the hull of a surface ship. The subject has been investigated by the collaborative US-NL project Mechanisms and Prediction of Surface Ship Radiated Flow Noise. The project aimed to determine and quantify the sources of flow noise generated by surface ships in terms of ship speed and hull shape. The focus was on the underwater noise generated by turbulence excited hull plating vibration and noise resulting from breaking bow waves. A combination of full scale, large scale, model scale tests and computational fluid dynamics was used. The paper will discuss the test set-up and results of the model scale experiments which were made in a (highly reverberant) towing tank. A special silent towing carriage was developed to reduce the background noise, while an acoustic antenna was used to measure the noise due to the breaking bow waves. The pressure fluctuations due to the turbulent boundary layer were measured at several hull locations.

de Klerk, D.

• Application of Operational Transfer Path Analysis on a Classic Car
  D. de Klerk, M. Lohrmann, M. Quickert and W. Foken [I]
[A]

This paper discusses the analysis of a classic car with the Operational Transfer Path Analysis (OTPA). This method poses a fast and efficient way of identifying critical vibration paths and source contributions within vehicle design. Only measurement data of the operating vehicle are needed to perform this analysis. Real measurement situations for example are several vehicle run-ups, full load, halve load, run down. These kinds of measurements are commonly performed in the automotive industry and can therefore be easily adopted. The method OTPA is introduced and discussed based on a classic car Wartburg 311. This vehicle is described in the project sketch ''method comparison of the transfer path analysis'' (Fraunhofer IWU Dresden) and provided by Westsächsische Hochschule Zwickau, Institut für Kfz- Technik (IfK).

de Mey, F.

• On two possible mechanisms for call directionality steering in the rufous horseshoe bat, Rhinolophus Rouxi
  F. de Mey, D. Vanderelst and H. Peremans [C]
[A]

In this paper we investigate a possible echolocation strategy for the rufous horseshoe bat (Rhinolophus rouxi). As this is a cf-fm bat, it concentrates most energy in the single frequency cf-component of its call. Therefore it cannot exploit a frequency dependent emission pattern focusing energy into different spatial regions, as is done by fm-bats using frequency modulated calls. This limits spatial information gathered from a single call. It has been conjectured that the typical pinna movements displayed by cf-bats would enhance their localization capacity. In this paper, we focus on two alternative means of generating additional spatial information, based on call directivity modification: on the one hand performing a controlled noseleaf movement, on the other, changing the phase between the emissions produced from the two nostrils. The latter mechanism would in effect constitute a phased array (2 elements approximately half a wavelength apart) allowing simple beam steering. We show how realistic noseleaf movements affect the call directivity. To this end, we simulate the acoustic field produced by the bat's emission system using a 3D boundary element method. In addition, we explore what phase shifts are required to produce useful beam steering when considering the emitter a phased array.

de Roo, F.

• Uncertainty of Close-Proximity (CPX) tyre-road noise measurements - Round Robin test results
  F. de Roo, J. Telman, G. van Blokland, H. van Leeuwen, J. Reubsaet and W.-J. van Vliet [I]
[A]

Practical experience in the Netherlands with the Close-Proximity (CPX) measurement method for the influence of road surfaces on traffic noise (ISO/CD 11819-2) showed that the measurement uncertainty of this so-called trailer method is relatively large. In order to gain more knowledge about the determining factors a well-coordinated Round-Robin test was organised and combined with a statistical investigation of the primary parameters that contribute to the uncertainty. Five organisations that perform CPX-tests on a commercial basis participated in the Round-Robin test and made their measurement trailer and its technical specifications available for this research. This paper will describe the test set-up of the Round-Robin test and the results achieved. Based on these results recommendations for a stricter specification of the test conditions and the design parameters of measurement trailers will be presented.

de Ruiter, E.

• A tool for environmental noise control in urban planning: The Population Annoyance Index
  E. de Ruiter [C]
[A]

In the planning stage often studies (like EIA) are conducted to compare alternative locations for a new urban area or the alignment of a new road. Comparison of the noise load of the alternatives is part of the studies. In simple cases e.g. where a generic noise reduction is achieved the comparison is simple. In most cases however comparing the alternatives will be more complicated. How to weigh a small increase of the number of dwellings exposed to high noise loads, to a large decrease of the number of less noise loaded dwellings? This problem demands a metric for the total amount of noise annoyance to be expected in a certain context (planning of a new residential section or road). For this purpose a metric is proposed that is easy to calculate, called Population Annoyance Index (PAI). The PAI is based on dose-response relationships and the density of dwellings/people in the noise loaded areas; the method then results in a one-number metric for the ''total annoyance'' of the specific variant of the project. If minimising the amount of annoyance is the target of noise control, the PAI is a useful indicator in comparing the quality of the alternatives of the project.

de Vos, P.

• An estimate of the global exposure to environmental noise
  P. de Vos [C]
[A]

The data produced in the frame of the European Noise Directive, incomplete as it is, provides a useful basis to estimate (1) the relative importance of various sources of environmental noise, (2) the distribution of noise exposure in different classes relevant to health effects, and (3) the influence of demographic parameters to the noise exposure. Based on these relationships, an extrapolation for the world population's exposure to noise has been made. The result supports the conclusion that environmental noise, in spite of ambitious legislative efforts, is a major and growing concern with respect to both public health and quality of life. This notice ought to be sufficient justification for a much higher political urgency, priority and interest than is expressed today.

de Vries, D.

• Plane wave decomposition: an inverse problem in array processing
  L. Hörchens and D. de Vries [I]
[A]

The decomposition of wave fields into plane waves is a standard tool for the analysis of impulse responses measured with an array of microphones. If the geometry is appropriately chosen, the decomposition can be obtained by spatio-temporal deconvolution. Several measures can be taken to tackle the ill-posedness of the associated mathematical problem. Typical examples are given by beamforming techniques or plane wave decomposition using circular or spherical harmonics. These approaches are compared based on the presented framework. The stability of the procedure for plane wave decomposition is strongly influenced by the choice of the array setup, which is extensively used for current approaches. On the processing side, regularisation techniques and prior knowledge can be employed to overcome instabilities and improve the achievable resolution. Exemplary results based on this approach are shown, pointing out the potential of further research in this field.

• Deriving Room Acoustical Parameters Using Arrays and Hearing Models
  J. van Dorp Schuitman and D. de Vries [I]
[A]

Using microphone arrays, room acoustical parameters can be determined by measuring room impulse responses along a line, circle or sphere. With this technique the distribution of the parameters within (parts of) a room can be evaluated. The method also revealed that some of the parameters may fluctuate as a function of microphone position, in such an amount that it does not correspond to human perception. In recent research, attempts were made to apply hearing models on the measurements to yield more perceptually motivated results. Listening tests were set up to find correlations between subjective attributes and the model results.

• On the applications of array technology in room acoustics
  A. Schlesinger and D. de Vries [I]
[A]

Room acoustic parameters vary much with the position of the receiver and the source as a consequence of the interferences in the anisotropic part of the sound field. It is thus difficult to yield exhaustive information of the room properties from independent single-point measurements. Using an array measurement permits to evaluate the sound field with a high spatial resolution and leads to a more precise assessment of the room acoustic properties. We propose a technique to investigate room acoustics by reconstructing the volumetric sound field from measurements taken on a sphere. The latter are obtained by a virtual spherical array using a microphone mounted at a robotic arm. The spatially sampled room impulse response upon the sphere is used to reconstruct the sound field within a particular volume of reconstruction by applying the methods of Nearfield Acoustical Holography (NAH). Subsequently, acoustical properties are derived from the reconstructed volumetric room impulse response. A virtual spherical microphone-array was constructed and tested in a volumetric modal analysis and a volumetric sound intensity probe. The results are discussed and an outlook on the potential of the presented method is given.

• Analysis of Enclosed Sound Fields using Multichannel Impulse Response Measurement
  M. Strauß, J. Nowak and D. de Vries [I]
[A]

When evaluating the performance of spatial audio systems, the knowledge about the sound propagation process inside the listening area gives additional information to assess the influence of room acoustics on the reproduction quality. Wave field analysis based on multichannel impulse response measurement delivers insight to the temporal and spatial properties of a sound field. Closely spaced impulse responses are usually recorded along a line, a circle or a grid resulting in a coherent acoustic data set. This enables the possibility for analyzing the spatial characteristics of a sound field. By applying a Fourier transform or Radon transform it is possible to identify room reflections.

As an example a set of impulse responses was collected at a two dimensional grid inside the listening plane of a small enclosure. Because of the fact that a 2D array recording contains 3D information, it was expected to identify the strongest reflections and to obtain directional information from the data.The analysis methods were validated against a multitrace impulse response set calculated from the simulation of a simple room model. In a second step the functions were applied to real measurement data of a car compartment.

• Spherical Array Systems - On the Effect of Measurement Errors in Terms of Perceived Auralization Quality
  F. Melchior, Z. Kuang, D. de Vries and S. Brix [I]
[A]

The analysis of acoustic fields using an open cardioid spherical microphone array enables the extrapolation of the measured field to arbitrary positions inside the reconstruction area of the array. An application of such a system is the generation of virtual stereo main microphone setups, which can be used for auralization purposes. Practical limitations of a array system will lead to several types of errors. In this work the effects of measurements errors on the perceptual characteristics of diffuse fields is investigated. In a listening experiment two channel ideal microphone setups are compared with corresponding virtual sensor setups based on spherical microphone array measurements.

de Vries, J.H.

• A source path contribution analysis on tire noise using particle velocity sensors
  H.-E. de Bree, D. Bekke and J.H. de Vries [C]
[A]

Road tire noise is an important topic of research where acoustic particle velocity based testing techniques can be expected to bring new insights. Modal analysis can be carried out using non contact particle velocity sensors, and PU sound probes can be used to measure the radiated sound without a need to use anechoic testing conditions. A further breakdown of the overall sound pressure levels measured in to its various sources can be made by applying a source path contribution analysis, using the PU probes to measure velocity, sound intensity and, for determining the reciprocal transfer path, the sound pressure. The concept of using this type of transfer path analysis will be outlined and illustrated by a tire noise case.

Deblauwe, F.

• Capturing a noise source in an interior enclosure
  F. Deblauwe and M. Robin [I]
[A]

Sound source localization (SSL), such as NAH and beam-forming, have been around for the last decennia. SSL techniques have been mainly used for free field conditions and only the last couple of year they have found their way for interior acoustic applications where non-free-field conditions are met. In this paper a two step test procedure is proposed to perform a detailed interior SSL localization. In a first step, a solid spherical antenna is used to perform a SSL on the complete interior compartment. The uniqueness of this solution lays in the fact that the SSL propagation is not only based upon a beam- forming solution, but takes also into account the acoustic diffractions that happens around the solid sphere. The first step gives an overview of all sources present in the interior enclosure, but with a limited spatial resolution. In a second step this spatial resolution is improved by taking a second measurement focusing in at certain areas of the enclosure. For these measurements a cylindrical array is used and the data is processed using focalization. The combination of the two methods leads to an accurate SSL in an interior enclosure in a more efficient way that has been done till now using masking techniques.

Dejonckere, P.

• Simulation of differences between male and female vocal fold configuration during phonation
  M. Kob, P. Dejonckere, E. Calderón and S. Kaynar [I]
[A]

Statistically, some vocal fold lesions such as vocal fold nodules - considered as a localized tissue reaction to voice loading - are much more frequent in female ( 95%) than in male ( 5%) adult humans. This can only partially be explained by the about two times higher oscillation frequency during voice production in females. In previous studies, gender-related differences in the shape of the vocal fold edge have been observed that can be related to different geometries of the laryngeal framework. With a time-domain model of vocal fold function the oscillation patterns and resulting forces between and inside the vocal folds have been simulated, and the conditions for eliciting a hourglass-shaped vibration pattern (necessary for determining the localized tissue reaction of the nodules) could be clearly defined. Our simulations and measurements support the assumption that collision and tearing forces on the vocal fold edges during phonation are higher and more specifically located in females compared to males.

Delfs, J.

• Broadband Simulation Of Flow-Induced Noise Generation in Air Conditioning Ducts
  C. Spehr, R. Ewert, T. Ahlefeldt and J. Delfs [C]
[A]

Air conditioning systems in cabins are possible noise sources influencing the passenger comfort in civil aircrafts. The current study continues an initiative to predict the generated noise on components of air distribution systems. The approach presented here is based on the numerical inexpensive RANS-simulation. Using a perturbation technique the turbulent flow fluctuations are represented by artificial turbulences injected into the mean flow field. In the first part of the study the time dependent perturbation was induced by injecting individual artificial vortices which allow qualitative comparison of the sound generation of different arrangements on the basis of the overall sound power level. In the current work a broadband turbulence is induced in the source region. The turbulence is thereby generated by filtering white noise to reproduce the autocorrelations and integral length-scales provided by a RANS-simulation. The interaction of the turbulence with AC-components generates pressure fluctuations with broadband character which are compared with corresponding measurements. Arrangements of two consecutive orifices were investigated as typical components in the air conditioning system. The sound generation on the downstream orifices depends therefore not only on its shape, but also on the flow disturbance, which in turn depends on the distance to the upstream orifice.

Delgado, A.

• Sound Generation in a Flow-Induced Vibrating Human Vocal Folds Model
  S. Becker, S. Kniesburges, S. Müller and A. Delgado [I]
[A]

For investigating the fluid-structure-acoustic coupled (FSA) process of human phonation, synthetic models of the vocal folds (VF) were designed which showed flow-induced oscillations and produced a tonal sound. Two different kinds of models were developed: one for observing the flow field vibrating at constant frequency and a more realistic which could be stretched in longitudinal direction to simulate the prestress within human VFs. Besides the investigation of the flow field using PIV the major aim was to observe the FSA process to determine the acoustic sources. Therefore two approaches were applied: Firstly the vibration of the VFs was recorded using a high-speed camera triggered by the subglottal pressure to determine phase shift between the two signals. Secondly synchronous measurements of the subglottal pressure, the flow velocity behind the VFs and the acoustic pressure were done to determine their coherence. The results support the existence of the Coanda-effect. Furthermore the vibration frequency and the phase shift between the subglottal pressure and the glottal width depended on the stiffness and the tension in the VFs. The synchronous measurements determined the pulsating flow rate as main acoustic source.

Demi, L.

• Medical acoustical array expertise at Delft University of Technology
  K.W.A. van Dongen, E.J. Alles and L. Demi [I]
[A]

The history of medical ultrasound dates back as far as the mid 30s, first as a therapeutic tool (e.g. deep tissue heating) and a decade later as a diagnostic tool based on pulse-echo reflections. Nowadays, a majority of all medical imaging modalities are based on ultrasound as the technique is often non-invasive, harmless and relatively cheap and still results in high image quality.

High image quality is obtained using array technology, i.e. a large number of active elements (preferably smaller than a wavelength) are positioned either in a single line (1-D array) or in a surface (2-D array). By applying a predefined time delay to the signal of each element, electronically or via software on the measured data, it becomes feasible to focus or steer the beam to a volumetric region of interest.

At the TU Delft in the section Acoustical Imaging and Sound Control we apply array technology in various medical applications. Among the applications are breast cancer detection, hyperthermia breast cancer treatment, transesophageal echo cardiography (TEE) and intravascular ultrasound. During our presentation, we will show for each of the above applications the challenges (and solutions) of implementing array technology in clinical applications.

Desanghere, G.

• QCITY: providing cities validated mitigation measures for noise action plans
  G. Desanghere [C]
[A]

QCITY (Quiet City Transport) is an integrated research project sponsored by the EC-sixth research framework - priority: sustainable surface transport. A major objective of the project is to provide municipalities with a guide with noise mitigation tools to establish action plans for compliance with Directive 2002/49/EC. This paper describes the development of a validated selection of mitigation measurers, fully documented with information such as the expected noise reduction, cost estimations, limitation in applicability. In addition to the selection list, datasheets with further documentation and references are provided. This guide focuses on practical and validated actions that can be undertaken by the municipalities themselves in their search for quiet city transportation. It involves mitigation measures such as traffic planning, noise reduction of road and rail surfaces, propagation paths and barriers, economic incentives, ... .

Descheemaeker, J.

• Surface and guided waves in porous materials
  J. Descheemaeker, L. Boeckx and W. Lauriks [C]
[A]

The last decades, a lot of research has been done about the propagation of sound in heterogeneous and porous materials. Especially the two-phase nature of the poro-elastic material leads to interesting physical phenomena. During recent years, several measuring methods have been developed to determine the material parameters. However, a lot of problems aren't solved yet. Because of the visco-elastic behavior of a lot of these materials, the elastic moduli will become frequency dependent (rubberlike behavior at low frequencies and glasslike behavior at high frequencies). Experimental data about the frequency- and temperature dependence of the elastic moduli are scarce because of experimental difficulties. Some experimental and numerical results about this subject will be presented . Porous materials still have to be studied in a lot of configurations. One of them is a porous material in an elastic cylinder. Some numerical results about the dispersion curves of the wave propagating in the system will be presented.

Desmet, W.

• A Ring-based Tyre Model for the Prediction of Structure-borne Interior Tyre/Road Noise
  P. Kindt, P. Sas and W. Desmet [I]
[A]

The tyre/road interaction causes a structural excitation of the tyre. This vibrational energy is transmitted through the suspension towards the vehicle body. The resulting body panel and window vibrations cause noise radiation in the vehicle passenger compartment. In order to predict this interior noise during the vehicle development phase, an accurate tyre model is required. This paper presents a structural tyre model, based on a three-dimensional flexible ring on an elastic foundation. The ring represents the belt and the elastic foundation represents the tyre sidewall. The model is valid till 300 Hz and includes a submodel of the wheel and the air cavity. Unlike most ring models, which only consider in-plane modes, the presented model also predicts modes that involve torsion of the belt in circumferential direction. The parameterization of the model, which does not require detailed knowledge of the tyre construction, is based on the main geometrical properties of the tyre and a limited modal test. The unloaded tyre response is calculated and validated with test data. As the presented model is physical, it can be applied to describe other operational conditions such as loading and rotation. In this paper, the response of the loaded tyre is also considered.

• Iterative learning control for an active exhaust noise attenuation valve for internal combustion engines
  R. Boonen, G. Pinte, W. Desmet and P. Sas [I]
[A]

An active silencer has been developed to attenuate combustion engine exhaust noise. The silencer consists of an electrically controlled valve connected to a buffer volume. The pulsating flow from the engine is buffered in the volume. The valve connected to it is controlled such that only the mean flow passes to the atmosphere. This flow is free from fluctuations and consequently free of sound. An iterative learning controller has been developed to control the active valve. In the control algorithm, non-causal filters are implemented. These filters allow to control systems with high dynamics over a broad frequency range. The active silencer has been experimentally validated on a cold engine simulator, which generates realistic exhaust noise and gas flow using compressed air. The exhaust noise has been reduced in a frequency range starting from 5 Hz until 800 Hz. Depending on the rotational speed of the engine, the typical reductions ranges from 13 dB to 24 dB.

• Quantification of the Equivalent Sources of Different Tire Noise Models by means of the ASQ Technique
  D. Berckmans, P. Sas and W. Desmet [I]
[A]

Tire noise is a predominant noise contributor in many traffic situations nowadays. Accurate models are therefore needed to predict the spectrum that results at receiver locations due to the tire rolling under different operating conditions. Predicting the spectrum relies on a source - transmission path - receiver model which involves the modeling of all noise sources and the determination of all relevant transfer paths.

Locating and quantifying the different noise sources on an operating tire remains an inherent difficult task. The sources are thought to be numerous and complex, with their relative magnitudes changing with tire specifications, road specifications and operating conditions.

A rolling tire was experimentally characterized on a roller bench by means of the substitution monopole technique: the running tire is substituted by the non-operating tire covered by monopoles. The monopoles are quantified by means of the airborne source quantification (ASQ) technique, based on operating indicator pressure measurements in the neighborhood of the rolling tire. The role of regularization techniques, important in the quantification process, is highlighted and results for different tire models are presented. Moreover, the contributions of the different tire segments to the resulting spectrum at the receiver position are evaluated.

Díaz, D.

• Reworking spoken dialogue systems with context awareness and information prioritisation to reduce driver workload
  J.L. Blanco, Á. Sigüenza, D. Díaz, M. Sendra and L. Hernández [C]
[A]

An entire host of new information, communication and entertainment applications and security systems are assaulting our cars. Interaction with onboard systems is moving beyond the classic switches, knobs, sticks, pedals and dashboard lights and sounds, to incorporate speech, touch-screens and a variety of other displays. While some of these systems are intended to help with the task of driving (e.g., navigation aids, lane departure and vehicle-in-blind-spot warnings), a constellation of displays, messages and requests threaten to overload the driver with information that is too rich and may distract him. Speech being eyes- and hands-free, it is increasingly being considered as a safe communication channel. However, interactive dialogue can increase the driver's mental workload, especially when robustness problems occur (e.g., recognition errors). Here we present our research results on dialogue as a secondary task, to gain understanding on how it may affect driving, on what dialogue strategies foster safer driving and on which applications, messages and interaction modalities should be prioritised over which others in each driving context. The proposed contribution is related to our current research activities within MARTA (Mobility for Advanced Transport Networks; www.cenitmarta.org), a Spanish, publicly-funded project. Experimental results from tests with a simulated driving platform are presented.

Dible, S.

• Temporal and Spectral Characteristics of a Marine Piling Operation in Shallow Water
  P. Lepper, S. Robinson, J. Ablitt and S. Dible [I]
[A]

Analysis of the underwater radiate acoustic characteristics for marine piling operations for two pile diameters, 2m and 4.74m, in a relatively shallow water site are presented. Measurements of the entire piling sequence for several piles were conducted at ranges from 10m to 22km for piles in 10-20 m water depth. Variations in the temporal and spectral characteristics of radiated energy are analysed in context of pile size, range from source, hammer energy used and pile penetration depth.

The hammer energy used shows a strong interdependence between mechanical strike 'hammer' energy and underwater radiated acoustic energy. This process appears 'coarsely' linear for individual piling operations although considerable variation in overall gradient were observed between operations. Temporal and spectral variations in radiated energy due to pile penetration are also examined for fixed hammer energy and range.

Simultaneous recordings of radiated energy made at increasing distances from the pile showed evidence of temporal and spectral dispersion effects consistent with relatively shallow water propagation. Correlation of received levels at various ranges in differing seabed topographies were made suggesting complex shallow water modal propagation dependant on both the source and environment characteristics including seabed topography, sediment type and water column acoustic properties.

Dietrich, P.

• Prediction of the sound radiation from a plate excited by a structure-borne sound source
  M. Lievens, P. Dietrich and C. Höller [I]
[A]

The radiation of structure-borne sound into a medium is generally predicted by linear system theory. In characterising a structure-borne sound source and its radiation, the goal is to predict the sound pressure at an arbitrary point in the medium using only mobilities and frequency responses of the individual system components. This seems straight forward but in practice the prediction often fails due to one or more of three causes: an incompletely measured coupling situation, a high measurement uncertainty of the mobilities, or the non-linear source behaviour.

To investigate the validity of the multipole theory a small one-dimensional measurement setup was built. It consists of a small electrical motor connected to a plate that radiates into a receiving room of one cubic metre. In this way many positions, plates and motor excitations can be analysed with limited effort.

The analyses will compare power measured in situ to power predicted using in situ measured mobilities and source excitations. The sound pressure in the receiving room is used to verify the assumption of a one dimensional excitation process and to monitor the error in the prediction.

• How to obtain high quality input data for auralization?
  P. Dietrich and M. Lievens [C]
[A]

Auralization of structure-borne and airborne noise problems contributes to understanding of sound transmission in a significant way. One expects the auralization results to be as close to real world result as possible. In hybrid models where measurement and simulation data are both used to generate the final audible result, the measurement data obviously has to be very precise. Even highly accurate numeric simulations require measured material data, e.g. structure-borne impedances. Therefore, the measurement of input data is of high interest regarding the simulation quality.

Measuring structure borne impedances up to high frequencies for ongoing coupling calculations often lacks of insufficient signal to noise ratio and phase errors. Mostly measured force and acceleration signals are directly used for impedance calculations without exploiting the advantage of impulse response.

This contribution aims to clarify the influences of special post-processing of the measurement data on the obtained impedance by means of deconvolution, bandpass filtering and time-windowing.

Based on a measurement setup to study the prediction of the sound-radiation of a small structure-borne sound source, the signal processing for the measured impedances will be presented and discussed. Nevertheless, this methodology can be applied for airborne impulse responses and other transfer paths as well.

• Comparison of Sound-Source Localization Techniques for Vibrating Structures
  M. Müller-Trapet and P. Dietrich [C]
[A]

A thorough comparison of sound-source localization methods for vibrating structures, exploring their theoretical and practical limitations, will be presented. The beamforming technique will be focused on as this method has become more and more popular and various approaches and algorithms exist that have to be evaluated for their accuracy and usability. This work concentrates on the planar case, which is assumed to be the most common one.

As a practical example, velocity measurements on vibrating plates will be compared to the output of different beamforming algorithms. To thoroughly investigate the signal processing, a virtual measurement setup based on the measured velocity data will be used. The microphone arrays and their input will be modeled using Boundary-Element-Method (BEM) simulations in order to be independent of microphone directivity and measurement uncertainties. The different results of the beamforming algorithms, the velocity measurements and the simulated sound intensity will be compared and discussed.

Dietz, M.

• The Effect of Envelope Waveform on Lateralization
  M. Klein-Hennig, M. Dietz, S.D. Ewert and V. Hohmann [C]
[A]

Interaural timing disparities are important cues for lateralization in the human auditory system. The auditory system is sensitive to timing disparities in the fine-structure and the envelope of sounds. At high frequencies, only envelope disparities can be exploited due to the lack of phase-locking to the fine-structure. In previously published data [e.g. Bernstein and Trahiotis in ''Hearing - from basic research to applications'', Springer 2007] parameters like the modulation frequency, or the power of a modulating sinusoid have been altered in order to investigate the role of the envelope waveform on lateralization. However, these ''classical'' parameters simultaneously change several ''secondary'' envelope parameters like attack and decay times, as well as pause and hold durations, which might conceal the fundamental cues exploited by the auditory system. In this study, psychophysical measurements were conducted with customized envelope waveforms in order to investigate the effect of isolated secondary parameters on lateralization. For high-frequency tones centred at 4 kHz with systematic envelope modifications, the just noticeable differences of interaural envelope time differences were measured. The results indicate that attack times and pause durations prior to the attack are the most important envelope features. The results are compared to predictions by a binaural auditory model.

• Azimuthal Localization of Concurrent Speakers Employing Interaural Coherence
  M. Dietz, S.D. Ewert and V. Hohmann [I]
[A]

A computational model for the localization of concurrent speakers is presented that extends the binaural model presented by Dietz et al. (Fortschritte der Akustik - DAGA 2007, 107-108). The processing is motivated by mammalian physiology and human psychoacoustic data. A wide range of interaural parameters is exploited, namely the interaural level difference (ILD), the interaural phase difference (IPD) of the fine structure, the IPD of the fundamental modulation frequency and the interaural coherence (IC). The parameters are determined for each auditory filter which allows a simultaneous estimation of several sound directions. By focusing on instants with high interaural coherence the robustness of the estimation could be greatly improved (Faller and Merimaa, 2004, J. Acoust. Soc. Am., vol. 116, 3075-3089). For two well-separated speakers in quiet the estimation error is generally less or equal 10. In noisy environments a single speaker can be localized with an error of less than 20, even for a signal-to-noise ratio of 0 dB.

Dietzel, R.

• Measuring Frequencies with Historic Resonators from SCHAEFER
  R. Hoffmann, D. Mehnert, R. Dietzel and G. Fuder [I]
[A]

Measuring frequencies of acoustic processes within the audible range required large technical expense in the early times of experimental phonetics and acoustics. Indirect methods like KUNDT's tube, the monochord, the interpretation of kymographic recordings, or the interference tube from QUINCKE allowed such kind of measurement, but had been too complicated for practical purposes. Therefore simpler measurement methods have been looked for in early times. E. g., it is possible to apply one-sided closed pipes of variable length as resonators. The acoustic perception of the experimentator was used as indicator.

This paper investigates this measuring method at first within the historic arrangement. Then, the transfer behaviour of single tube resonators from SCHAEFER will be measured in the free sound field.

These results as well as the ability of judgement of the listener are used to estimate the measuring uncertainty of such historic methods.

Dijckmans, A.

• A Wave Based Model to predict the Airborne and Structure-Borne Sound Insulation of Finite-Sized Multilayered Structures
  A. Dijckmans, G. Vermeir and W. Lauriks [C]
[A]

In this paper, a Wave Based Method (WBM) is used to predict the airborne and impact sound insulation of finite-sized multilayered structures. The WBM is a Trefftz-based deterministic prediction technique for the steady-state dynamic analysis of coupled vibro-acoustic systems. The field variables (sound pressures and plate displacements) are expanded in terms of acoustic and structural wave functions, which are exact solutions of the homogeneous dynamic equations. The contributions of the wave functions depend only on the boundary and continuity conditions. Because the modal behaviour of plates and rooms is taken into account, the WBM is not limited to the high-frequency range like Statistical Energy Analysis. The enhanced computational efficiency, as compared to other deterministic methods like finite element models, allows predictions in a broad frequency range (typically 50-5000 Hz in building acoustics). As many problems in building acoustics can be simplified to a rectangular geometry, the WBM is developed for this specific case. Predictions for multilayered structures containing air layers and porous layers are compared with analytical prediction models and measurements. The influence on the sound reduction index of different parameters, like concentrated absorption in sending or receiving room, is discussed.

Dittmar, C.

• Preprocessing methods for rhythmic mid-level features
  C. Dittmar, M. Gruhne and D. Gaertner [I]
[A]

Rhythmic mid-level features are an important pre-requisite in Music Information Retrieval. They can be deployed to describe the rhythmical gist of songs for various tasks, such as genre classification and music similarity search. Although different computation strategies have been proposed in the literature, rhythmic mid-level features commonly represent the most salient rhythmic periodicities in a music signal. Major shortcomings of rhythmic mid-level features are their dependency on the actual tempo of the songs and their susceptibility to degradation caused by interference with non-rhythmic signal components, such as melodic-sustained instruments. This publication describes preprocessing strategies applicable to rhythmic mid-level features. The problem of interference with melodic instruments is addressed by emphasizing the rhythmic content via a signal decomposition related to Drum Transcription. Tempo changes are tackled by logarithmic re-sampling of the features' lag a xis. An evaluation of the proposed methods is conducted via genre classification using both artificial and real-world music signals. The evaluation results are presented and discussed accordingly.

Dittrich, M.

• Procedure and applications of combined wheel/rail roughness measurement
  M. Dittrich [I]
[A]

Wheel-rail roughness is known to be the main excitation source of railway rolling noise. Besides the already standardised method for direct roughness measurement, it is also possible to measure combined wheel-rail roughness from vertical railhead vibration. This is a different quantity which offers its own particular applications. This measurement technique has significant practical benefits and has been in use for several years. In this paper the method is outlined and it is discussed how it could be implemented in a procedure. Applications are discussed and some illustrative data are presented.

Doclo, S.

• Preserving binaural hearing of hearing impaired subjects with binaural noise reduction systems for hearing aids
  T. van den Bogaert, S. Doclo, J. Wouters and M. Moonen [I]
[A]

Hearing aid users experience great difficulty understanding speech in noisy environments. Therefore noise reduction algorithms are introduced in hearing aids. The development of these algorithms is typically done monaurally. However, the human auditory system is a binaural system. Providing two monaural, independently operating, noise reduction systems to the hearing aid user may disrupt binaural information needed to localize sound sources correctly and to improve speech perception in noise.

We first examined the influence of commercially available, bilateral, noise reduction algorithms on binaural hearing. Objective and perceptual evaluations with normal hearing and hearing impaired subjects show that fixed and adaptive directional microphone (ADM) configurations, two of the most commonly used noise reduction configurations in hearing aids, can significantly distort the binaural properties of the sound signals. As a second step, binaural noise reduction schemes based on a multichannel Wiener filter (MWF) approach were developed and evaluated. It is observed that a binaural hearing aid design significantly increases noise reduction performance owing to the larger number of microphones used. Moreover, the binaural MWF and the binaural MWF with partial noise estimation provide a better combination of noise reduction performance and preservation of binaural cues compared to the bilateral ADM algorithm.

Döbler, D.

• Comparision of various Array Geometries with Respect to their Depth of Field for acoustic Mappings
  O. Jaeckel, R. Schröder and D. Döbler [I]
[A]

The wide-spread use of planar microphone arrays for acoustic 2D-beamforming mappings does not consider the influence of actual threedimensional object surface structures on the calculated maximum sound pressure levels. To quantify the effects of the very complex interactions between array- and objectgeometry in more detail, several investigations of common two- and threedimensional arrays with varying focus distances have been performed. All simulations were done using the correct spherical wave model for point sources for the distance calculations. In the simulations, ring- and spiral arrays as well as spherical 3D-arrays will be compared. Measurement examples for some of the arrays will also be shown and will be related to the simulations. The results show that most arrays have a remarkably strong dependence on the focal distance actually used for the beamforming calculation. Especially for the case of beamforming within 3D-interiours, which is very difficult to handle due to strong individual distance variations, it is therefore inevitable to take the actual object-geometry into account to get useful results at all. So the use of 3D-CAD-models for acoustic 3D-mappings is a well-founded decision, even if it makes software-implementations much more complex than the easy but oversimplifying use of just planar beamforming maps.

• Threedimensional Beamforming considering the Effects of covered Microphones
  O. Jaeckel, D. Döbler, R. Schröder and A. Meyer [I]
[A]

Beamforming onto complex geometric structures in three dimensions must take into account surface areas from which the emitted soundwaves can not reach the microphone array directly. Similar to optics, the beamformer is faced with the problems of ''full shade'' (no microphone has a direct path to the acoustic source) and ''half shade'', i.e. the acoustic source is only visible partly by a certain subset of all microphones. The article presents two methods to geometrically take those areas into consideration in 3D-beamforming. The first method is a relatively simple one that can only determine either full shading or full visibility of some parts of the object geometry. It gives good results in most measurement scenarios when the microphone array is compact or has a comparatively small diameter, e.g. with a small spherical array in car interiours. The second solution presented is computationally far more intensive but also more general. This method does also consider the ''half shade'' baffle effects and therefore it also allows measurements with distributed microphone arrays onto 3D-structures from the outside. The necessary modifications of the beamforming-algorithm for both cases are presented and the resulting effects will be discussed using practical measurement examples.

Döllinger, M.

• Numerical Modelling of Vocal Fold Dynamics by a 3D Multi-Mass-Model
  A. Yang, J. Lohscheller, M. Stingl, D. Voigt, U. Eysholdt and M. Döllinger [I]
[A]

The emitted human speech signal originates due to oscillations of both vocal folds. So called functional dysphonia being only observable during phonation yield voice disorders, i.e. hoarseness. In historical voice research, the endoscopic vocal folds movements were mostly simulated via 1D Two- or 2D Multi-Mass-Models. However, by using these models only vocal fold movements in lateral and longitudinal directions are taken into account.

This work presents an enhancement of these mathematic biomechanical models: a Multi-Mass-Model expanded from 2D to 3D displacements. It consists of five mass planes arranged in vertical direction. Each plane contains five longitudinal coupled mass-spring oscillators. Via the new proposed model not only the vertical movements but also the diffusion of the mucosal wave from inferior to superior can be simulated.

In order to better simulate vocal fold vibrations, the time-invariant model parameters are adjusted in a developed hybrid optimization procedure. The evaluation of optimization procedure is performed by using lateral symmetric data sets synthetically generated by the suggested model. Different glottis closure-types are also considered.

The future goal is to describe the physiological relevant characteristics of 3D vocal fold dynamics, which can be extracted via high-speed imaging in combination with a new developed laser-spot projection system.

• Numerical study of the human phonation process by the Finite Element Method
  S. Zörner, M. Kaltenbacher, G. Link, R. Lerch and M. Döllinger [I]
[A]

The basis of the human phonation process is given by the complex interaction of air and flow in the larynx together with the structural mechanics of the vocal folds. Our goal is to study the influence of the geometric shapes of vocal folds on the quality of the generated sound. Since direct measurements on such variations are not feasible, we developed an enhanced numerical scheme based on 2d Finite Element Methods, capable to compute this coupled field problem. The geometric variations on the vocal folds change the main frequency component and, furthermore, influence the occurrence of the Coanda effect (the jet is attaching randomly to either side of the trachea). With our simulations, we can demonstrate, that this effect causes a broadband acoustic signal. Currently, we perform numerical simulations with asymmetric vocal folds to investigate different types of dysphonia.

• Optimizing Material Properties and Geometry of a Physical Multi-Layered Vocal Fold Model
  B. Schmidt, G. Leugering, M. Döllinger, R. Lerch, M. Kaltenbacher and M. Stingl [I]
[A]

Understanding the physical fundamentals of human voice generation is a challenging and interdisciplinary task. The present work takes a closer look at a self-sustained physical model of the human larynx, in which the vocal folds are modeled by multi-layered elastic bodies, i.e. silicone layers. We search for optimal material and geometry parameters in the sense that the deformation of the multi-layer model at its surface is as similar to given deformation patterns arising from hemilarynx experiments as possible. For this purpose, we employ an optimization concept based on a 2D finite element simulation in which the vocal fold is modeled as linear elastic body. Keeping the layer geometry fixed, we use the material optimization approach in order to identify optimal material parameters for each layer by assuming isotropic or transversal-isotropic material properties. Additionally, we employ techniques from shape optimization, in order to optimize geometry of the boundaries between the different layers parameterized as spline curves, while keeping the material tensors fixed. We consider the simultaneous solution of both problems, discussing optimization of material and geometry concurrently. We conclude this presentation by giving some numerical results using a 2D vocal fold model.

Dörfler, M.

• On the Structure of the Phase around the Zeros of the Short-Time Fourier Transform
  F. Jaillet, P. Balazs, M. Dörfler and N. Engelputzeder [C]
[A]

The short-time Fourier transform is a time-frequency representation widely used in audio signal processing. At first sight, the phase of the transform may appear to be difficult to interpret, a fact that led to the development of many processing schemes in audio processing relying on the modulus of the transform only. However, it has been shown, that the phase information can be successfully exploited for improved analysis and processing. This is illustrated for example by the reassignment method or the phase vocoder algorithm.

In order to better exploit the phase information in audio processing, a better understanding of its structure is desirable. This study describes a particular behavior of the phase which can be observed when looking at the phase derivative over time or frequency around the points where the transform equals zero. This recurrent behavior is illustrated using numerical simulations on synthesized and real audio signals. The performed simulations show the influence of the parameters of the transform on the observed structures. Some consequences of this phase structure for audio processing are also presented.

• Removing Components from a Time-Frequency Representation
  M. Dörfler, P. Balazs and F. Jaillet [C]
[A]

Time-frequency representations such as the spectrogram or short-time Fourier transform seem to be well suited to the task of removing certain components with approximately disjoint support in the time-frequency domain. For example, one might be interested in suppressing a certain instrument's contribution from a music signal. Such approaches are used in Computational Auditory Scene Analysis by the name of Time-Frequency masks. However, the trivial approach of just deleting the corresponding component in the time-frequency representation leads to artefacts such as ''ghost-tones'', i.e. a modified version of the deleted component is still audible, or to the perception of unnatural sound coloration comparable to a ''phasing effect''. We compare different approaches to tackle this problem: a soft-thresholding procedure is applied and compared to a method using statistical models. On the other hand, we draw conclusions from considering the optimal separating mask for known signal components. In this situation, a model promoting sparsity in the representation can lead to favourable results by yielding a feasible approximation of the optimal mask. We perform informal listening experiments to evaluate the resulting removal of certain components in a complex audio-signal.

Döring, W.H.

• Application of binaural technology in an adaptive auditory speech test for children
  J. Fels, F. Coninx and W.H. Döring [I]
[A]

This contribution reports on the development of a screening procedure for the identification of hearing loss in children by using a speech test in various spatial conditions.

A study looking at the performance of the Adaptive Auditory Speech Test (AAST) under spatial locatable noise as a possible diagnostic procedure for Auditory Processing Disorders (APD) has been carried out. The AAST was developed to determine the Speech Recognition Threshold (SRT) in a fast and reliable way under quiet and noisy conditions for children of 3 to 4 years of age.

In the study the AAST is used to develop a diagnosis method with a noise made by a talk between children (Two-Talker-Noise) or talking noise (person reading a newspaper) to test the intellectual abilities, separation and localization.

The test was carried out under various spatial conditions with the signal and noise coming from different positions in the horizontal plane. The signals were generated using head-related transfer functions which are appropriate for this age group. Therefore, the test words and the talker-noise were convolved with the corresponding children HRTFs. The scenarios were presented with the help of headphones.

Significant differences were detected between the different spatial situations. Results and experiences from current screening programs will be presented.

Doesburg, C.

• History and development of Pipe Organs in The Netherlands
  C. Doesburg [I]
[A]

An introduction will be given about the history and development of Pipe Organs in The Netherlands and their place in the Organ History of Europe.

Starting with the oldest Church Organ installed in the 15th Century, we talk about Organs in The Golden Age, influence of the German Baroque in the 18th Century, influence from the Romantic from Germany and France in the 19th Century, influence of the Modern Times from the United States concerning electric action and sound especially concerning Concert- and Theatre Organs, influence from The Organ Reform from the Elzas and the Organ Movement from North Germany. State of the Art in the Netherlands about Organ Building in this 21th Century.

Some essential information will be given about the changes in sound and technical solutions in Organ Building. Some examples with pictures and sound of the most remarkable Organs in the Netherlands will be presented.

Domitrovic, H.

• Influence of Short Term Noise on Concentration and Human Performance
  M. Horvat, K. Jambrosic and H. Domitrovic [C]
[A]

It is well known that short-term exposure to noise causes fatigue, insomnia, headache, loss of concentration and the occurrence of negative emotions. In order to examine certain aspects of this problem, a subjective testing has been made with the emphasis on the ability of human beings to maintain concentration and to perform well in a noisy environment. The test itself included the evaluation of the subjects' psychophysical state before and after completing the required task in order to examine the changes in psychophysical state caused by exposure to noise. In the central part of the test the subjects were asked to solve a simple task that did not require any specific knowledge, only a certain amount of concentration. Simultaneously, the subjects were exposed to four types of artificially generated noise. The type of noise varied from group to group and the loudness of the signal varied from one test séance to the next. The results of the tests reveal that the subjects were able to maintain their concentration and a certain level of performance when exposed to short-term noise, although they experienced significant changes in their psychophysical state.

• Evaluation Of Acoustical Suitability Of Lecture Rooms With Respect To The Quality Of The Sound Reinforcement System And The Level Of Background Noise
  M. Horvat, H. Domitrovic and S. Grubeša [C]
[A]

In order to evaluate acoustical suitability of several lecture rooms located on the Faculty of EE and Computing in Zagreb, Croatia, a questionnaire was given to the students who take their classes in those particular rooms. The first part of the investigation is focused on the quality of sound reinforcement system located in two largest lecture halls, to be replaced in the near future with a new one of significantly better quality. The goal is to examine whether this change is reflected in the grades of acoustical suitability given by the students. The second part of the examination deals with two smaller lecture rooms of identical size and acoustic treatment, but with significantly different levels of background noise due to positions of these rooms relative to the nearby street as the source of traffic noise. The overall level and spectral content of noise are measured in these lecture rooms. In addition, speech intelligibility is examined in all four rooms. The results of these measurements are then interpreted with respect to the subjective grades given by the students.

Dommes, E.

• Investigation on Auditory Cortex Activation by Low-Frequency Sound and Infrasound Using Magnetic Resonance Imaging: Stimulus Generation and Control, and Noise Assessment
  J. Hensel, E. Dommes and G. Scholz [C]
[A]

Following the demonstration of the impact of infrasound in the cochlea [Hensel et. al. 2007, Hear. Res. 233, 67-76], in the present study the stimulation of the auditory cortex down to 12 Hz was investigated by means of functional magnetic resonance imaging (MRI).

The neuroradiological evidence of cortical excitation will be reported elsewhere. In this report, details of stimulus application and noise assessment are described.

In order to exclude the other sensory paths, and to avoid the difficulties of free-field stimulation, the stimuli were delivered directly into the ear canal.

This would ideally have required a metal-free earphone, which was not available. 12 meters of silicon tube and an ear-plug were used instead. To control the influence of standing waves, the sound pressure at the ear was measured in-situ by a metal-free optical microphone, embedded at the interface of tube and ear-plug. At very low frequencies, where high sound pressures were needed even near the hearing threshold, it turned out to be a demanding task to keep the harmonics small.

The microphone position near the ear canal entrance allowed to measure the MRI noise entering the system up to that point. Spectral analysis showed that this noise already contained some infrasound.

Domont, X.

• Combining Auditory Inspirations and Hierarchical Feature Extraction for Robust Speech Recognition
  M. Heckmann, X. Domont, F. Joublin and C. Goerick [C]
[A]

Drotleff, H.

• Methods to characterize the acoustic properties of periodic surfaces
  K. Hengst, H. Drotleff, R. Wack and M. Blau [C]
[A]

Sound striking a periodic surface is absorbed and scattered depending on the surface properties. In comparison to non-periodic surfaces, periodic surfaces cause a higher sound absorbtion, due to their nearfield: If the nearfield is large, the scattered soundfield will be small and the amount of the absorbed sound high. If the scattered sound field is large, the nearfield is becomming smaller and the periodic surface is less absorbing because more sound energy is reflected. The mechanisms of sound absorption and acoustic scattering by periodic surfaces consisting of bands of alternating impedance are demonstrated first by means of a numerical model. Subsequently, measuring methods are presented to determine the acoustic properties of periodic surfaces at normal, diffuse and directed sound incidence. These measuring methods are assessed with regard to accuracy, informational value and practicability and compared to the theoretically calculated values. It is shown that it is possible (with varying accuracy) to determine the structural absorption at normal, diffuse and directional incidence. The geometrical reflection can be separated from the scattered field at diffuse and directional sound incidence. None of the measuring methods presented allows the separate examination of the near and farfield. The measurements essentially confirm the calculated values.

Druyvesteyn, E.

• Near field sound source signal separation using a single acoustic particle velocity vector sensor
  E. Druyvesteyn, J. Wind and M. Ligtenberg [I]
[A]

A method is presented to determine the contributions of incoherent sound sources based on measurements from a single acoustical vector sensor (vx,vy,vz). Well-known signal processing techniques such as Principal Component Analysis (PCA) can separate multiple incoherent contributions. In our method, however, we use the property that the acoustic particle velocity is a vector. It is an extension of a previously reported method [1], where several acoustical particle velocity sensors were necessary to a method where only one single acoustic particle velocity sensor is required. Source separation using a single sensor is achieved by applying the procedure in [1] to multiple frequency bands. When at one measuring point the distances to the various sources are different the frequency characteristics of the various sources in the near field will be different (for spherical waves this is the extra term (1+1/jkr), k wavenumber) and the procedure described in reference [1] is applied for different frequency bands. Experiments, using a three dimensional Microflown probe, and a simple model (with point sources) confirm that this method works well. W.F. Druyvesteyn, R. Raangs, Acta Acustica 91,932-935, 2005

Düster, A.

• Optimization of lightweight floors in the low frequency range with a FEM based prediction model
  A. Rabold, A. Düster, J. Hessinger and E. Rank [I]
[A]

The impact noise transmission at low frequencies is a well known problem of lightweight floors, which is treated in many publications. A satisfying solution, considering the different construction principles of lightweight floors, could not be found so far. To overcome this problem a FEM based prediction model for the optimization of the floor construction and the improvement of the impact sound insulation has been developed and applied in a current research project at the TU München. The overall approach of the prediction model consists of the three-dimensional modelling of the structure and the excitation source (standard tapping machine), the subsequent modal- and spectral analysis and the computation of radiated sound from the ceiling. The validation of the prediction model has been carried out by comparing the evaluated impact sound pressure levels with results from measurements on 25 different floor constructions. In the next step the prediction model was used for the improvement of established lightweight timber floors. Finally these constructions were tested in a laboratory test stand according to ISO 140-6 at ift Rosenheim centre for acoustics and the ibp Stuttgart. This contribution shows the results of the computations and the construction rules developed for optimized lightweight floors.

Dybek, M.

• Squeal Noise - Long Time Examination of Stochastically Occurence on Tramways passing Curves
  P. Fürst and M. Dybek [C]
[A]

In the last years exactly parametrically examination on squeal noise has demonstrated, that many factors of influence are existing. In this paper we will demonstrate, that also stochasti-cally occurrence has an influence of the sound emission of a track curve. Observations in a track curve during one year show, that the frequency of noise emission under same condi-tions don't achieve exactly comparable results. So we say, that there is only a limited prob-ability depending on the different parameters that squeal noise is appearing or not. Other-wise we found, that the curve is squealing, but there is no typical situation for the appearance of squeal noise. We will present results within variable trams and weather parameters on the same curve. At the end we will compare our observations at another kind of tram curves.

Ebersold, M.

• Effect of airborne sound on installation noise - Part 1: Basic investigations
  M. Ebersold, L. Weber, S. Öhler and M. Blau [C]
[A]

The noise excitation of buildings by service equipment takes place by both structure-borne and airborne sound. Mostly sound transmission is determined by the structure-borne portion. For some sources, however, the contribution of airborne sound can't be neglected. A typical example is the EMPA-hammer, defined in the swiss standard SIA 181 as a standardized source for the simulation of user noise. If airborne sound contributes substantially, conventional measures for noise reduction based on elastic isolation between source and building won't work properly. In order to develop efficient measures for that kind of excitation, structure-borne and airborne sound must be separated and investigated in context with each other.

The separation of structure-borne and airborne sound was performed by means of a simplified experimental set-up consisting of two parallel plates acoustically connected by an adjustable structure- borne sound bridge and by the air gap between the plates. Since the receiving plate was bedded on an elastic support the transmitted sound power could be determined directly from the mean velocity level on the surface of the plate. For investigation of sound transport the experimental conditions such as distance of the plates or damping and division of the air gap were varied systematically.

Eccardt, P.-C.

• Evaluation of Spatial Approaches for Improvement of Patient Speech Acquisition at Magnetic Resonance Tomography Systems
  M. Friedrich, K. Porzelt, M. Bechtold and P.-C. Eccardt [C]
[A]

During operation of magnetic resonance (MR) tomography systems, switching noise is produced with high sound pressure levels. This noise is a disturbance for patient-to-operator communication, as well as for examinations that require synchronous recording of patient speech (fMRI). Signal processing algorithms can be applied for the removal of MR and other noise sources from patient speech. However, the disturbed speech signal should be obtained with the highest possible input SNR, which is complicated by limitations in practical application. On a typical MR installation, the room acoustics and sound field characteristics were examined, with the objective of optimising the microphone solution with regard to patient speech pick-up. Applying coherence based measures, the noise field is characterised on and around the MR system. General conclusions are drawn about possibilities to employ spatial information for optimising the speech level with regard to interfering noise sources, such as using microphones with directional characteristics, or multiple microphones and source separation algorithms. The choice of suitable microphones and their positioning on a MR installation are discussed. In addition, experiment results are presented for SNR improvement on a real MR installation, using microphones with different directional characteristics on suitable position.

Eckers, C.

• Investigation of voice production in death metal singers
  C. Eckers, D. Hütz, M. Kob, P. Murphy, D. Houben and B. Lehnert [I]
[A]

Singers in ''metal'' style often use a unique form of voice production called ''growling''. The underlying mechanisms of voice production as well as possible potentially harmful influences on the singer's voice have not yet been intensively studied. Supra glottal laryngeal constriction appears to play a major role in metal singing. This constriction results in difficulties to apply standard laryngeal endoscopy for visualization of vocal fold function during this particular phonation style. In this study we present an examination method using a questionnaire including the voice handicap index as well as a combined simultaneous endoscopy, electroglottography and acoustic analysis to investigate in particular the production of death metal singing in seven subjects. None of the subjects complained about persistent voice problems and none had signs of secondary organic voice disorders like vocal fold nodules or inflammation. However, most showed minor irregularities on stroboscopic examination. Further findings reveal that at least two forms of supra glottal laryngeal constriction are used to obtain the typical death metal sound which differ in use of vocal folds and ventricular folds as well as in supra-glottal adduction effort. The study shall help to answer the question whether death metal singing can be performed such that the risk of voice damage can be reduced or avoided.

Eder, A.

• Comparison of Coding Techniques for Photoacoustic Coded Excitation
  M. Mienkina, A. Eder, C.-S. Friedrich, N. Gerhardt, M.R. Hofmann and G. Schmitz [C]
[A]

In photoacoustics (PA) ultrasound is generated by laser irradiation of tissue. Usually, PA signals are generated by Nd:YAG lasers but it would be attractive to use less expensive laser diodes instead. Although laser diodes exhibit low pulse energy, the pulse repetition frequency (PRF) of pulsed laser diodes is much higher than the PRF of Nd:YAG lasers thus averaging can be used to increase the SNR. The applied PRF for averaging is limited by the acoustical time-of-flight. To further increase the SNR we propose to use coded excitation for PA imaging. In optical-time-domain-reflectometry, which is similar to PA imaging, Simplex Codes (SC) are widely used for coding. In ultrasound imaging Golay Codes (GC) are successfully employed. Therefore, we compare the performance of SC and GC for PA coded excitation (PACE) by a simulation study. For both codes, the coding gain normalized to equivalent averaging procedures is computed as a function of code length and PRF of the laser diode. Assuming a time-of-flight of 60 s and e.g. a PRF of 250 kHz the coding gain of GC is 1.6 dB higher than the coding gain of SC. Consequently, GC is more suitable for PACE than SC.

Ehrig, L.

• Electro Active Materials for Flat Panel Transducers
  L. Ehrig, D. Beer and S. Brix [C]
[A]

Electroactive materials are a class of materials that change their geometry due to the presence of an electrical field. A well known effect that occurs in crystals, polarized ceramics and plastic materials is piezoelectricity. Since there is no center of symmetry and a preferred direction in such materials respectively an electrical voltage applied on the materials surface yields a deformation. The same macroscopical behaviour can be observed in relatively new materials called ferroelectrets. Voids in these cellular polymer foams are internally charged and form giant dipoles. The dipoles are deformed by an electrical field likewise in piezoelectric materials. Electroacoustic transducers based on electroactive materials may have a unique feature: driver and membrane are one part. Thus it is possible to build extremely simple, compact and flat transducers. Using the material by means of a thickness oscillator even the rearward air spring, causing a major part of problems for thin transducer constructions, can be avoided. In this work several approaches for transducers using electroactive materials are presented.

Eichenlaub, C.

• Standstill Measurements on Open Track
  M. Tinter, N. Meunier and C. Eichenlaub [C]
[A]

Since the entry into force of the Technical Specification for Interoperability of Conventional Rolling Stock - Subsystem Noise (short TSI Noise), the homologation process of new vehicles includes a test of their noise emission. TSI Noise specifies in particular limits for stationary noise, starting noise, pass by noise and driver's cab noise. The conformity of rolling stock has to be assessed within type testing.

Modern vehicles developed under consideration of the TSI requirements may present a quite low noise emission, which involves that the measuring site has to be chosen carefully, since the exterior noise measurements are performed at a distance of 7.5m from the track axis and the background noise has to be at least 10dB lower than the vehicle noise. The lecture demonstrates - using the example of the TSI certification of the electric trainset Coradia Continental - the advantages of carrying out standstill measurements at the test site for acceleration and pass by tests, resulting in benefits for precision, time and costs of the measurements.

Eichler, M.

• Superdirective Beamforming Using an Extended Modal Subspace Decomposition
  M. Eichler and A. Lacroix [C]
[A]

Beamforming techniques allow to control the characteristic of a microphone array in order to achieve a desired directivity. One of the most general formulations is the filter-and-sum beamformer which has readily been generalized by the concept of modal subspace decomposition. This approach finds optimum FIR filter coefficients for each sensor by solving an eigenvalue problem and projecting the desired beam pattern to the set of eigen-beam patterns found. However, the resulting beamformers fail to be directive at very low frequencies where the wavelength is large in comparison to the overall size of the sensor array. In this contribution, we propose an extension to the mentioned concept of modal subspace decomposition which helps to overcome this problem by making the algorithm superdirective.

Eilers, R.

• Evalution of a Nonlinear Auditory Filterbank with Instantaneous Frequency Control
  R. Eilers, S.D. Ewert, J.L. Verhey and V. Hohmann [C]
[A]

Psychoacoustical studies suggest that the increase in suppression as a function of suppressor level is greater for a suppressor spectrally below than above the signal frequency. This pattern was investigated by measuring suppression for a tonal signal with two different psychoacoustical methods in the same subjects. The first method was the pulsation threshold technique that was used in a two-tone suppression paradigm with suppressor frequencies below and above the signal frequency. The second method was a forward-masked paradigm where the masker was a bandpass noise and the parameter the masker bandwidths. The noise was either centred at the signal frequency or was asymmetric with a larger portion of the spectrum on either the low or the high frequency side. The results of the two experiments were consistent with the hypothesis of different suppression growth functions for the different frequency regions. The measured data was modelled with a nonlinear filter approach. The model uses a modified dual-resonance filter by including a control of the nonlinear filter gain by the sub-band instantaneous frequency. It is investigated, if the model is able to quantitatively predict suppression and the effect of suppressor frequency.

Elmahdy, M.

• Survey on common Arabic Language forms from a speech recognition point of view
  M. Elmahdy, R. Gruhn, W. Minker and S. Abdennadher [C]
[A]

Arabic language is the largest still living Semitic language based on the number of speakers which exceeds 250 million first language speakers. Arabic language has two main forms: Standard Arabic and Dialectal Arabic. Standard Arabic includes ''Classical Arabic'' and ''Modern Standard Arabic (MSA)'' while dialectal Arabic includes all forms of currently spoken Arabic in day life and it vary among countries and deviate from standard Arabic to some extent. While there are many forms of Arabic, there still many common characteristics on the acoustic level and the language level. In this paper we investigate the main differences between Arabic forms from a speech recognition point of view to focus on similar and different acoustic properties among Arabic forms as well as the language properties. For standard Arabic we have chosen both classical and MSA forms and for dialectal Arabic we have chosen Egyptian colloquial Arabic as an example since it is the most popular dialect among Arabic speakers. The main purpose of this paper is to summarize all characteristics of Arabic and previous research results in Arabic speech recognition from a speech recognition point of view in one paper so researchers in this field can use it as one reference.

Embrechts, J.-J.

• An interactive and real-time based auralization system for room acoustics, implementing directional impulse responses and multiple audio reproduction modules for spatialization (the AURALIAS project)
  L. Bos and J.-J. Embrechts [I]
[A]

The AURALIAS project aims at developing an interactive and real-time auralization system for room acoustics. The user is immersed in a virtual sound field, looking at the virtual room and interacting with it through an appropriate interface. The real-time auralization system first imports pre-computed Directional Room Impulse Responses (DRIRs) according to the current situation in the virtual world. Those DRIRs, previously obtained by a ray tracing software, are then convolved with the anechoic signal using a frequency block segmented convolution. Filtered signals are then distributed either to a VBAP or an HRTF sound field reproduction module. The first prototype allows the auralization of multiple isotropic sound sources and the displacement of the virtual listener including head's orientation. Some other features like changing ''on the fly'' the anechoic signals and the current room are also implemented. Further developments will offer the support for moving sources, mobile sweet spot, user's own speech auralization and both real-time acoustical and geometrical modifications of the virtual room. Besides, the software's architecture should be compatible with several room acoustics algorithms, convolution modules or sound field reproduction systems.

Emich, M.

• Earprotection Project Military Music
  M. Logar-Holzer, M. Emich, W. Wahler, C. Gruber, A. Graf-Langheinz and S. Spendelhofer [C]
[A]

The past experiences in the Austrian military music band with the adaption of a customized earprotectors have so far been unsatisfactory. Different aspects felt negatively by the musicians (subjectively) lay particularly in the range between the false interpretation of the sound image and/or the difficulty in the assessment of the dynamics, volume and Intonation of the sounds/tones.This observation study was conducted in three representative Austrian Military Music Bands. The results were gained out of a total of 146 persons by means of a standardised questionnaire in various subjects with regard to personal handling and the use, deployment and acceptance of the personal protection of the ears.The daily length of music performance was documented during the period of 8 weeks. Parallelly volume decibel measurements in different practical-relevant sound situations of the musicians (orchestra, single exercises,...) were gathered.

Engel, G.

• Creating Temporary Venues for High Demanding Classical Concerts using a Room Enhancement System
  G. Engel and M. Blome [I]
[A]

In the field of classical concerts an increasing tendency of arranging concerts in unusual and sometimes unsuitable venues can be observed. There are various possibilities to acoustically improve these venues, but most of them turn out to be impracticable due to the limited preparation time and an unacceptable appearance of the acoustical measures. In many cases the only solution can be provided by the use of room enhancement systems, which counteract the acoustic deficiencies with electro acoustical means. The lecture presents case studies of two festivals in St. Moritz and in Abu Dhabi, where ball rooms, dining halls or auditoria of first class hotels are used as venues for top level orchestras and soloists. Beside the benefit for the musicians and concert audience the use of enhancement systems helps acousticians to improve the knowledge about the correlation between measurement results and the audible impression.

Engelputzeder, N.

• On the Structure of the Phase around the Zeros of the Short-Time Fourier Transform
  F. Jaillet, P. Balazs, M. Dörfler and N. Engelputzeder [C]
[A]

The short-time Fourier transform is a time-frequency representation widely used in audio signal processing. At first sight, the phase of the transform may appear to be difficult to interpret, a fact that led to the development of many processing schemes in audio processing relying on the modulus of the transform only. However, it has been shown, that the phase information can be successfully exploited for improved analysis and processing. This is illustrated for example by the reassignment method or the phase vocoder algorithm.

In order to better exploit the phase information in audio processing, a better understanding of its structure is desirable. This study describes a particular behavior of the phase which can be observed when looking at the phase derivative over time or frequency around the points where the transform equals zero. This recurrent behavior is illustrated using numerical simulations on synthesized and real audio signals. The performed simulations show the influence of the parameters of the transform on the observed structures. Some consequences of this phase structure for audio processing are also presented.

Enghardt, L.

• Experimental Investigation of Mode Scattering at C-Shaped Bypass-Duct Sections of Aero-Engines
  R. Bauers, C. Stöhr, U. Tapken and L. Enghardt [C]
[A]

Modern aero-engines feature high bypass ratios to reduce fuel consumption and jet noise. As a consequence, fan noise has become one of the most annoying noise sources. Over a wide frequency range, fan tones excited by the rotor-stator interac-tion propagate in form of higher order modes through the bypass duct, which is in mosta cases radially divided by a pylon. The pylon is often paired with a supporting strut at perpendicular circumferential position. Both bifurcations form two independent c-shaped ducts, respectively, the boundary conditions of which substantially differ from the regular case of a cylindrical bypass-duct with hub. At DLR in Berlin, a di-ploma thesis was undertaken to experimentally investigate the impact of the py-lon/strut combination on the propagation of modes in a bypass-duct. The experimen-tal setup comprised different geometrical bypass duct configurations. Modes were excited by means of a loudspeaker ring in a cylindrical duct section, which bifurcated downstream to a c-shaped duct. Measurements of the modal field were accom-plished in both the cylindrical and the c-shaped duct sections. The paper will present the experimental results with emphasize on mode scattering and transmission loss at the c-shaped duct section.

• Experimental and Numerical Characterisation of a Non-Locally Reacting Liner
  S. Busse, C. Richter, C. Kückens, U. Müller, L. Enghardt and F. Thiele [C]
[A]

A new type of aero-engine liner was investigated. It is manufactured by combining a cell structure of assembled metal strips with a perforated face sheet and a solid backing plate by using a point welding process. The structure is completed with an additional face sheet layer of meshed metal baffle. Since neighbouring cells are not air tight the damper is non-locally reacting and inter-cellular communication is possible. The acoustic properties in terms of the insertion loss were experimentally analysed by integrating a liner sample in a side wall of a grazing flow acoustic test facility. Measurements with microphones and plane wave excitation were conducted in order to calculate the reflection and transmission coefficients. The influence of parameters like the flow Mach number and the liner length on the damping behaviour were investigated. By using a numerical method which incorporates the Extended Helmholtz Resonator model (EHR) the impedance was educed from the measurement values utilising an optimisation algorithm. It is shown that the method provides reasonable results even for this special type of liner.

Epp, B.

• Comodulation masking release with time-varying spectra
  H. Heidemann, J.L. Verhey and B. Epp [C]
[A]

The detectability of a tone masked by noise bands can be enhanced if noise bands with common level fluctuations are used i.e comodulated bands. This effect is commonly referred to as ''Comodulation Masking Release''(CMR). Since comodulation is a property of many natural sounds like e.g. speech, it is generally assumed, that this sensitivity is relevant in the differentiation of natural sound sources. So far, CMR has been investigated using stationary spectra. However, speech and other natural sounds are also characterized by spectral and spectro-temporal modulations. In order to test the robustness of CMR in the presence of these properties a modified CMR paradigm is used where the stimulus spectrum changes over time. In the first experiment sweeps rather than tones are used for the signal and masker band center frequencies. In the second experiment simultaneous manipulation along the spectral and the temporal axes is used in order to test the sensitivity to spectro-temporal patterns. The results of the first experiment show that CMR is still observed if the CMR is calculated relative to a condition with the same spectral properties but uncorrelated bands. The results of the second experiment indicate that changing from temporal modulation to spectro-temporal modulations abolishes CMR.

• Modulation perception and threshold fine structure: Simulations with a nonlinear cochlea model
  B. Epp, J.L. Verhey, S.J. Heise and M. Mauermann [C]
[A]

Many natural sounds show temporal level fluctuations, i.e. amplitude modulations. Thus, a detailed knowledge about mechanisms underlying modulation perception is essential for the understanding of auditory signal processing. Most psychoacoustical studies investigated amplitude modulation processing for carrier levels well above threshold in quiet. Recently, Heise et al. (presented at DAGA2008) showed that the quasi-periodic fine structure in the threshold in quiet commonly observed in normal-hearing listeners affects the modulation detection of sinusoidally modulated tones with low carrier levels. The influence of threshold fine structure on modulation detection thresholds is not predicted by current models of modulation perception. This is essentially due to the simplification of the peripheral stages in these models. In contrast to the previous studies, the present study uses a one-dimensional active and nonlinear model of the cochlea which simulates a fine structure in the threshold in quiet. The representation of amplitude modulation in regions of fine structure at the level of the basilar membrane is analyzed and compared to the performance of subjects in the modulation detection task. The simulations can further enhance our understanding of the relative contribution of temporal and spectral mechanisms underlying modulation perception.

Ertl, M.

• Vibroacoustics and sound emission characteristics of thin-walled, oil-immersed transformer vessels
  M. Ertl and H. Landes [C]
[A]

The fluid-structure-interactions, vibro-acoustics and sound emission characteristics of power transformers are investigated by numerical methods.

Oil-immersed power transformers consists of thin-walled, rib-stiffened tank structures. This geometric setup and the three-phase acoustical excitation results in a complex spatial sound emission characterisics in the low frequency range of 100Hz to400Hz. Sound emission lobes are observed in measurement and simulation.

The effect of additional damping, mass load and stiffening to the vibroacoustic behaviour and overall sound level is investigated. Further, the effect of cavity modes inside the vessel to the sound emission is quantified. The influence of these parameter to the structure-borne sound level, the transmission ratio and radiation loss factor of the vessel construction is analysed.

Ewert, R.

• Broadband Simulation Of Flow-Induced Noise Generation in Air Conditioning Ducts
  C. Spehr, R. Ewert, T. Ahlefeldt and J. Delfs [C]
[A]

Air conditioning systems in cabins are possible noise sources influencing the passenger comfort in civil aircrafts. The current study continues an initiative to predict the generated noise on components of air distribution systems. The approach presented here is based on the numerical inexpensive RANS-simulation. Using a perturbation technique the turbulent flow fluctuations are represented by artificial turbulences injected into the mean flow field. In the first part of the study the time dependent perturbation was induced by injecting individual artificial vortices which allow qualitative comparison of the sound generation of different arrangements on the basis of the overall sound power level. In the current work a broadband turbulence is induced in the source region. The turbulence is thereby generated by filtering white noise to reproduce the autocorrelations and integral length-scales provided by a RANS-simulation. The interaction of the turbulence with AC-components generates pressure fluctuations with broadband character which are compared with corresponding measurements. Arrangements of two consecutive orifices were investigated as typical components in the air conditioning system. The sound generation on the downstream orifices depends therefore not only on its shape, but also on the flow disturbance, which in turn depends on the distance to the upstream orifice.

Ewert, S.

• Towards timbre-invariant audio features for harmony-based music
  S. Ewert, M. Müller and M. Clausen [I]
[A]

In the context of automatic music processing, chroma features have turned out to be a powerful midlevel representation with a wide range of applications such as content-based retrieval, music synchronization, and chord transcription. An important step of the feature calculation is the grouping of spectral energy components that belong to the same pitch class or chroma of the equal tempered scale. Here, the octave identification introduces a high degree of invariance to changes in timbre and instrumentation. In this paper, we introduce a strategy to further increase this invariance by combining the concept of chroma features with the well-known concept of mel-frequency cepstral coefficients (MFCCs). The lower MFCCs are known to capture information on timbre. Therefore, to enhance robustness to timbral changes, we discard this information by only keeping the upper coefficients. Furthermore, using a pitch scale instead of a mel scale allows us to group the remaining coefficients into twelve chroma bins. The resulting audio features, as indicated by our systematic experiments, have indeed gained a significant boost towards timbre invariance without a degradation of discriminative power.

• Combination of Onset-Features with Applications to High-Resolution Music Synchronization
  P. Grosche, M. Müller and S. Ewert [I]
[A]

Many different methods for the detection of note onsets in music recordings have been proposed and applied to tasks such as music transcription, beat tracking, tempo estimation, and music synchronization. Most of the proposed onset detectors rely on the fact that note onsets often go along with a sudden increase of the signal's energy, which particularly holds for instruments such as piano, guitar, or percussive instruments. Much more difficult is the detection of onsets in the case of more fluent note transitions, which is often the case for classical music dominated by string instruments. Here, one has to develop methods that allow for detecting smooth temporal and spectral changes in the audio signal. In this paper, we present various methods for onset detection including energy-based and phase-based techniques, which account for different properties of the audio signal. In particular, we present a recent audio feature that combines the high temporal accuracy of onset features with the robustness of chroma features. Finally, we show how a combination of various onset features can be used to significantly improve the temporal resolution in music synchronization tasks.

Ewert, S.D.

• Modeling the masking of tones by Schroeder-phase harmonic tone complexes
  E. Rasumow, M. Hansen and S.D. Ewert [C]
[A]

The amount of masking from harmonic tone complexes (HTC) with Schroeder phase on a sinusoidal signal varies up to over 30 dB when the phase curvature of the masker is altered. This masking differences can be explained by the dispersion on the basilar membrane (BM) which, in interaction with the phase of the HTC components, influences the temporal ''peakiness'' of the BM excitation. When modeling the phase curvature dependent amount of masking the use of a linear BM-filter with optimized phase-curvature and magnitude response is insufficient in terms of the resulting dynamic range of the simulated masking thresholds. One important factor might be the inclusion of peripheral compression into the model. However, current studies using the dual-resonance nonlinear filter (DRNL) [Meddis et al., J. Acous. Soc. Am, 109(6):2852-2861,2001] have shown no improvement of the simulations. This study deals with the refinement of the response characteristic of the DRNL filter. The effect of modifications of the DRNL filter was evaluated in the framework of the CASP-model [Jepsen et al., J. Acoust. Soc. Am., 124(1):422-438, 2008]. Furthermore, the dependence of the model-parameters concerning the dynamic range of the internal representation was investigated.

• Subband instantaneous-frequency analysis to determine masking with high temporal resolution for use in audio codecs
  N. Koppaetzky, S.D. Ewert, B. Kollmeier and V. Hohmann [C]
[A]

Algorithms for perceptual audio coding use psychoacoustical masking models to compress audio data with minimal impact on the perceived audio quality. Latest psychoacoustical research emphasizes the excellent spectro-temporal resolution of the auditory system. Most common audio codecs, however, apply classical spectral masking models which generally reach the required high spectral resolution at the expense of an insufficient temporal resolution. An application of a current psychoacoustical model which simulates the high spectro-temporal resolution of the auditory system, might offer an increase of the coding efficiency or of the perceptual coding quality, respectively. To analyze the general effect of an enhanced time resolution, a current nonlinear auditory filterbank model with instantaneaous-frequency conrol was tested in this constribution. The applicability of the model for the classification of the tonality of the signal and for forming an estimate of the spectro-temporal masking pattern was examined. The masking model as well as the resulting perceptual coding quality which can be achieved by a combination of the model with a simple perceptual coder are presented and discussed.

• Evalution of a Nonlinear Auditory Filterbank with Instantaneous Frequency Control
  R. Eilers, S.D. Ewert, J.L. Verhey and V. Hohmann [C]
[A]

Psychoacoustical studies suggest that the increase in suppression as a function of suppressor level is greater for a suppressor spectrally below than above the signal frequency. This pattern was investigated by measuring suppression for a tonal signal with two different psychoacoustical methods in the same subjects. The first method was the pulsation threshold technique that was used in a two-tone suppression paradigm with suppressor frequencies below and above the signal frequency. The second method was a forward-masked paradigm where the masker was a bandpass noise and the parameter the masker bandwidths. The noise was either centred at the signal frequency or was asymmetric with a larger portion of the spectrum on either the low or the high frequency side. The results of the two experiments were consistent with the hypothesis of different suppression growth functions for the different frequency regions. The measured data was modelled with a nonlinear filter approach. The model uses a modified dual-resonance filter by including a control of the nonlinear filter gain by the sub-band instantaneous frequency. It is investigated, if the model is able to quantitatively predict suppression and the effect of suppressor frequency.

• Classification of Reverberant Acoustic Situations
  J. Schröder, T. Rohdenburg, V. Hohmann and S.D. Ewert [C]
[A]

In daily communication, speech intelligibility depends on the acoustic surrounding or acoustic situation. Particularly for hearing impaired persons, speech understanding is often problematic if speech is distorted by (room) reverb, noise or competing talkers. Acoustic situations are characterized by different dominating types of distortion. Hearing aids might provide appropriate algorithms to enhance speech intelligibility in the different acoustic situations. A robust and fast automatic classification of the acoustic situation should therefore select the appropriate hearing aid algorithm without requiring an action of the hearing aid wearer. This study is concerned with the automatic estimation of the reverberation time (T60) in natural situations and with unknown excitation signal. Acoustic test situations were generated by convolving speech signals with artificial and real room impulse responses with T60-times ranging from 0.1 to 4 s. Three different features derived from the cepstral mean, the autocorrelation function and from the distribution of modulation energy were used to train and classify multiple groups of different reverb times with Gaussian mixture models (GMM). Additionally, information about the unknown excitation signal was gained by estimation of its (audio-) spectral content and used to improve the classification. The classifiers, features and results are presented and discussed.

• Measurement, modelling and compensation of nonlinearities in hearing aid receivers
  T. Schaer, S.D. Ewert, J. Anemüller and B. Kollmeier [C]
[A]

Hearing aid receivers operate over a large dynamic range and reach output levels of up to 130 dB. Electromagnetic receivers are commonly used in hearing aids since they offer a very high efficiency. Receivers are typically characterized by their linear transfer function. These linear distortions are easily measured with standard methods and if compensation is applied it is often limited to the magnitude transfer function. The current study addresses a more complete characterization of the hearing aid receiver by measurement of the nonlinear transfer function. Nonlinear distortions of the receiver are output level dependent and might negatively influence measures like speech intelligibility of an aided hearing impaired person, particularly at high levels. Moreover, nonlinear distortions are problematic in the context of future ''high-fidelity'' hearing aids and compensation methods are of interest. Here, a fast and efficient sinesweep method proposed by Farina was employed to estimate the linear and nonlinear transfer function. The measurement was repeated at different output levels. From the data, nonlinear input-output functions can be constructed or alternatively, the diagonal elements of a Volterra-series based description of the nonlinear transfer function can be derived. A nonlinear receiver model and compensation methods are suggested and tested.

• The Effect of Envelope Waveform on Lateralization
  M. Klein-Hennig, M. Dietz, S.D. Ewert and V. Hohmann [C]
[A]

Interaural timing disparities are important cues for lateralization in the human auditory system. The auditory system is sensitive to timing disparities in the fine-structure and the envelope of sounds. At high frequencies, only envelope disparities can be exploited due to the lack of phase-locking to the fine-structure. In previously published data [e.g. Bernstein and Trahiotis in ''Hearing - from basic research to applications'', Springer 2007] parameters like the modulation frequency, or the power of a modulating sinusoid have been altered in order to investigate the role of the envelope waveform on lateralization. However, these ''classical'' parameters simultaneously change several ''secondary'' envelope parameters like attack and decay times, as well as pause and hold durations, which might conceal the fundamental cues exploited by the auditory system. In this study, psychophysical measurements were conducted with customized envelope waveforms in order to investigate the effect of isolated secondary parameters on lateralization. For high-frequency tones centred at 4 kHz with systematic envelope modifications, the just noticeable differences of interaural envelope time differences were measured. The results indicate that attack times and pause durations prior to the attack are the most important envelope features. The results are compared to predictions by a binaural auditory model.

• Azimuthal Localization of Concurrent Speakers Employing Interaural Coherence
  M. Dietz, S.D. Ewert and V. Hohmann [I]
[A]

A computational model for the localization of concurrent speakers is presented that extends the binaural model presented by Dietz et al. (Fortschritte der Akustik - DAGA 2007, 107-108). The processing is motivated by mammalian physiology and human psychoacoustic data. A wide range of interaural parameters is exploited, namely the interaural level difference (ILD), the interaural phase difference (IPD) of the fine structure, the IPD of the fundamental modulation frequency and the interaural coherence (IC). The parameters are determined for each auditory filter which allows a simultaneous estimation of several sound directions. By focusing on instants with high interaural coherence the robustness of the estimation could be greatly improved (Faller and Merimaa, 2004, J. Acoust. Soc. Am., vol. 116, 3075-3089). For two well-separated speakers in quiet the estimation error is generally less or equal 10. In noisy environments a single speaker can be localized with an error of less than 20, even for a signal-to-noise ratio of 0 dB.

Eyben, F.

• Blind Enhancement of the Rhythmic and Harmonic Sections by NMF: Does it help?
  B. Schuller, A. Lehmann, A. Weninger, F. Eyben and G. Rigoll [I]
[A]

Non-Negative Matrix Factorization is well known to lead to considerable successes in the blind separation of drums and melodic parts of music recordings. Such splitting may well serve as enhancement when it comes to typical Music Information Retrieval tasks as automatic key labelling or tempo detection. In this respect we introduce the combination of an NMF based blind music separation into several isolated audio tracks in combination with Support Vector classification to assign each obtained track to be either rhythmic or melodic. Thereby optimal parameterization and performances are discussed. Next, stereophonic information is further used to eliminate the key melody for chord labelling. We then analyze the potential for the named tasks by an extensive number of experiments carried out on the MTV Europe Most Wanted of the 80ies and 90ies in MP3 format.

Eysholdt, U.

• Numerical Modelling of Vocal Fold Dynamics by a 3D Multi-Mass-Model
  A. Yang, J. Lohscheller, M. Stingl, D. Voigt, U. Eysholdt and M. Döllinger [I]
[A]

The emitted human speech signal originates due to oscillations of both vocal folds. So called functional dysphonia being only observable during phonation yield voice disorders, i.e. hoarseness. In historical voice research, the endoscopic vocal folds movements were mostly simulated via 1D Two- or 2D Multi-Mass-Models. However, by using these models only vocal fold movements in lateral and longitudinal directions are taken into account.

This work presents an enhancement of these mathematic biomechanical models: a Multi-Mass-Model expanded from 2D to 3D displacements. It consists of five mass planes arranged in vertical direction. Each plane contains five longitudinal coupled mass-spring oscillators. Via the new proposed model not only the vertical movements but also the diffusion of the mucosal wave from inferior to superior can be simulated.

In order to better simulate vocal fold vibrations, the time-invariant model parameters are adjusted in a developed hybrid optimization procedure. The evaluation of optimization procedure is performed by using lateral symmetric data sets synthetically generated by the suggested model. Different glottis closure-types are also considered.

The future goal is to describe the physiological relevant characteristics of 3D vocal fold dynamics, which can be extracted via high-speed imaging in combination with a new developed laser-spot projection system.

Falch, C.

• Efficient Parametric Audio Coding for Interactive Rendering: The Upcoming ISO/MPEG Standard on Spatial Audio Object Coding (SAOC)
  L. Terentiev, C. Falch, O. Hellmuth and J. Herre [C]
[A]

Spatial Audio Object Coding (SAOC) is one of the recent standardization activities in the ISO/MPEG audio group. SAOC is a technique for bitrate efficient parametric coding and flexible, user-controllable rendering of multiple audio objects. An extremely high audio compression efficiency is reached due to transmission of only a backward compatible mono (or stereo) downmix signal and a small amount of additional side information describing the perceptual properties of the sound objects. At the SAOC decoder, all sound objects can be rendered interactively to different spatial reproduction set-ups of choice. The new technology extends the MPEG Surround standard in a natural way and inherits its advantages and spatial rendering capabilities. Both bitrate efficiency and rendering interactivity makes SAOC attractive for broad adoption in a variety of interactive applications, for example music 2.0 (personal remixing), next generation teleconferencing and broadcasting. The publication provides an overview of the SAOC reference model architecture and its basic underlying concepts. It describes modes of operation, shows typical application scenarios and reports the current status of the ISO/MPEG standardization process.

Falk, T.H.

• Quality Prediction for Synthesized Speech: Comparison of Approaches
  S. Möller and T.H. Falk [C]
[A]

Text-To-Speech (TTS) technology has reached a level of maturity which seems to be sufficient for a number of telephony applications. In order to assess TTS quality, system developers need to carry out auditory tests where participants are asked to transcribe what they have heard, or to rate certain aspects of the auditory event. To overcome the temporal and financial effort involved in auditory testing, it is desirable to estimate the quality on the basis of the synthesized speech signal alone.

Two such approaches will be compared in this talk: The first has proven successful to estimate the impact of transmission degradations on telephone networks. It generates an artificial (high-quality) reference, compares the TTS signal to this reference, and combines the result with a noisiness, clipping and robotization detector in a parametric analysis to derive an overall quality estimate. The second approach is based on a Hidden Markov Model trained on natural (male or female) speech; features derived from TTS signals are compared to the model for natural speech and quality is derived from the normalized log-likelihood between the reference models and the extracted features. The performance of both approaches will be presented and proposals for improvements will be outlined.

Fallast, K.

• Traffic Noise Annoyance on Roads and Rail (TNAR) in an experimental laboratory setup
  M. Cik, K. Fallast, R.B. Raggam, E. Marth, E. Gallasch, A. Lackner and M. Fellendorf [C]
[A]

At present, according to the current standards and specifications, the impact of road and rail traffic noise is represented by the A-weighted energy-equivalent sound level (LA,eq). This quantity does not account for the subjective annoyance of sound events as it is perceived by the affected persons.

Pass-by road and rail vehicle noise samples were binaurally recorded under predefined conditions with a dummy head measurement system and synthetically composed to defined vehicle ensembles. A total of 200 persons were selected and exposed to these defined vehicle ensembles in different sessions in a hearing laboratory. Annoyance was rated on an 11-graded interval scale and by a hand power-dynamometer which measures the intensity of the applied strength. Psychoacoustic parameters of the defined vehicle ensembles were calculated. Annoyance ratings were combined with psychoacoustic parameters by multiple regression analysis. The relevant parameters were statistically combined into the index TNAR.

The analyses revealed that loudness is the parameter describing annoyance best, significantly better than the LA,eq (P0.0001). Further significant influencing parameters are sharpness and a new designed parameter representing the frequency characteristic of traffic noise. These three parameters together (R=0.807), correlate significantly better with the subjective estimation of noise-induced annoyance than the LA,eq (R=0.672) does (P0.0001).

Fastl, H.

• Measurement of Head Related Impulse Responses for psychoacoustic research
  F. Völk, M. Straubinger, L. Roalter and H. Fastl [C]
[A]

In recent years, the importance of Head Related Impulse Responses (HRIRs) for psychoacoustic experiments and especially for studies in directional hearing has been growing, mainly caused by the availability of fast computers and the resulting possibility of very realistic dynamic binaural synthesis. A necessary prerequisite for all these methods is the correct measurement of HRIRs.

In this contribution a setup will be presented which is capable of measuring HRIRs for psychoacoustic research by two different techniques: maximum length sequences and exponential sine sweep. To reduce the measurement time, an interpolation algorithm can be used. The resulting HRIR-sets are compared by typical physical parameters as well as judged in a localization experiment. Based on these data, parameters for high quality HRIR measurements for psychoacoustic research are proposed.

• Binaural Sky – Examination of Different Array Topologies
  K. Laumann, G. Theile and H. Fastl [C]
[A]

The ''Binaural Sky'' system is known as a novel reproduction technology for binaural signals. The essential part can be characterized by the term ''virtual headphone''. Movements of a listener are compensated using dynamic WFS-based rendering of focussed sources. A first prototype was constructed with a circular speaker array mounted above the listener. Details are given in: Menzel, Wittek, Theile, Fastl - ''Binaurale Raumsynthese mittels Wellenfeldsynthese - Realisierung und Evaluierung", DAGA 2006. Several parameters influence the quality of reproduction, one of them being the array topology. For a versatile practical application of the concept it is necessary to quantify this influence. A number of experiments are presented in this paper, examining different array topologies with up to 24 speakers.

• Psychoacoustic Evaluation of Music Reproduction in Passenger Cars
  H. Fastl, M. Ammler and A. Sulowski [C]
[A]

A method to assess the quality of music reproduction in passenger cars is developed and evaluated. In essence, the subject sits in a standing car with the motor off, and is presented from a CD music of different genres via the car's audio system. The subject has to rate different aspects of the performance on a 10 step scale using a PDA. From the ratings, a net like a spider net, with different attributes like natural, squeeking, dull, reverberant, acceptable, good etc. on its rays is constructed. The goal is a net with a large area, i.e. as many as possible of the attributes should attain the maximum rating.

Examples from a pilot study are given for car audio systems of different sophistication, and aspects like gender differences or possible biases due to product relationship will be discussed.

• Comparison of Speech Intelligibility by EAS, Bimodal, Uni- and Bilateral Cochlear Implant Patients in a ’Multi-Source Noise Field’ (MSNF)
  T. Rader, U. Baumann and H. Fastl [C]
[A]

Speech perception in noise is one of the most difficult tasks for people suffering from hearing impairment. The Oldenburg Sentence Test (OLSA) is a useful tool to investigate speech intelligibility threshold in noise. In the present study, a multi-source noise field introduced by Rader et al. (DAGA2008) consisting of a four loudspeaker array with independent noise sources was combined with the OLSA. The MSNF allows to present a more realistic noise environment and shows a higher effect of binaural interaction regarding the separation of signal and noise from different directions.

Three different noise characteristics were applied in two conditions (S0N0: Signal and noise from the front; MSNF) to investigate speech reception threshold: Oldenburg noise (generated by superposed sentences of OLSA), CCITT noise, and speech simulating noise according to Fastl (1987).

Four different groups of cochlear implant patients separated into listening conditions unilateral, bimodal, bilateral and EAS (electric-acoustic stimulation) served as subjects in the present study.

Results showed a clear discrepancy between the examined patient groups for Fastl noise in MSNF ( SNR=11.0dB). For OLSA noise in S0N0-condition, the results differ less ( SNR=4.7dB). The MSNF with Fastl-noise is particularly suitable to measure the performance of various aided CI-patients in realistic sound environments.

• Headphone Reproduction via Loudspeakers using Inverse HRTF-Filters
  T. Weissgerber, K. Laumann, G. Theile and H. Fastl [C]
[A]

The ''Binaural Sky'' is a loudspeaker array above the listener's head serving as a ''virtual headphone'' for binaural room synthesis. WFS focussed sources are rendered close to listener's ears and used for HRTF inverse filtered signal reproduction (''HRTF inverse filtered sources''). Their locations are constant relative to the ears for every head direction by means of head tracking data. Hence, there is no need to change the inverse HRTF (''crosstalk cancelling'') filters due to head rotations, which could create audible artefacts and instabilities. The present study is concentrating on the stability of the virtual headphone characteristics due to the number and positions of the focussed sources. In a first step the optimum configuration of ''real'' loudspeakers without using WFS is examined. This requires measurements of nearfield-HRTFs for different azimuth and distance of the speaker. HRTF inverse filters based on these data are applied to a number of speaker configurations. The speakers are placed in two circles of different diameter around the listener's head for playback. Transfer functions of the virtual left and virtual right headphone channel as well as both channels are measured and the quality of the virtual headphone evaluated.

• Algorithmic modelling of the disturbance impact of background sounds
  S. Schlittmeier, T. Weissgerber, S. Kerber, H. Fastl and J. Hellbrück [I]
[A]

Mentally demanding tasks very rarely take place during silence. However background sound can reduce cognitive performance, even if it is irrelevant to the task and is intended to be ignored. This so-called Irrelevant Sound Effect (ISE) has been verified in a multitude of behavioral experiments for verbal short-term memory performance.

Although a multitude of cognitive psychological experiments have explored the ISE, no psychoacoustically based instrumental procedure existed to predict its occurrence. This poses problems in an applied context: In office environments, for example, the potential beneficial - or less advantageous - effects of noise abatement on cognitive performance could not be calculated before their realization.

The talk presents an algorithm which models performance data in ISE experiments on the basis of instrumental measurements of the hearing sensation fluctuation strength. It was verified with a database consisting of about 50 background sounds and corresponding performance data that have been collected in cognitive psychological experiments at the KU Eichstätt-Ingolstadt. The algorithm is able to reproduce the performance results in about 90 % of cases within the interquartile ranges. It will be discussed within the scope of cognitive short-term memory models, which claim to explain the ISE and with respect to practical implications.

• Crying colours and their influence on loudness judgments
  D. Menzel, T. Dauenhauer and H. Fastl [I]
[A]

The German term ''schreiende Farbe'' denotes a very salient colour which stands out among other colours. It can be translated as ''loud colour'' or literally as ''crying colour''. As it is known from previous studies that the colour of objects like trains or cars can influence loudness judgments, experiments were performed to determine if such crying colours lead to higher loudness ratings compared to non-crying colours. As a first step, subjects were asked to synthesise colours which they perceived as ''crying'' by adjusting values in the Lab colour-space. A representative sample of these colours was then combined with non-crying colours. The resulting group of 15 colours was rated on a five-point scale ranging from ''not-crying'' to ''very crying''. In a final listening experiment, the same 15 colours were presented simultaneously with uniform exciting noise of different sound pressure levels to asses the influence of the visual stimuli on the subjective loudness ratings. Results indicate that some crying colours where able to elicit - at same SPL of the noise - higher loudness ratings compared to a neutral colour.

Favrot, S.

• A loudspeaker-based room auralisation system for auditory perception research
  J.M. Buchholz and S. Favrot [I]
[A]

Most research on basic auditory function has been conducted in anechoic or almost anechoic environments. The knowledge derived from these experiments cannot directly be transferred to reverberant environments. In order to investigate the auditory signal processing of reverberant sounds, a loudspeaker-based room auralisation (LoRA) system is proposed here. The LoRA system efficiently combines modern room acoustic modelling techniques with higher-order Ambisonic auralisation. Thereby, aspects of the auditory precedence effect are utilized to realise highly authentic room reverberation. This system provides a flexible research platform for conducting auditory experiments with normal-hearing, hearing-impaired, and aided hearing-impaired listeners in a fully controlled and realistic environment. This includes measures of basic auditory function (e.g., signal detection, distance perception) and measures of speech intelligibility. A battery of objective tests (e.g., reverberation time, clarity, interaural correlation coefficient) and subjective tests (e.g., speech reception thresholds) is presented that demonstrates the applicability of the LoRA system.

Fedtke, T.

• Investigations on measurement control and signal processing for the pressure reciprocity calibration of microphones
  T. Fedtke and T. Rewig [C]
[A]

The realization of the unit of sound pressure is done by determining the open-circuit pressure sensitivity of laboratory standard microphones using the reciprocity technique with microphones coupled in pairs by means of cylindrical couplers. The metrological centrepiece of this process is the measurement of the ratio between the open-circuit output voltage of the receiver microphone and the current through the transmitter microphone: the electrical transfer impedance. For its determination, which is based on a computer controlled reference resistor, four alternating voltages are measured, and their quotient is used to calculate the transfer impedance from the resistance value of the reference resistor.

Therefore, the measurement uncertainty of the electrical transfer impedance is influenced by the uncertainties of the reference resistor and of the voltage ratio. In order to obtain an optimal stability and to suppress disturbances, signal analysis and signal processing are used for the voltage measurements, such as averaging procedures and signal stability criteria. Measurement control and signal processing were optimized with regard to the time required for the measurement and the uncertainty. For this purpose, techniques derived from statistical process control, such as control charts, were used and adapted to provide objective criteria for the quality of measurement signals.

Fehndrich, M.

• Influence of impurities between train wheels and wheel damper on the effectiveness of the damper
  M. Fehndrich [C]
[A]

Wheel vibration absorbers by Bochumer Verein Verkehrstechnik are successfully used with train wheels for vehicles of the light and heavy rail to avoid curve squeal and to optimise the level reduction of the rolling noise.

Anyhow, curve squeal occurred at individual vehicles in distinct cases. An investigation of the concerning wheel dampers did not lead to explanatory results. These results could not be reproduced by measurements at similar wheels on the test stand, either.

As working hypothesis, we assumed that impurities on the contact surface of wheel and damper had reduced the damping behaviour. Measurements with a layer of a model dirt are presented. They can explain the effect, as the damping of the wheel was reduced clearly.

Therefore, we developed new dampers, which are insensitive to such dirt layers. The effectiveness of the new damper can be shown on the basis of test results conducted during rides with trams in different cities.

Fehren, H.

• Active Attenuation of Squeal Noise at Train Wheels
  H. Siebald, H. Fehren, M. Schiedewitz, A. Cigada, S. Manzoni and M. Redaelli [I]
[A]

The present work represents preliminary experimental investigation aimed at the possibility to prevent the squealing noise of a resilient train wheel under real operating conditions. The control effectiveness of two different kind of piezoelectric actuators have been tested in a non-rotating test stand designed with the aim of reproducing the wheel working conditions as closely as possible. It is not possible to fully simulate the rotating wheel behaviour, but promising results and important information have been obtained with collocated control. In the next step piezoelectric stack actuators have been adopted on a rotating wheel, on the condition that slip rings have been installed both for signals and power transmission. Tests with a prepared wheel set have been performed on a rotating test stand. A good vibration reduction at the wheel and also a noise reduction in the nearfield of the wheel have been achieved.

Fellendorf, M.

• Traffic Noise Annoyance on Roads and Rail (TNAR) in an experimental laboratory setup
  M. Cik, K. Fallast, R.B. Raggam, E. Marth, E. Gallasch, A. Lackner and M. Fellendorf [C]
[A]

At present, according to the current standards and specifications, the impact of road and rail traffic noise is represented by the A-weighted energy-equivalent sound level (LA,eq). This quantity does not account for the subjective annoyance of sound events as it is perceived by the affected persons.

Pass-by road and rail vehicle noise samples were binaurally recorded under predefined conditions with a dummy head measurement system and synthetically composed to defined vehicle ensembles. A total of 200 persons were selected and exposed to these defined vehicle ensembles in different sessions in a hearing laboratory. Annoyance was rated on an 11-graded interval scale and by a hand power-dynamometer which measures the intensity of the applied strength. Psychoacoustic parameters of the defined vehicle ensembles were calculated. Annoyance ratings were combined with psychoacoustic parameters by multiple regression analysis. The relevant parameters were statistically combined into the index TNAR.

The analyses revealed that loudness is the parameter describing annoyance best, significantly better than the LA,eq (P0.0001). Further significant influencing parameters are sharpness and a new designed parameter representing the frequency characteristic of traffic noise. These three parameters together (R=0.807), correlate significantly better with the subjective estimation of noise-induced annoyance than the LA,eq (R=0.672) does (P0.0001).

Fels, J.

• Application of binaural technology in an adaptive auditory speech test for children
  J. Fels, F. Coninx and W.H. Döring [I]
[A]

This contribution reports on the development of a screening procedure for the identification of hearing loss in children by using a speech test in various spatial conditions.

A study looking at the performance of the Adaptive Auditory Speech Test (AAST) under spatial locatable noise as a possible diagnostic procedure for Auditory Processing Disorders (APD) has been carried out. The AAST was developed to determine the Speech Recognition Threshold (SRT) in a fast and reliable way under quiet and noisy conditions for children of 3 to 4 years of age.

In the study the AAST is used to develop a diagnosis method with a noise made by a talk between children (Two-Talker-Noise) or talking noise (person reading a newspaper) to test the intellectual abilities, separation and localization.

The test was carried out under various spatial conditions with the signal and noise coming from different positions in the horizontal plane. The signals were generated using head-related transfer functions which are appropriate for this age group. Therefore, the test words and the talker-noise were convolved with the corresponding children HRTFs. The scenarios were presented with the help of headphones.

Significant differences were detected between the different spatial situations. Results and experiences from current screening programs will be presented.

Fernandez Astudillo, R.

• Performance Analysis of Wavelet-based Voice Activity Detection
  M. Jeub, D. Kolossa, R. Fernandez Astudillo and R. Orglmeister [I]
[A]

The objective of this paper is to analyse the performance of wavelet-based voice activity detection algorithms (VAD) and to contrast it with that of the VAD standardized for the AMR-WB (Adaptive Multi-Rate Wideband) codec. Experimental results show that wavelet approaches lead to good results with respect to speech clipping and offer a much lower computational complexity. Integration of these algorithms into a Hidden Markov model (HMM) speech recognizer shows that the recognition performance using the AMR VAD can also be obtained or improved upon by wavelet based approaches, again at a notably reduced computational effort.

Fichtel, C.

• Conception of test setups for investigations on the sound transmission from landings in the staircase test facility
  C. Fichtel, T. Schneiderhan and J. Scheck [I]
[A]

The German standard DIN 4109 doesn't allow an accurate prediction of the normalised impact sound pressure level from resiliently supported heavy stairs and landings. Investigations on representative landing constructions are carried out in the staircase test facility to identify and quantify the parameters that are relevant for the sound transmission. The conception of the test setups requires an analysis of the whole transmission system including the landing, resilient layer and the test environment itself to enable a proper transformation of the laboratory data into building situations. A particular difficulty is the simulation of varying pressure conditions on the resilient elements as found in buildings. The presentation contains the latest results of the ongoing investigations.

Fiebig, A.

• The Potential and Benefit of Audible Noise Maps
  K. Genuit, A. Fiebig and P. Marla [I]
[A]

The importance of several (psycho-)acoustic parameters to noise evaluation and noise rating was demonstrated in numerous studies. However, in conventional noise maps neither information about further acoustic parameters nor comprehensive auralization options are available. In the context of the European research project ''Quiet City Transport'' the development of a traffic noise synthesizer was initiated, which deals with the idea of generating time signals representing specific traffic scenarios. On the basis of such generated time signals various acoustical parameters can be calculated and in consideration of sound propagation effects the noise for certain receiver positions can be determined. The advantage of noise maps providing audible examples of noise exposure in important immission points would be that experts and decision makers can listen to the environmental noise and can base their decisions with respect to required noise mitigation measures on a more significant basis. Of course, several criteria like accuracy, uncertainty, practicability, feasibility, interpretability of the maps must be critically discussed before considering the implementation of more extensive auralization tools in noise maps. The presentation will show some case study results. Moreover, the benefit of advanced noise maps will be highlighted and realization impediments will be discussed.

• Psychoacoustic Evaluation of Traffic Noise
  A. Fiebig, S. Guidati and A. Goehrke [I]
[A]

In the European research project ''Quiet City Transport'' the quantitative characterization of annoyance caused by traffic noise using psychoacoustic descriptors for a perception-related environmental noise description was a major objective. In the first project stage single pass-by noises have been evaluated by subjects. The aim was to discover components and patterns, which are responsible for specific annoyance effects. The results showed that certain noise properties such as diesel knocking are particularly annoying regardless of their absolute sound pressure levels. In the second part of the project complete traffic noise scenarios were evaluated to detect indicators which allow for an adequate assessment of environmental noise quality. Using a traffic noise synthesizer as well as artificial head recordings a wide variety of stimuli were created and evaluated by subjects. By means of a principal component analysis links between specific noise properties and subjective evaluation patterns were found and provided the basis for the development of a metric. The results of the study dealing with the evaluation of traffic noise will be presented and discussed under the perspective of their applicability in environmental noise policy and for soundscape studies.

Fingerhuth, S.

• Consonance/dissonance - evaluation of electrical machine noise
  S. Fingerhuth [C]
[A]

In this work the result of two listening tests of consonance/dissonance will be presented. The listeners had to evaluate several sounds with high tonal components. The sounds were synthesized based on recordings of electrical machines. The stimuli represented two machines driven at different speeds generating a more or less consonant/dissonant sound, depending on the ratio of the speeds. Each listeners had to evaluate the sounds in two different experiments, i) using an absolute magnitude estimation (AME) method, without a reference, and ii) using a category partitioning (CP) scale with 5 main divisions (from ''not at all dissonant'' to ''very dissonant'') each one with 10 subdivisions.

The differences and implications of using both methods will also be presented and some conclusions of consonance/dissonance as a psychoacoustic magnitude and consequences for sound design targets will be shown.

Fingscheidt, T.

• Spectral Restoration of Narrowband Speech Recordings Supported by Phonetic Transcriptions
  P. Bauer and T. Fingscheidt [C]
[A]

Due to limitations of the acoustic bandwidth, historic and telephone recordings suffer from poor speech quality and intelligibility. Nevertheless they are widely found in radio, television, internet, as well as in storage media in the context of newscast reports, documentations, or simply in large databases with archived speech material. Spectral restoration of such narrowband speech recordings improves the auditory impression as well as intelligibility. As the recordings are available offline, phonetic transcriptions already exist or can be extracted from the speech data manually or automatically and support the spectral restoration process. Hence, subjective quality and intelligibility much closer to wideband speech can be expected. In this paper, an artificial bandwidth extension (ABWE) is presented to spectrally restore 8 kHz sampled speech signals by upsampling to 16 kHz and estimating further frequency regions of interest. It makes use of the phonetic transcriptions in order to improve the partially insufficient ABWE performance of critical phonemes, i.e., mainly fricatives /s/, /z/ and /f/. We found the spectrally restored speech significantly enhanced, particularly the typical lisping effect disappeared in many instances.

• A Comparison of Instrumental Measures for Wideband Speech Quality Assessment of Hands-free Systems in Echoic Condition
  K. Steinert, S. Suhadi and T. Fingscheidt [I]
[A]

Instrumental speech component quality assessment presumes the availability of the processed speech portion of the enhanced output signal mixture along with its clean input counterpart. As the filtered speech component is unavailable in the context of black box measurements of speech enhancement systems, we have proposed a signal separation technique in earlier publications along with instrumental and subjective evaluations for the narrowband case. In this paper we apply the technique for a black box speech component quality assessment of two wideband hands-free systems with some instrumental measures. In particular, the echo reduction performance in the double-talk case will be investigated. Our findings are compared with the results of white box measures and of a subjective listening test.

Fischer, D.

• In-line monitoring of polymer extrusion processes with ultrasonic attenuation spectroscopy
  J. Müller, S. Kummer, K. Sahre and D. Fischer [C]
[A]

Fischer, G.

• Inverse reconstruction techniques and their applicability for acoustic travel-time tomography
  G. Fischer, M. Barth and A. Ziemann [C]
[A]

Remote sensing by acoustic travel-time tomography enables one to determine area-averaged sound speed fields in an area of interest, e.g. the atmospheric boundary layer. Due to the dependency of the sound speed on temperature, the distribution of this meteorological quantity can be reconstructed by inverting the given travel times of sound signals along defined sound ray paths through a scanned medium. Over the past years, different inverse approaches were used for acoustic tomography. In this study, the results achieved with matrix inversion, simultaneous iterative reconstruction technique, and stochastic inversion are compared. To identify the algorithm which yields the optimal results by reconstructing two-dimensional temperature fields, both synthetically generated and experimental travel times were considered. On the one hand, the temperature fields were compared concerning their reconstruction accuracy and spatial resolution. The results show that the stochastic approach yields spatially highly resolved and accurate temperature fields. Furthermore, to estimate the applicability of the different methods during operational studies, the computational effort was analysed. Here, the stochastic inversion offers some deficits because of the relatively long computational time. Oppositely, the matrix inversion and the iterative reconstruction technique are suitable for online interpretations by providing temperature fields with acceptable accuracy.

• Adapting a sound ray model of the atmosphere to simulate sound propagation in urban environment
  A. Ziemann and G. Fischer [C]
[A]

Noise exposure in urban areas is an important factor which influences the well-being and quality of life of urban residents. According to recent studies, the population is most frequently disturbed by road traffic noise. Particularly in residential areas near main roads this problem is essential, because prescribed noise levels are often exceeded here. When analysing this problem, the influence of the atmosphere on sound propagation under specific urban environmental conditions also has to be investigated. Therefore, to operationally answer questions of noise protection, the sound propagation model of the atmosphere using ray-tracing (SMART) has been developed. In the present study, this model, which considers the influence of the vertical atmospheric structure on sound refraction and absorption, was enhanced to allow simulations of sound propagation in urban areas. Thereby, investigations of the dependency of the sound propagation on temporally variable weather conditions were performed. Furthermore, as a consequence of noise exposure near residential areas, noise barriers are often installed along busy roads. To quantify the effect of this active noise protection, a barrier was implemented in SMART. The results of first studies illustrate that special weather conditions might reduce the effect of such noise barriers.

Fischer, H.-M.

• Direct sound insulation of double-leaf separating walls
  R. Marin, M. Schneider, J. Scheck and H.-M. Fischer [C]
[A]

The simple calculation of the sound insulation of stiff double-leaf separating walls according to DIN 4109: 1989 Beiblatt 1 allows only a rough estimation of the in-situ sound reduction index. Construction parameters like wall material, cavity thickness, absorption material in the cavity and the flanking transmission are only considered by general requirements. An alternative approach is to obtain the in-situ sound insulation from the (undisturbed) direct sound insulation and the situation-dependant flanking sound reduction. The direct sound insulation of the double-leaf wall is obtained from the direct sound insulation of the first leaf and by adding the improvement of a second leaf with respect to the design of the cavity. Measurement results for the direct transmission are compared with results obtained by various prediction models.

• On force- and moment mobilities of a timber joist floor
  A.R. Mayr, B. Gibbs and H.-M. Fischer [I]
[A]

In wood framed buildings, floors are formed by fastening wood sheathing to joists spaced at a regular interval. The sheathing is typically fastened to the joist using screws so the resulting system is a complicated periodic point-connected plate-rib structure. It is shown that the point force mobility varies significantly with position. A machine installation generally is close to discontinuities, such as at: floor edges, joist-screw locations, joints between the sheathing plates, and due to workmanship. In such cases, moment excitation might become important at some frequencies and the neglect of moments a priori can lead to inaccurate prediction of the total emission. This paper presents measured point moment mobilities with respect to the distance to discontinuities for a timber joist floor where a single layer of chipboard forms the sheathing. It is shown that the measured point moment mobility indicates an infinite plate behaviour. This includes for positions above a joist but between screw positions. To determine the relative contribution of moments and perpendicular forces to the total structure-borne sound power, case studies of two sources, a fan unit and a whirlpool bath, are described, for various locations on a timber joist floor.

• Vibration behaviour and structure-borne sound transmission of a resiliently supported landing
  E. Taskan, J. Scheck and H.-M. Fischer [I]
[A]

The expected normalized impact sound level for heavy stairs and landings is presently predicted for less and obsolete constructions and insulation in the latest version of the DIN 4109. Thus the calculation of the standard impact sound level for resiliently supported heavy stairs and landings is according to DIN 4109 - Beiblatt 1 impracticable. By analyzing the structure borne transmission paths, appropriate acoustical parameters can be related to the involved elements (i.e. flight of stairs, stair landing or decoupling elements) and included into a calculation model, for example DIN EN 12354. The vibration behavior of the landing and the transmission through the resilient layer are of particular importance. The vibration behavior of a representative landing construction was investigated by means of experimental and computational modal analysis and a finite element model was set up. It is intended to advance the FEM model in order to simulate the transmission from the landing through the resilient layer into the wall to enable parameter studies.

• Study of the Influence of Adjacent Elements on the Sound Level Decay of Heavy Building Structures by Means of Transient SEA
  M. Schneider, F. Mack and H.-M. Fischer [I]
[A]

The sound reduction index of heavy constructions depends on the energy losses from this element into the respective adjacent building elements. For each in-situ situation different energy offtake into the coupled elements may occur. When measuring the sound reduction index of a building element in the laboratory, it is therefore necessary to determine the loss factor of this element in its respective laboratory condition. But this loss factor measurement may not only contain energy flux from the considered element into the adjacent elements but also energy flux from the adjacent elements into the considered one. The energy flux back into the considered element depends on the energy levels and energy level difference between the elements and therefore on the internal losses of the individual elements. This time dependent energy flow is modeled with a transient SEA approach. This approach gives us an idea about the circumstances under the energy flux from the adjacent elements to the considered element will influence or determine the level decay. Furthermore measured reverberation times of heavy walls are compared with the decay calculated by the transient SEA.

• Acoustical characteristics of lightweight solid gypsum walls and their implementation into prediction models
  A. Ruff and H.-M. Fischer [C]
[A]

In multi storey buildings gypsum blocks are often used to construct solid inner walls without static requirements. These gypsum walls with a mass per unit area of only about 90 kg/m are not connected rigidly to the adjacent building elements. They are commonly decoupled from the adjacent construction elements with elastic interlayers made of bitumen, cork or polyethylene foam. The direct sound insulation of the gypsum walls is depending on the kind of the used elastic interlayers. The elastic interlayers have also a significant influence on the flanking sound transmission of the gypsum walls. Therefore the decoupled walls have to be treated different to rigidly connected walls.

Within a research project the direct and the flanking sound transmission of gypsum walls is investigated in different test facilities in the laboratory. For different kinds of elastic interlayers the sound reduction index of gypsum walls is measured in a test facility for direct sound transmission. Measurements of the vibration reduction index are carried out in two test facilities for flanking transmission (vertical and horizontal) and also in different building situations. These measurement results are used as input data for the prediction model according EN 12354-1.

Fischer, R.-L.

• Application of the Binaural Speech Intellegibility Model for prediction of intelligibility in in-car noise
  R.-L. Fischer, S. Meyer, J. Otten and C. Meinecke [I]
[A]

Conventional speech prediction models are based on monaural input, which limits the application in spatial conditions. However, in vehicles signal and noise sources are spatially distributed. Hence, an application of a binaural model is likely to increase the accuracy of speech intelligibility prediction. The Binaural Speech Intelligibility Model (BSIM, Beutelmann & Brand, 2006) was used to check, whether this binaural SII gives advantage over monaural approaches as SII and STI (speech transmission index). The advantage of the BSIM is the combination of the monaural speech intelligibility index (SII, ANSI, 1997) with the binaural equalization- cancellation processing (Durlach, 1984). The model is taking into account the binaural release of masking, if there is a directional separation of speech and interfering noise. The evaluation of speech prediction accuracy was based on measurements with normal hearing subjects. The collected speech related threshold data for several in-car noise conditions are compared with speech indices predic- tions named above. Results of the study will be presented and discussed concerning implications for future work.

Fischer, S.A.

• Acoustical environment in open-space offices - How to achieve the field of confidence?
  S.A. Fischer and E.-J. Völker [C]
[A]

The open-space office has a sensitive acoustical behavior and has therefore to be designed carefully. There is e.g. disturbing noise from adjacent working places. The falling off of noise-level with increasing distance is too small. Rooms are reverberant and noisy. Intelligibility of speech is too high and leads to a lack of concentration. Many sound reflections from walls, from the ceiling, and the furniture arrive at the working place and increase the sound levels. The open space is a sensitive system with many parameters which form an acoustical field of confidence. It must be provided at every working-place. It includes the criterions of privacy defined for different requirements and expectations such as team-workers or single working places. Background noise may be introduced to mask interfering noise from adjacent places. It should be a non-informative noise comparable with pink noise. The paper deals with the different parameters and their application in today's office design. Measurement results will be included, such as reverberation times, sound reflections, decrease of sound levels, and background noise levels. Evaluations will be presented.

Fischer, S.

• Predicting the acoustics of individual ears for hearing aid and audio applications - model framework and future work
  M. Blau, T. Sankowsky, P. Roeske and S. Fischer [C]
[A]

Applications involving sound delivery to the ear, such as hearing aids, mobile phones, ear- and headphones, will always have to cope with the problem of varying acoustic loads, due to interindividual differences in shape of the outer ear and mobility of the ear drum. It has been known for a long time that this variability can lead to considerably different sound pressure levels at the ear drum for frequencies above a few kHz. An additional challenge is the more and more frequent use of ear molds which do not acoustically seal the ear canal (''open fittings'' in hearing aid terminology). In this paper, a simple 1-dimensional model framework is presented, which may be used to predict individually different acoustics (including the pressure at the ear drum and acoustical feedback), or for real-time simulations of hearing instruments. The discussion will include known model sub-components, as well as open questions concerning the needed ''individuality'' parameters, or resulting for instance from violations of the 1D assumption in some parts of the model. Some of the open questions will be dealt with in the companion papers by Sankowsky et al. and by Roeske et al.

Fleischer, H.

• Characteristic Tones and Modes of a Church Bell
  H. Fleischer [I]
[A]

Fleischer, M.

• A Three Dimensional Model of the Organ of Corti of the Guinea Pig
  J. Baumgart, M. Fleischer, R. Gärtner and A. Voigt [C]
[A]

For the function of the inner ear the mechanical linkage of basilar membrane displacement to the deflection of the inner hair cell stereocilia is essential and investigated here. Besides the passive excitation by a pressure on the basilar membrane also the excitation by electromotile outer hair cells as well as by the hair bundles of the outer hair cells is studied.

Based on geometrical data of the guinea pig and experimental velocity measurements the organ of Corti is modelled by means of the finite element method. The fluid and the structure are modelled as linear material with small displacements. The interaction of the fluid and the structure is implemented as a matrix coupling. The geometry is extracted from laser scanning microscope stacks and material properties are taken from the literature and adjusted to match experimental results.

This model provides a helpful tool to understand experimental data more in detail. It is possible to look for the influence of geometrical variations and material properties. Different configurations and boundary conditions are modelled and the relation of excitation to inner hair cell stereocilia displacement due to bending is analysed.

Foken, W.

• Methods for the Transfer Path Analysis
  T. Stichling, M. Quickert and W. Foken [I]
[A]

Transfer Path Analysis (TPA) is a well known technique to evaluate and to separate the different transfer paths describing how noise gets from the sources into the compartment of a passenger car, i.e. to the driver's ear. It is not only a measurement technique but a prediction method as well. Based on previous measurements and assumptions TPA provides the opportunity to predict the interior sound for a virtual prototype. Due to the progress in computational technology and measurement techniques, TPA is being seen as an important tool in automotive acoustic design. The paper gives a brief survey of the mathematical background of different TPA methods like the Complex-Stiffness-Method or the Inertance-Matrix-Method. With the help of a 5:1 scaled simplified car model differences between the methods of calculation are verified and discussed. In conclusion the paper presents the measurement setup for a real car with a dominating transfer path of the structure borne engine noise into the passenger compartment.

• Application of Operational Transfer Path Analysis on a Classic Car
  D. de Klerk, M. Lohrmann, M. Quickert and W. Foken [I]
[A]

This paper discusses the analysis of a classic car with the Operational Transfer Path Analysis (OTPA). This method poses a fast and efficient way of identifying critical vibration paths and source contributions within vehicle design. Only measurement data of the operating vehicle are needed to perform this analysis. Real measurement situations for example are several vehicle run-ups, full load, halve load, run down. These kinds of measurements are commonly performed in the automotive industry and can therefore be easily adopted. The method OTPA is introduced and discussed based on a classic car Wartburg 311. This vehicle is described in the project sketch ''method comparison of the transfer path analysis'' (Fraunhofer IWU Dresden) and provided by Westsächsische Hochschule Zwickau, Institut für Kfz- Technik (IfK).

Fournelle, M.

• Contrast Enhanced Optoacoustics for Molecular Imaging
  M. Fournelle, W. Bost, S. Wagner, F. Stracke and R. Lemor [I]
[A]

Optoacoustic imaging can provide valuable information about tissue structures by combining the high intrinsic contrast of optical modalities with the good resolution and the high penetration depth of ultrasound. When biological tissue is irradiated with ultrashort laser pulses of durations of a few nanoseconds, the light is absorbed according to the local optical properties of the tissue, and is converted into pressure by means of the thermoelastic effect. The resulting pressure waves can be acquired and analyzed with techniques similar to conventional ultrasound and provide an image of the distribution of absorbed light in tissue. Although endogenous chromophores such as melanin or haemoglobin result in remarkably high optoacoustic contrast, the administration of additional targeted nanoparticular contrast agents becomes necessary if a more selective molecular imaging is needed. Different types of gold nanoparticles are particularly suitable in this context since their strong plasmon resonance results in a high absorption coefficient. An overview on results regarding the usability of the different particle types as optoacoustic contrast agents will be given and the technology and its applications will be discussed. Further, first results on contrast-enhanced in-vivo imaging with a system for real-time optoacoustics will be presented.

Franz, S.

• Multi-channel noise reduction for binaural hearing aids by using short-time spectral attenuation combined with noise estimators for non stationary noise
  S. Franz, J. Bitzer and U. Simmer [C]
[A]

Multi-channel noise reduction (NR) with binaural output is an ongoing research topic for future hearing aids. Especially the preservation of interaural level and time differences is important in order to keep the spatial impression intact. In this contribution we present an extended noise reduction algorithm based on the two-channel NR scheme from Kim et al. The algorithm has a structure similar to the generalized sidelobe canceler (GSC) structure, however, instead of reducing the noise components by cancellation, short-time spectral amplitude filtering is used. We extended and tested this algorithm with different compensation filters to reduce the estimation error as well as musical noise. As a new approach a compensation filter based on minimum statistics will be presented. This method has the advantage that no voice activity detector is necessary. The evaluation will compare different filter estimation techniques on a realtime developing platform for hearing aid algorithms. Finally, we will discuss the advantages and drawbacks of the compared methods.

Franzen, T.

• Experimental investigations on the sound transmission of drywall constructions at low frequencies
  R. Völtl, T. Franzen and U. Schanda [C]
[A]

For a better understanding of sound transmission of drywall constructions in the frequency range below 100 Hz several measurements in a laboratory according to ISO 140-1 were carried out, the receiving part of the laboratory had been changed to semi free-field conditions. Sound intensity levels in the near field of the wall as well as acceleration levels were mapped by excitation either with airborne sound or with a shaker. With the mapped values the radiation efficiency of the wall was calculated. In the frequency levels below approx. 250 Hz a clear rise of the radiation efficiency was observed, contrary to different theoretical models. Furthermore the measured and mapped eigenmodes of the drywall show significant differences when exited by airbone sound or by the shaker in the frequency range below 50 Hz. In order to find out the reasons of these observations many experimental investigations have been carried out und willl be reported.

Fremerey, C.

• Handling Scanned Sheet Music and Audio Recordings in Digital Music Libraries
  C. Fremerey, D. Damm, M. Müller, F. Kurth and M. Clausen [I]
[A]

Significant digitization efforts have resulted in large music collections, which typically contain various types and modes of data. In particular, such collections may comprise audio data in form of CD recordings as well as image data in form of scanned sheet music, thus concerning both the auditory and the visual modalities. In this paper, we review various techniques for automatically processing, analyzing, and organizing such multimodal music collections. We then present a music player system along with various interfaces that allow a user to conveniently search, navigate, and browse across the two types of music data. For example, during audio playback our system synchronously highlights the corresponding musical measures within the sheet music. Furthermore, a user may formulate a query by marking a region of measures within the scanned score to trigger an audio search. The system retrieves all available audio recordings along with the temporal positions of the audio segments that correspond to the queried measures. The user may then freely navigate across the various retrieved interpretations. Our techniques and interfaces are prototypically integrated into the library service system of the Bavarian State Library as part of the PROBADO project.

Frese, D.

• Simulation of curve squeal considering the vehicle dynamics
  D. Frese and T. Kohrs [C]
[A]

Friebe, S.

• Design and Tests of Thermoplastic Textile-reinforced Composite Trays for Vibro-acoustic Relevant Applications
  W. Hufenbach, M. Dannemann, S. Friebe, F. Kolbe and M. Krahl [C]
[A]

Driven by increasing customer demands, environmental and economical requirements, lightweight engineering gains in importance for various industrial and transport applications. Here, composite materials based on thermoplastic hybrid yarns offer new possibilities for multifunctional design due to its adjustable properties. Requirements concerning high stiffness and damping as well as low weight and short cycle times during manufacture can be met by an adapted composite lay-up and manufacturing process. In the presented work, the vibro-acoustic behaviour of selected textile-reinforced thermoplastics was characterized dependent on the kind of reinforcement, number of layers and its fibre orientation. Therefor, models based on Finite Element Method and Boundary Element Method for the calculation of modal and acoustic parameters of textile-reinforced structures were developed. For the verification of this simulation models as an example composite trays were manufactured by a novel hot pressing technology. The modal parameters and transmission loss of the produced trays were measured. Subsequently, the experimental results were used for the verification of the numerical simulations. The performed investigations illustrate that the vibro-acoustic behaviour of composite trays is influenced by the complex interaction of geometry and material parameters. Thus, an optimal sound radiation behaviour can only be achieved by a comprehensive material-adapted vibro-acoustic design.

Friedrich, C.-S.

• Comparison of Coding Techniques for Photoacoustic Coded Excitation
  M. Mienkina, A. Eder, C.-S. Friedrich, N. Gerhardt, M.R. Hofmann and G. Schmitz [C]
[A]

In photoacoustics (PA) ultrasound is generated by laser irradiation of tissue. Usually, PA signals are generated by Nd:YAG lasers but it would be attractive to use less expensive laser diodes instead. Although laser diodes exhibit low pulse energy, the pulse repetition frequency (PRF) of pulsed laser diodes is much higher than the PRF of Nd:YAG lasers thus averaging can be used to increase the SNR. The applied PRF for averaging is limited by the acoustical time-of-flight. To further increase the SNR we propose to use coded excitation for PA imaging. In optical-time-domain-reflectometry, which is similar to PA imaging, Simplex Codes (SC) are widely used for coding. In ultrasound imaging Golay Codes (GC) are successfully employed. Therefore, we compare the performance of SC and GC for PA coded excitation (PACE) by a simulation study. For both codes, the coding gain normalized to equivalent averaging procedures is computed as a function of code length and PRF of the laser diode. Assuming a time-of-flight of 60 s and e.g. a PRF of 250 kHz the coding gain of GC is 1.6 dB higher than the coding gain of SC. Consequently, GC is more suitable for PACE than SC.

Friedrich, M.

• Evaluation of Spatial Approaches for Improvement of Patient Speech Acquisition at Magnetic Resonance Tomography Systems
  M. Friedrich, K. Porzelt, M. Bechtold and P.-C. Eccardt [C]
[A]

During operation of magnetic resonance (MR) tomography systems, switching noise is produced with high sound pressure levels. This noise is a disturbance for patient-to-operator communication, as well as for examinations that require synchronous recording of patient speech (fMRI). Signal processing algorithms can be applied for the removal of MR and other noise sources from patient speech. However, the disturbed speech signal should be obtained with the highest possible input SNR, which is complicated by limitations in practical application. On a typical MR installation, the room acoustics and sound field characteristics were examined, with the objective of optimising the microphone solution with regard to patient speech pick-up. Applying coherence based measures, the noise field is characterised on and around the MR system. General conclusions are drawn about possibilities to employ spatial information for optimising the speech level with regard to interfering noise sources, such as using microphones with directional characteristics, or multiple microphones and source separation algorithms. The choice of suitable microphones and their positioning on a MR installation are discussed. In addition, experiment results are presented for SNR improvement on a real MR installation, using microphones with different directional characteristics on suitable position.

Friedrich, P.

• On the Classification of Acoustic Sequences for Intervention in Essential Hypertension
  P. Friedrich, T. Kohler and B. Wolf [C]
[A]

Acoustic signals can modulate the human metabolic and central-nervous functions and evoke physiological effects. Especially the anti-hypertensive effect of certain iterative sound-patterns as possible intervention for essential hypertension is examined in many recent studies as well as our own research work. It was shown that the acoustic sequences and sound clusters used for intervention can decrease blood pressure significantly. For a therapy, however, it is necessary to identify the musical active ingredients of the sounds in the sense of active pharmaceutical ingredients. This contribution discusses the systematic analysis of the musical features that are responsible for the anti-hypertensive effect. More than 400 features were extracted and investigated in terms of their relevance concerning a blood pressure lowering effect. The 17 most significant characteristics were used to develop a classifier based on the discriminant analysis that decides whether a sound pattern is sedative or stimulating. This will be clarified in acoustic demonstrations. With this tool it is possible to filter from a large selection of music sequences the most suitable therapeutical sound patterns. We now have the starting basis to provide individualized and personalized therapies while respecting personal preferences of users.

Fritzsch, A.

• Extended Multi-Plane Pressure Mode Matching for CFD/CAA Coupling
  C. Weckmueller, A. Fritzsch and S. Guerin [C]
[A]

The objective is to extend the pressure mode matching method proposed by Ovenden and Rienstra in order to improve the aerodynamic/acoustic splitting at the interface between CFD and CAA computational domains. The numerical calculation and propagation of BPF tones in bypass ducts is the aimed application. CFD/CAA hybrid approaches have been developed for reducing the costs of aeroacoustic computations. Within the sound source region (rotor/stator stage), the unsteady Reynolds Averaged Navier-Stokes equations are solved, whereas in the rest of the computational domain, the sound propagation is computed with CAA methods solving the Euler equations. At the CFD/CAA interface, only the acoustic fluctuations should be transfered into the CAA domain to avoid stability issues in the CAA simulation. Ovenden and Rienstra proposed a so-called multiplane pressure mode matching method to remove the vortical part of the fluctuations. As suggested by De Roeck, this method can be extended with an aerodynamic model to improve the splitting. Here we have studied an aerodynamic model consisting of the same radial eigenfunctions as the ones taken for the acoustic modal basis, while the axial phase-velocity is the mean flow convection velocity. This is analytically investigated and applied to some numerical data.

Fritzsche, C.

• Silent Owl Flight: Setup for Flyover Noise Measurements
  C. Fritzsche, E. Sarradj and T. Geyer [C]
[A]

It is common knowledge that most genera of owls fly silently in order to be able to catch their prey. To investigate the mechanisms leading to that quiet flight, several basic fly-over measurements were carried out by biologists in the past. But the very low noise that is produced when the owls are in gliding flight makes it very hard to do exact acoustic measurements.

The presentation describes an attempt to do fly-over noise measurements of trained birds on natural conditions. It is focused on the measurement setup which consist of a small linear microphone array on a ground plate. The microphone array data is processed using a simple weighted-sum beamforming algorithm. Birds pass over the array in gliding flight at very low speed. A camera setup is used for flight speed measurement. A number of tests were conducted in a wildlife park. Preliminary results are presented from several raptor genera, including owls. Necessary improvements of the measurement setup will be discussed.

• Silent owl flight: experiments in the aeroacoustic wind tunnel
  T. Geyer, E. Sarradj and C. Fritzsche [C]
[A]

It is common knowledge that most genera of owls fly silently in order to be able to catch their prey. To investigate the mechanisms leading to that quiet flight, several basic fly-over measurements were carried out by biologists in the past. But the very low noise that is produced when the owls are in gliding flight makes it very hard to do exact acoustic measurements, especially when using the measurement techniques available at that time.

The presentation describes the measurements of the noise generation at the wings of silent and nonsilent flying birds in the aeroacoustic wind tunnel of the BTU Cottbus. The measurement setup will be explained, especially regarding the usability of conventional measurement techniques for experiments on wing specimen. A 56 channel microphone array is used to measure the noise generated at the wings in the fluid flow. The acoustic data is processed with advanced beamforming algorithms to obtain the radiated sound pressure levels and the location of the noise sources on the surface of the wings.

First results will be given that indicate that the magnitude, the spectral shape and the location of the noise produced by owls wings is different than those of nonsilent flying birds.

Fuder, G.

• Measuring Frequencies with Historic Resonators from SCHAEFER
  R. Hoffmann, D. Mehnert, R. Dietzel and G. Fuder [I]
[A]

Measuring frequencies of acoustic processes within the audible range required large technical expense in the early times of experimental phonetics and acoustics. Indirect methods like KUNDT's tube, the monochord, the interpretation of kymographic recordings, or the interference tube from QUINCKE allowed such kind of measurement, but had been too complicated for practical purposes. Therefore simpler measurement methods have been looked for in early times. E. g., it is possible to apply one-sided closed pipes of variable length as resonators. The acoustic perception of the experimentator was used as indicator.

This paper investigates this measuring method at first within the historic arrangement. Then, the transfer behaviour of single tube resonators from SCHAEFER will be measured in the free sound field.

These results as well as the ability of judgement of the listener are used to estimate the measuring uncertainty of such historic methods.

Fürst, P.

• Squeal Noise - Long Time Examination of Stochastically Occurence on Tramways passing Curves
  P. Fürst and M. Dybek [C]
[A]

In the last years exactly parametrically examination on squeal noise has demonstrated, that many factors of influence are existing. In this paper we will demonstrate, that also stochasti-cally occurrence has an influence of the sound emission of a track curve. Observations in a track curve during one year show, that the frequency of noise emission under same condi-tions don't achieve exactly comparable results. So we say, that there is only a limited prob-ability depending on the different parameters that squeal noise is appearing or not. Other-wise we found, that the curve is squealing, but there is no typical situation for the appearance of squeal noise. We will present results within variable trams and weather parameters on the same curve. At the end we will compare our observations at another kind of tram curves.

Fuß, S.

• Modal Analysis of a Fluid inside and around a Recorder
  S. Fuß and S. Marburg [I]
[A]

This talk presents the three-dimensional numerical simulation of the sound spectrum and the propagation of the acoustic noise inside and around a recorder. The fluid inside and close to the recorder is meshed by Lagrangian tetrahedral finite elements. To optain results in the far field of the recorder, complex conjugated Astley-Leis infinite elements are used. To apply these infinite elements the finite element domain has to be meshed either in a spherical or an ellipsoidal shape. Advantages and disadvantages of both shapes regarding the recorder will be shown in this talk. Examples of the modal analysis of the fluid for different tones will also be presented.

Gaafar, A.

• The new electronic report of noise-exposure-measurements and audiometric-examinations in the AUVA
  W. Wahler, K. Körpert and A. Gaafar [C]
[A]

The AUVA is the Austrian Social Insurance for occupational risks for approximately 3 million employed persons and 1.3 million school children and students. Therefore, a central motivation of the AUVA is the prevention of occupational accidents and diseases, e.g. the noise-induced hearing loss. One major element of the prevention activities are the visits of prevention workers in companies to improve the employees' conditions of hearing by an audiometric examination. To pick out the employees which are heavily exposed, it is necessary to determine the noise exposure level at work places by measurement and calculation. The data entry of the detailed noise-exposure and audiometric studies can be now electronically supported by a new computer application called EPOS. The objective of this paper is to report on the challenges and experiences gained during a five year multidisciplinary software development project and the essential components of the software will be presented.

Gade, A.C.

• Relationship between room shape and Early Lateral Energy Fraction in rectangular concert halls
  A.K. Klosak and A.C. Gade [C]
[A]

The acoustic properties of 24 rectangular rooms representing ''shoe-box'' type concert halls with volumes between 8000m3 and 16000m3 have been analysed using Odeon (v.8) computer simulation software. Among several objective measures calculated in numerous receiver positions (between 300 and 850) in each hall, this paper concentrates on the behaviour of Early Lateral Energy Fraction (LF80). The dependency of LF80 on source-receiver distance, receiver location within the room and room dimensions will be discussed. It is found that the distribution of LF80 in rectangular rooms can be described as a three-dimensional ''tongue-shape''. The simulation results have also been compared with regression models based on measurements in real halls. In rectangular halls, the variation in LF80 values can be described by the influence of the overall room dimensions and their mutual ratios. Formulas for prediction of LF80 in rectangular halls have been proposed, which takes into account both the width and the length of the hall.

Gaertner, D.

• Preprocessing methods for rhythmic mid-level features
  C. Dittmar, M. Gruhne and D. Gaertner [I]
[A]

Rhythmic mid-level features are an important pre-requisite in Music Information Retrieval. They can be deployed to describe the rhythmical gist of songs for various tasks, such as genre classification and music similarity search. Although different computation strategies have been proposed in the literature, rhythmic mid-level features commonly represent the most salient rhythmic periodicities in a music signal. Major shortcomings of rhythmic mid-level features are their dependency on the actual tempo of the songs and their susceptibility to degradation caused by interference with non-rhythmic signal components, such as melodic-sustained instruments. This publication describes preprocessing strategies applicable to rhythmic mid-level features. The problem of interference with melodic instruments is addressed by emphasizing the rhythmic content via a signal decomposition related to Drum Transcription. Tempo changes are tackled by logarithmic re-sampling of the features' lag a xis. An evaluation of the proposed methods is conducted via genre classification using both artificial and real-world music signals. The evaluation results are presented and discussed accordingly.

Gärtner, R.

• A Three Dimensional Model of the Organ of Corti of the Guinea Pig
  J. Baumgart, M. Fleischer, R. Gärtner and A. Voigt [C]
[A]

For the function of the inner ear the mechanical linkage of basilar membrane displacement to the deflection of the inner hair cell stereocilia is essential and investigated here. Besides the passive excitation by a pressure on the basilar membrane also the excitation by electromotile outer hair cells as well as by the hair bundles of the outer hair cells is studied.

Based on geometrical data of the guinea pig and experimental velocity measurements the organ of Corti is modelled by means of the finite element method. The fluid and the structure are modelled as linear material with small displacements. The interaction of the fluid and the structure is implemented as a matrix coupling. The geometry is extracted from laser scanning microscope stacks and material properties are taken from the literature and adjusted to match experimental results.

This model provides a helpful tool to understand experimental data more in detail. It is possible to look for the influence of geometrical variations and material properties. Different configurations and boundary conditions are modelled and the relation of excitation to inner hair cell stereocilia displacement due to bending is analysed.

Gajdatsy, P.

• Innovative Approaches to Fast Transfer Path Analysis
  H. van der Auweraer, P. Gajdatsy, K. Janssens, P. Mas and L. Gielen [I]
[A]

Over the last decennia, Transfer Path Analysis (TPA) has become an accepted tool for NVH troubleshooting and internal load estimation. A major bottleneck preventing its even more widespread use in the actual vehicle development process is the test time to build the full data model, requiring extensive Frequency Response Function testing next to the obvious in-operation tests. Recent innovations on instrumentation and new procedures such as Fast and Multilevel TPA and Operational TPA address this limitation, however often introducing new constraints. For example, OTPA only requires operational response data at reference and target locations. But, being basically a transmissibility method, its application requires special care regarding the inherent ill-conditioning and reference-completeness problems. It essentially assesses a co-existence relationship between targets and references and does allow causality and contribution interpretations. Hence a new approach is proposed, using a parametric model for the estimation of loads. This makes the method scalable, enabling the engineer to use a simpler model based on a small amount of measurement data for quick troubleshooting or increase accuracy by using a more complex model together additional measurements. The method is applied to a customer problem and compared with the existing approaches, showing the advantages in real life.

Gallasch, E.

• Traffic Noise Annoyance on Roads and Rail (TNAR) in an experimental laboratory setup
  M. Cik, K. Fallast, R.B. Raggam, E. Marth, E. Gallasch, A. Lackner and M. Fellendorf [C]
[A]

At present, according to the current standards and specifications, the impact of road and rail traffic noise is represented by the A-weighted energy-equivalent sound level (LA,eq). This quantity does not account for the subjective annoyance of sound events as it is perceived by the affected persons.

Pass-by road and rail vehicle noise samples were binaurally recorded under predefined conditions with a dummy head measurement system and synthetically composed to defined vehicle ensembles. A total of 200 persons were selected and exposed to these defined vehicle ensembles in different sessions in a hearing laboratory. Annoyance was rated on an 11-graded interval scale and by a hand power-dynamometer which measures the intensity of the applied strength. Psychoacoustic parameters of the defined vehicle ensembles were calculated. Annoyance ratings were combined with psychoacoustic parameters by multiple regression analysis. The relevant parameters were statistically combined into the index TNAR.

The analyses revealed that loudness is the parameter describing annoyance best, significantly better than the LA,eq (P0.0001). Further significant influencing parameters are sharpness and a new designed parameter representing the frequency characteristic of traffic noise. These three parameters together (R=0.807), correlate significantly better with the subjective estimation of noise-induced annoyance than the LA,eq (R=0.672) does (P0.0001).

Garbers, J.

• Bridging Music Information Retrieval and Folk Song Research - The Computational Setup of the WITCHCRAFT Project
  J. Garbers [I]
[A]

The WITCHCRAFT project sets as its objective to develop a fully functional content-based retrieval system for folk song melodies. Besides of making the Netherlands folk song collection of the Meertens Institute available for the general public, we aim to support their research in oral variation.

Folk song researchers traditionally know many songs by heart and intuitively find correlations between them. For large or international collections, however, researchers cannot know all the song instances, so computer aided support is indispensable to reveal related items. This in principle means to take existing models about oral variation and design and implement similarity measures that can be used in a retrieval system. This, however, proves to be difficult in practice.

In this paper we present our information retrieval setup consisting of performed folk song recordings, their symbolic encoding, an expert classification of variant tunes (ground truth), annotation data about the particular expert classification reasons and our approaches to similarity. As an example we present how we translated the expert concept of ''pitch stability'' into an alignment-based similarity measure and how we evaluated our method and the original ''pitch stability'' hypothesis with respect to the given annotations.

Garcia, I.

• Soundscape characterization in the city of Bilbao
  G. Memoli, I. Garcia and I. Aspuru [I]
[A]

The first round of noise mapping and action plans has left a lot of unresolved questions, mainly addressing the design of perception-oriented action plans and the classification/preservation of quiet areas. While the interest in the practical use of the soundscape concept to address this problems is now growing all through Europe, psychoacoustic indicators linked to perceived noise are the key to determine the path to follow. This work presents the soundscape characterization of two different green areas in the Municipality of Bilbao (E), with in common the presence of a heavily trafficked road nearby. The acoustical experience of passers-by will be mapped using an indicator related to the time history of sound energy, related in previous studies to people's perceptions. Comparison with the same description performed by classical psychoacoustic analysis and perspectives for innovative, positive soundscape based actions and measurement techniques will be discussed.

Gaul, L.

• Fast FE-Analysis and Measurement of the Hydraulic Transfer Function of Pipes with Non-Uniform Cross Section
  J. Herrmann, M. Spitznagel and L. Gaul [C]
[A]

Automotive piping systems are often characterized by a non-uniform cross-section. Orifices in fuel charge assemblies are a typical example. The goal of this research is to quantify the influence of diameter changes on the hydraulic transfer function between the dynamic pressure at the beginning and at the end of the pipe. A finite element based substructuring and component mode synthesis technique is used to assemble the spatial piping system including the orifice. The adaptation of the Craig-Bampton method on fluid-structure coupled piping systems with an additional reduction of the remaining interface degrees of freedom leads to a considerable model order reduction. A local, frequency dependend fluid damping model is integrated in the finite element code and the result of the harmonic analysis is compared to measurements. A hydraulic test bench with a dynamic pressure source is used to estimate the hydraulic transfer function for different orifice diameters.

Geebelen, N.

• The Use of Reciprocity in determining Rainfall Noise
  N. Geebelen and H. Cauberg [C]
[A]

The theorem of reciprocity exists in many forms and is therefore applied in a wide range of scientific fields. With regard to building acoustics, a relation exists between the sound radiation and the vibration response of a structure. From this, a formula can be derived that connects the air-borne and structure-borne sound insulation of the structure, being only dependent on frequency. This way it is possible to calculate the air-borne sound insulation when the structure-borne sound insulation is known and vice versa, creating lots of practical possibilities. Rainfall noise comes from the mechanical excitation of light structures by rain drops. The measurement procedure for determining rainfall noise is however quiet elaborate. The question rises if reciprocity can be used to simplify this specific set-up? Or else said, when the type of rain (number of drops, intensity, etc.) is known, is it possible for a light structure to predict rainfall noise from the air-borne sound insulation? The use of reciprocity creates many practical advantages, but the relations and formulas have a lot of preconditions. In this paper it is investigated if these preconditions are met in the case of rainfall noise or if they can be overcome be using certain techniques?

Geerlings, A.

• Acoustic services at Philips Applied Technologies
  A. Geerlings and B. Roozen [C]
[A]

Within the Philips community and beyond, Philips Applied Technologies (formerly CFT) provides acoustic services under the umbrella of our mechatronics activities. Combined with Philips ATC in Drachten we service most divisions of Philips Consumer Lifestyle, Healthcare and Lighting, as well as various technology groups inside and outside Philips, worldwide. Acoustics knowhow and a fast access to related technologies e.g. flow, control and thermal management, for a broad range of application domains, is available within the grounds of the Philips High Tech Campus. The type of products we work on, in joint effort with application experts in production groups, range from MRI systems to small surface mounted components. The acoustics therefore range from over 90 dB(A) to below 20 dB(A) in soundpower level. This requires proper measurement facilities, such as the reverberant room, the anechoic room and an acoustic imaging tool. Since acoustic solutions are often restricted by the application we need innovative analysis and breakthrough solutions, to improve the quality of the products under consideration. Operating closely with the application experts we provide a way for acoustic solutions to be designed into the application. The poster presented addresses some cases from the Philips product range and explains how new technologies are used for challenging acoustic cases.

• Source Identification and Noise Reduction of a Reciprocating Compressor at Grasso
  B. Roozen, J. van den Oetelaar, A. Geerlings and T. Vliegenthart [I]
[A]

Source identification is an important aspect of noise control engineering. A noise source can be very complex in nature. For control purposes it is rather important to identify the strongest contributing noise sources. This paper discusses the use of several structured methods on a reciprocating compressor in order to reduce the noise level significantly.

The measurement approach is quite logical, structured and practical, allowing an efficient identification of the contributing noise sources. Once the most contributing noise source is identified, the measurement technique also allows predicting the maximum achievable noise reduction of the machine as a whole and the way how to achieve this reduction. A BEM analyses shows areas for further optimizing the most significant noise source.

As a result, the sound power level of the reciprocating compressor is not only reduced with 7 dBA. Moreover, from a perceptual point of view the sound itself can be characterized as ''solid'' as well. Another 2 dB reduction is achievable.

Geier, M.

• Binaural Monitoring of Massive Multichannel Sound Reproduction Systems Using Model-Based Rendering
  M. Geier, J. Ahrens and S. Spors [I]
[A]

In model-based spatial sound reproduction a number of input signals is positioned in a virtual space according to a scene description. Using soundfield reproduction methods like Wave Field Synthesis or Ambisonics a scene can be rendered with a loudspeaker array. Suitable loudspeaker arrays typically consist of a huge number of loudspeakers and are often installed in heavily booked venues.

During the creation of an audio scene for playback in such a system it is desirable to listen to a representation of the scene via headphones without needing access to the actual loudspeaker system. This can save production resources because the target system is only needed for final adjustments. Alternatively, the same scene can also be monitored in a studio with much fewer loudspeakers using Vector Base Amplitude Panning.

We present different methods for reproducing a given audio scene via headphones, either by binaural simulation of a loudspeaker-based (possibly virtual) system or by binaural rendering of the original scene. Binaural simulation can be implemented efficiently by combining the loudspeakers' driving functions with Head Related Transfer Functions.

Gemmeke, H.

• Results of an Experimental Study for 3D Ultrasound CT
  N. Ruiter, G. Schwarzenberg, M. Zapf, A. Menshikov and H. Gemmeke [I]
[A]

Gemmeke, J.

• Evaluation of missing data techniques for in-car automatic speech recognition
  Y. Wang, R. Vuerinckx, J. Gemmeke, B. Cranen and H. van Hamme [I]
[A]

Background noise has a detrimental effect on the accuracy of automatic speech recognition (ASR) systems because it changes the statistical properties of the spectrally related features extracted from the speech. If the recognizer's acoustic model that describes these statistics is trained under noise conditions different from those of the test data, the likelihoods of speech hypotheses will deviate and recognition errors result. Many practical systems therefore compensate the spectral features for noise with techniques like spectral subtraction, modify the statistics by acoustic model adaptation or train the acoustic model under multiple noise conditions. In our ASR system which is based on missing data techniques (MDT), spectral regions that are dominated by noise are replaced by an estimate that depends on the speech hypothesis of the decoder. In this paper, we discuss the pro's and con's of this approach in terms of computational requirements, training data requirements and versatility. In a benchmark study using in-car recorded data, as well as other non-stationary noise types, we compare the accuracy of a MDT system with that of the commercially available Nuance VOCON 3200.

Genender, P.

• Acoustic Challenges of Rolling Bearings in Combustion Engines
  P. Genender, K. Wolff and C. Steffens [C]
[A]

In addition to the efficiency of gas exchange and combustion, fuel consumption is influenced to a large extent by engine friction. The distribution of mechanical losses shows that with one third, the main and conrod bearings have the largest share in engine friction apart from the piston subassembly. From a given 1,6 L 4-cylinder plain bearing engine changed to roller bearings a proved 5,4% (NEDC) improvement of the fuel consumption resulted from reduced friction. The challenges resulting from the use of roller bearings in internal combustion engines are related to the disciplines acoustics, durability, and fabrication. Within this article the main influencing factors on NVH will be identified by means of simulation and experimental investigations. It will be shown in which way the critical engine NVH as a matter of principle can be improved with the help of design modifications. Based on this knowledge, an advanced test engine was set up. Vehicle measurements, which have examined the development stages of the rolling bearing engine, verify the significant NVH improvements. Finally, interior noise level as well as psycho-acoustic parameters are on the same level as in the series production engine with plain bearings.

Genuit, K.

• The Potential and Benefit of Audible Noise Maps
  K. Genuit, A. Fiebig and P. Marla [I]
[A]

The importance of several (psycho-)acoustic parameters to noise evaluation and noise rating was demonstrated in numerous studies. However, in conventional noise maps neither information about further acoustic parameters nor comprehensive auralization options are available. In the context of the European research project ''Quiet City Transport'' the development of a traffic noise synthesizer was initiated, which deals with the idea of generating time signals representing specific traffic scenarios. On the basis of such generated time signals various acoustical parameters can be calculated and in consideration of sound propagation effects the noise for certain receiver positions can be determined. The advantage of noise maps providing audible examples of noise exposure in important immission points would be that experts and decision makers can listen to the environmental noise and can base their decisions with respect to required noise mitigation measures on a more significant basis. Of course, several criteria like accuracy, uncertainty, practicability, feasibility, interpretability of the maps must be critically discussed before considering the implementation of more extensive auralization tools in noise maps. The presentation will show some case study results. Moreover, the benefit of advanced noise maps will be highlighted and realization impediments will be discussed.

Gerhardt, N.

• Comparison of Coding Techniques for Photoacoustic Coded Excitation
  M. Mienkina, A. Eder, C.-S. Friedrich, N. Gerhardt, M.R. Hofmann and G. Schmitz [C]
[A]

In photoacoustics (PA) ultrasound is generated by laser irradiation of tissue. Usually, PA signals are generated by Nd:YAG lasers but it would be attractive to use less expensive laser diodes instead. Although laser diodes exhibit low pulse energy, the pulse repetition frequency (PRF) of pulsed laser diodes is much higher than the PRF of Nd:YAG lasers thus averaging can be used to increase the SNR. The applied PRF for averaging is limited by the acoustical time-of-flight. To further increase the SNR we propose to use coded excitation for PA imaging. In optical-time-domain-reflectometry, which is similar to PA imaging, Simplex Codes (SC) are widely used for coding. In ultrasound imaging Golay Codes (GC) are successfully employed. Therefore, we compare the performance of SC and GC for PA coded excitation (PACE) by a simulation study. For both codes, the coding gain normalized to equivalent averaging procedures is computed as a function of code length and PRF of the laser diode. Assuming a time-of-flight of 60 s and e.g. a PRF of 250 kHz the coding gain of GC is 1.6 dB higher than the coding gain of SC. Consequently, GC is more suitable for PACE than SC.

Gerl, F.

• Joint Speaker Identification and Speech Recognition for Speech Controlled Applications in an Automotive Environment
  T. Herbig and F. Gerl [C]
[A]

Gerretsen, E.

• Vibration reduction indices at junctions with cavity walls
  H. Martin, J. Smits, J. Niggebrugge, E. Gerretsen and L. van Luxemburg [C]
[A]

By applying concrete or brick cavity walls in dwellings instead of single homogeneous walls, the sound insulation between dwellings can be increased. However, then the role of flanking sound transmission is becoming more important. Flanking transmission depend partly on the properties of the building junctions formed by the cavity walls and other building elements. The properties of junctions made of single homogeneous building elements are known for some time. Few data are available of junctions of which cavity walls are part. In calculation models empirical modules are used for the vibration reduction indices of these specific junctions. Their accuracy is not known yet because of missing data from practice. Therefore, vibration reduction indices of several junctions with cavity walls have been determined by field measurements, also in dwellings under construction. For this purpose, a quick and reliable measurement method had to be developed. Also the invariance of the vibration reduction indices concerning the same junction in different surroundings has been investigated.

• Measuring Sound Insulation under Extreme Conditions using Deconvolution Techniques
  N. van Hout, C. Hak, E. Slaat and E. Gerretsen [C]
[A]

To compare the sound insulation performance of a building element with the given requirements, measurements have to be done. Generally, a broadband noise source is used, according to international standards. This method does not always work in practice due to high sound insulation values or high background noise levels. From a practical point of view it is very inconvenient or even impossible to perform an accurate sound insulation measurement for all frequency bands. A solution to this problem can be found in deconvolution techniques using MLS or sweep signals. It is possible to increase the signal to noise ratio using these techniques by averaging measurements and spreading out the spectral sound energy in time. As a result an efficient use of available sound power is possible. In a laboratory it can be investigated how to use MLS or sweep as a source signal and deconvolution as a measurement technique to obtain the sound insulation under noisy conditions. This has been investigated under moderately reverberant conditions, presented at the conference 'Acoustics'08 Paris', while this paper describes the same investigation under more extreme conditions.

• Model-based Assessment Scheme for Acoustic Quality Classes in Buildings
  E. Gerretsen [I]
[A]

In the past the acoustic quality in buildings was mainly specified by minimum requirements deduced from the performance of existing qualitatively good buildings. Later changes and additions were often only partly based on additional research and social surveys but mainly on 'political' negotiations. During the last decade the interest has grown in specifying classes of quality, especially for dwellings, preferably for the buildings as a whole. A class could than still be used as minimum - legal - requirement while at the same time a desirable better quality for the future can be clearly indicated and quantified. In order to avoid endless discussions about a decibel less or a decibel more, a model for the subjective assessment of the acoustic climate in buildings has been deduced from the available information. Discussion can than focus on the various aspects of the subjective assessment and their quantification and the final requirements simply follow from the model. This model has been used to specify acoustic quality classes in the Dutch standard for the assessment of sound reduction in buildings, NEN 1070. The presentation will focus on that model and its backgrounds.

• Sound and vibration transmission through lightweight junctions: Numerical investigation and validation by laboratory tests
  S. Lentzen, E. Gerretsen, C. van Bentum and S. Bron- van der Jagt [C]
[A]

In this work the sound and vibration characteristics of several lightweight junctions are numerically investigated with FEM (for the low- and the mid-frequency range) and SEA (for the mid- and the high-frequency range). In order to validate these analyses, the junctions are tested in a laboratory setting.
For sound insulation the investigations are focused on obtaining the vibration reduction indices . These are not only investigated for the floor-floor flanking transmission path, but also for all the other possible paths (e.g. floor-wall, floor-ceiling and wall-ceiling). Although applied in different frequency ranges, the results of the FEM and the SEA analyses agree fairly well with those obtained in the laboratory tests.
Looking at the vibration behaviour and transmission, the main focus was put on obtaining the One-Step-RMS velocities of the floors. In these investigations the point- and transfermobilities are acquired either by numerical analyses or by measurements. Since these spectra are only required for frequencies up to 40Hz, the numerical study consists merely of FEM analysis. Here, both results agree fairly well.
From the investigations in both areas, the appropriate conclusions concerning good building practices in lightweight buildings concerning sound and vibration can be drawn.

Gerstenberger, C.

• Acoustic Streaming in a Viscous Fluid-Structure System
  F. Böhnke and C. Gerstenberger [C]
[A]

The numerical evaluation of fluid-structure coupled systems in complex geometries, e.g. the acoustical wave propagation in the inner ear of mammals, requires elaborate mathematical formulations. One of these is the consideration of the stimulating boundary condition by an acoustic radiator (JASA, Bradley C.E. 1996), which leads to linear behaviour in Lagrangian but is nonlinear in an Eulerian description. In coupled systems the interaction between the structure guided Rayleigh wave and the fluid causes an internal acoustic streaming (AS). AS effects are a result of the nonlinearity of the Navier-Stokes equations in combination with the viscosity. For a numerical evaluation different discretizations are used. A complete Finite Element (FEM) approach is computationally too extensive at least when three-dimensional formulations are used. The fast BEM-FEM mortar coupling for acoustic-structure interaction was recently formulated but the acoustic field is governed by the Helmholtz equation and therefore viscosity was not considered. Another approach which is mainly used for the simulation of AS effects in compressible, viscous gas filled two-dimensional enclosures is the Flux-Corrected Transport (FCT) algorithm to solve nonlinear, time-dependent continuity equations applying a Finite Volume Method (FVM). Special care has to be taken with the formulation of the density boundary condition in this case.

Getzmann, S.

• Cortical processing of auditory motion cue information
  S. Getzmann and J. Lewald [I]
[A]

This study investigated the neural correlates of auditory motion processing by employing high-density electroencephalography. Sound motion was implemented by (a) gradual shifts in interaural time (ITD) or (b) level difference (ILD); (c) motion of virtual 3D sound generated using HRTF filters (derived from a set of KEMAR measurements); or (d) successive activation of 45 loudspeakers along the horizontal plane. In a subset of trials, listeners (N = 19) performed a 2A-FC motion discrimination task. Each trial began with a period of stationary acoustic stimulation from a central position, immediately followed by a period of motion starting from this location. The onset of motion elicited a specific cortical response that was dominated by large negative and positive deflections, the so-called change-N1 and change-P2. The temporal dynamics of these components depended on the auditory motion cues presented: Freefield motion and virtual 3D sound were associated with larger and earlier cortical responses and with shorter reaction times than shifts in ITD or ILD. Also, topographical analyses of the motion onset response indicated differences in the pattern of activation related to the different motion cues. The differences in cortical responses may allow conclusions on the processing of dynamic auditory information in the human brain.

Geyer, T.

• Silent Owl Flight: Setup for Flyover Noise Measurements
  C. Fritzsche, E. Sarradj and T. Geyer [C]
[A]

It is common knowledge that most genera of owls fly silently in order to be able to catch their prey. To investigate the mechanisms leading to that quiet flight, several basic fly-over measurements were carried out by biologists in the past. But the very low noise that is produced when the owls are in gliding flight makes it very hard to do exact acoustic measurements.

The presentation describes an attempt to do fly-over noise measurements of trained birds on natural conditions. It is focused on the measurement setup which consist of a small linear microphone array on a ground plate. The microphone array data is processed using a simple weighted-sum beamforming algorithm. Birds pass over the array in gliding flight at very low speed. A camera setup is used for flight speed measurement. A number of tests were conducted in a wildlife park. Preliminary results are presented from several raptor genera, including owls. Necessary improvements of the measurement setup will be discussed.

• Silent owl flight: experiments in the aeroacoustic wind tunnel
  T. Geyer, E. Sarradj and C. Fritzsche [C]
[A]

It is common knowledge that most genera of owls fly silently in order to be able to catch their prey. To investigate the mechanisms leading to that quiet flight, several basic fly-over measurements were carried out by biologists in the past. But the very low noise that is produced when the owls are in gliding flight makes it very hard to do exact acoustic measurements, especially when using the measurement techniques available at that time.

The presentation describes the measurements of the noise generation at the wings of silent and nonsilent flying birds in the aeroacoustic wind tunnel of the BTU Cottbus. The measurement setup will be explained, especially regarding the usability of conventional measurement techniques for experiments on wing specimen. A 56 channel microphone array is used to measure the noise generated at the wings in the fluid flow. The acoustic data is processed with advanced beamforming algorithms to obtain the radiated sound pressure levels and the location of the noise sources on the surface of the wings.

First results will be given that indicate that the magnitude, the spectral shape and the location of the noise produced by owls wings is different than those of nonsilent flying birds.

Gibbs, B.

• On force- and moment mobilities of a timber joist floor
  A.R. Mayr, B. Gibbs and H.-M. Fischer [I]
[A]

In wood framed buildings, floors are formed by fastening wood sheathing to joists spaced at a regular interval. The sheathing is typically fastened to the joist using screws so the resulting system is a complicated periodic point-connected plate-rib structure. It is shown that the point force mobility varies significantly with position. A machine installation generally is close to discontinuities, such as at: floor edges, joist-screw locations, joints between the sheathing plates, and due to workmanship. In such cases, moment excitation might become important at some frequencies and the neglect of moments a priori can lead to inaccurate prediction of the total emission. This paper presents measured point moment mobilities with respect to the distance to discontinuities for a timber joist floor where a single layer of chipboard forms the sheathing. It is shown that the measured point moment mobility indicates an infinite plate behaviour. This includes for positions above a joist but between screw positions. To determine the relative contribution of moments and perpendicular forces to the total structure-borne sound power, case studies of two sources, a fan unit and a whirlpool bath, are described, for various locations on a timber joist floor.

Gielen, L.

• Innovative Approaches to Fast Transfer Path Analysis
  H. van der Auweraer, P. Gajdatsy, K. Janssens, P. Mas and L. Gielen [I]
[A]

Over the last decennia, Transfer Path Analysis (TPA) has become an accepted tool for NVH troubleshooting and internal load estimation. A major bottleneck preventing its even more widespread use in the actual vehicle development process is the test time to build the full data model, requiring extensive Frequency Response Function testing next to the obvious in-operation tests. Recent innovations on instrumentation and new procedures such as Fast and Multilevel TPA and Operational TPA address this limitation, however often introducing new constraints. For example, OTPA only requires operational response data at reference and target locations. But, being basically a transmissibility method, its application requires special care regarding the inherent ill-conditioning and reference-completeness problems. It essentially assesses a co-existence relationship between targets and references and does allow causality and contribution interpretations. Hence a new approach is proposed, using a parametric model for the estimation of loads. This makes the method scalable, enabling the engineer to use a simpler model based on a small amount of measurement data for quick troubleshooting or increase accuracy by using a more complex model together additional measurements. The method is applied to a customer problem and compared with the existing approaches, showing the advantages in real life.

Gier, J.

• Modal analysis of a recorder: Experiment and simulation
  F. Möbius, J. Gier, T. Grothe and S. Marburg [I]
[A]

Herein, the authors present the investigation of an alto recorder. In a first step, a modal analysis of the structure of a recorder tunes the simulation model. It is possible to adjust material parameters to set up an eigenspectrum close to the one which has been measured in the experiment. However, some deviations remain. In a second step, the first resonance frequencies - which correspond to the played tone - are compared with the eigenfrequencies of a simulation model. Even without adjustment of parameters, the resonance frequencies and first eigenfrequencies of 16 tones are close together. Some unexpected effects of the tones of the particular recorder will be shown. Simulation reveals that the pressure magnitude in the duct does not form a half-wave as expected. It is one result of this project that it can be clearly shown that structural modes will hardly affect the tonal spectrum of the fluid.

Giering, K.

• Dose-Response-Relationships (DDR) and the Environmental Noise Directive (END)
  K. Giering [C]
[A]

A Review of newer developments in dose-response relationships is given. It will be argued, that unique curves, applicable for all European countries are hardly to establish.

Gierlich, H.W.

• Challenges in the introduction of wideband hands-free in cars
  H.W. Gierlich [I]
[A]

The introduction of wideband speech services in the new generation networks will be one of the major steps in the near future. The support of such services by the different networks and services also forms the basis for the integration of wideband speech hands-free services in cars. Wideband speech quality would be mostly welcomed in cars since especially on the receiving side the poor (narrowband) speech quality is more obvious in cars than in other type of terminals. The narrowband speech is compared directly to the typically high quality audio presentation of all other audio equipment in the car and even less sensitive users ob-serve the difference in quality immediately. To achieve a superior speech quality inside the car all elements involved in the transmission have to provide adequate speech quality. The challenges faced when connecting the hands-free car system via Bluetooth to the mobile terminal, via the mobile terminal to the network and connecting the mobile network to the wideband NGN type network is introduced, the users expectation with respect to the different quality dimensions of wideband speech and the wideband presentation of narrowband speech signals is discussed. An overview of the standardization work in the ITU-T Focus

Giesler, J.

• Broadband noise investigation on rod-airfoil-configurations
  J. Giesler and E. Sarradj [C]
[A]

One important broadband noise generating mechanism in axial fans is the interaction between the turbulent inflow or a turbulent wake of a rotor vane and the leading edge of the downstream stator vane. Additionally, the interaction of the turbulent vortices with the trailing edge of the airfoil generates noise as well. These mechanisms are the subject of an experimental research at BTU Cottbus.

The setup for measurements in the aeroacoustic wind tunnel consists of a rod and an airfoil mounted downstream. The geometric parameters of the rod-airfoil configuration, namely the cylinder diameter, the gap between rod and airfoil and the airfoil geometry are variable. For an improved flow quality in the measurement area of the wind tunnel the configuration is mounted between two porous sound absorbing side plates, directly attached to the nozzle exit. Acoustic measurements are done using an array of 38 flush-mounted microphones. The flow is characterized using constant temperature anemometry.

The presentation deals with the details of the experimental setup as well as with a first subset of experimental results. Moreover, the influence of the parameters on the leading and trailing edge noise generation is discussed.

Goehrke, A.

• Psychoacoustic Evaluation of Traffic Noise
  A. Fiebig, S. Guidati and A. Goehrke [I]
[A]

In the European research project ''Quiet City Transport'' the quantitative characterization of annoyance caused by traffic noise using psychoacoustic descriptors for a perception-related environmental noise description was a major objective. In the first project stage single pass-by noises have been evaluated by subjects. The aim was to discover components and patterns, which are responsible for specific annoyance effects. The results showed that certain noise properties such as diesel knocking are particularly annoying regardless of their absolute sound pressure levels. In the second part of the project complete traffic noise scenarios were evaluated to detect indicators which allow for an adequate assessment of environmental noise quality. Using a traffic noise synthesizer as well as artificial head recordings a wide variety of stimuli were created and evaluated by subjects. By means of a principal component analysis links between specific noise properties and subjective evaluation patterns were found and provided the basis for the development of a metric. The results of the study dealing with the evaluation of traffic noise will be presented and discussed under the perspective of their applicability in environmental noise policy and for soundscape studies.

Gömmel, A.

• An approach for numerical calculation of glottal flow during glottal closure
  A. Gömmel, M. Kob, T. Niendorf and C. Butenweg [I]
[A]

Limitations of classical low-degree-of-freedom approaches (multiple-mass models) of modeling vocal fold (VF) oscillation have indicated the need for more advanced models. A couple of finite-element (FE) models have been presented in the last years, which include a number of numerically difficult solver algorithms. The presented model represents a coupled two-dimensional model of VF movement including fluid-structure interaction effects. A structural model of vocal fold tissues and a fluid model of the airflow are combined by the Arbitrary Lagrangian-Eulerian formulation. As solvers ANSYS was chosen for the structure and CFX for the fluid. A special problem is handling the stop of the airflow in the closed phase of the glottal cycle. Since the fluid elements in the glottal gap channel would have to be zero-volume elements, a special modeling strategy is necessary. In this model a loss coefficient is introduced when the glottal gap falls below a certain control volume. As exemplary area of use, results of comparative simulations are shown, which differ in driving pressure values and VF shape. It is shown that typical convergent/divergent shapes are only obtained when a minimum length of a channel between the VF remains.

Goenner, L.

• Speech intelligibility prediction for normal hearing and hearing impaired subjects – MCHI-S
  L. Goenner and J. Haubold [C]
[A]

Goerick, C.

• Combining Auditory Inspirations and Hierarchical Feature Extraction for Robust Speech Recognition
  M. Heckmann, X. Domont, F. Joublin and C. Goerick [C]
[A]

Goetze, S.

• Estimation of the Optimum System Delay for Speech Dereverberation by Inverse Filtering
  S. Goetze, M. Kallinger, A. Mertins and K.-D. Kammeyer [C]
[A]

Equalization of room impulse responses is an straightforward approach for dereverberation of speech signals in a hands-free scenario. The well-known least-squares equalization filter minimizes the euclidian distance between the concatenated system of room impulse response (RIR) and equalizer and a given target system. This target system usually is chosen as a delayed discrete pulse, a delayed band-pass or a delayed high-pass filter. In this contribution we address the choice of the this delay which has to be introduced w.r.t. maximum dereverberation performance. Since designing one equalizer (EQ) for each possible delay and choosing the best one is computationally inefficient we investigate the dependence of the optimum EQ delay on various parameters describing RIRs. A high correlation was found between the so-called central time of the room impulse response and the optimum EQ delay. Since the central time can be determined based on estimates of the initial peak of the RIR and the room reverberation time, we propose to use a very short filter for system identification and an estimate of the room reverberation time to identify the optimum equalizer delay.

Golliard, J.

• Towards Industry-Standard Flow-Induced Noise Prediction
  G. Guilloud, P. Martinez, J. Golliard and C. Schram [I]
[A]

Engineers are nowadays predicting flow-induced sound in order to obtain quantitative assessment of complex devices in industries that now encompass air and ground transportation, energy sectors or domestic appliances. However, a fair amount of know-how still has to be captured to reach the decibel-accuracy sought by industrials for realistic configurations. A particular concern is the assessment of the numerical requirements of both the flow and acoustic modeling, in relation with the desired prediction accuracy over a pre-defined frequency range.

The present work is focused on the hybrid approach inspired from the acoustic analogy. A crucial issue regards the reliability of the flow data. In that respect, designing a mesh and selecting a numerical scheme for a given targeted noise frequency range can be complicated. A related diffculty lies in the data transfer that takes place from the CFD code to the acoustic tool. The problem consists in transferring data from a fine mesh towards a considerably coarser mesh without significant loss of information. Hereto, conservative interpolation schemes are applied to optimize the effciency of the process. The optimized procedure is applied to the case of a rectangular diaphragm fitted in a square duct, for realistic Reynolds and Mach numbers.

Goossens, S.

• Loudness perception with headphone presentation compared to loudspeaker presentation in the diffuse field
  S. Goossens, G. Bonin and R. Stumpner [C]
[A]

The same sound pressure level in the ear-canal can lead to different loudness perception for headphone and loudspeaker presentation. This effect, which is known for a long time, was examined once more for different headphone types and signal presentations. During these tests the subjects were sitting in the reverberation chamber and had to adjust a third-octave band noise to the same loudness caused by loudspeaker and headphone reproduction. The headphone reproduction was carried out in a monotic, diotic and dichotic way. Significant differences of sound pressure levels had been measured in the ear-canal for the monotic and diotic case although the same loudness had been perceived. In contrary the level differences disappear largely for a dichotic stimulation. This is valid both for a binaural noise corresponding to the loudspeaker signal with spatial information, which was taken by artificial head recordings in the reverberation room and for a simple incoherent noise without spatial information at all. In the dichotic case equal levels in the ear-canal had been measured for the same loudness.

Grabinger, J.

• Prediction of Load Noise of Power Transformers by 3D FEM and Comparison to Measurements
  A. Hauck, M. Meiler, R. Lerch and J. Grabinger [C]
[A]

In absence of reliable theoretical formulas for noise emitted by electric power transformers, finite element simulation play an important role in the design of low-noise transformers. A major challenge in developing a FE model is to resolve the small scale windings structure, as well as the tank with the oil and the surrounding air.

In this work we make us of a two-scale homogenization approach for calculating the mechanical deformations of the windings by exploiting the periodicity of the winding structure. In a first step, a unit cell (circular ring) problem is solved on a very fine grid resolving the winding structure, which yields the mechanic eigenmodes. On a very coarse model of the complete setup, these eigenmodes are augmented with the standard finite element basis functions to model the homogenized winding structure.

The full structure-acoustic interaction of the windings, the oil and the tank is taken into account, as well as the acoustic radiation into air. The applicability of our method is demonstrated for a 3-phase power transformer. Comparisons of computed the sound pressure level, as well as the acoustic power with measurements are performed and show good agreement.

• Aeroacoustics taking Fluid-Structure Interaction into Account
  M. Kaltenbacher, S. Becker, J. Grabinger and F. Schäfer [I]
[A]

In the last years, manufacturers have started to consider the aerodynamic noise level in many industrial applications as a relevant design parameter (e.g. airplanes, cars, air conditioning systems, etc.). Because of the growing demand for reducing noise levels and for fulfilling noise regulations, there is a great motivation to investigate the basic aeroacoustic phenomena and mechanisms of sound generation and propagation. In many cases, the occurring noise is not only generated by the turbulent flow but also due to vibrations resulting from the interaction between the flow and the mechanical structure.

In the present work, we investigate the three field coupling for a flow past a thin flexible structure. Our developed numerical algorithms are capable to simulate the complex interaction between fluid, structure and acoustics. The scheme is based on a partitioned approach employing a finite-volume flow solver and a finite-element structural mechanics and acoustics solver. An important feature of the computational scheme is that it allows for a separate prediction of flow-induced and vibration sound. We will present detailed numerical studies of the interaction of the fluid flow with the flexible structure and the generated noise. Moreover, a comparison of computational results to our experimental investigations will be provided.

• Separation of Aero- and Vibroacoustic Noise from a Flexible Structure
  J. Grabinger, S. Müller and S. Becker [C]
[A]

The reduction of noise due to turbulent flow over a car's underbody is an important--yet complicated--issue in automotive engineering. Our objective is to study the mechanisms of noise generation at a simplified model of a car's underbody. For this purpose we mount a flexible rectangular plate into a rigid surrounding plate, which is then placed in a wind tunnel in parallel to the main flow direction. A square cylinder is mounted upstream of the flexible plate in order to model a strut in an underbody. This assembly generates noise by means of two mechanisms, namely the vibrations of the flexible plate and the turbulent flow.

These two mechanisms require different counter measures for noise reduction. Therefore, one has to separate the aeroacoustic and vibroacoustic contributions in the radiated sound. We measure the sound pressure with a microphone and, simultaneously, the normal deflections of the plate with a laser scanning vibrometer. On the basis of these data, we conduct FEM simulations to compute the vibrational sound at the position of the microphone. Subsequently, we estimate the contribution of the vibrational sound by comparing it to measured data. We find that the vibration generates dominant tonal components at the plate's resonance frequencies, while the flow-induced noise gives a broadband contribution.

Graf, J.

• Feedback Control for Active Noise Reduction in Headsets
  J. Graf and E. Reithmeier [C]
[A]

With regard to production costs, control strategies for commercial ANR-headsets (Active Noise Reduction) have to be realized simple and efficient. Therefore, the implementation of computationally expensive algorithms, which require fast and expensive digital signal processors, is avoided. Besides an expensive DSP, further production costs are caused by a reference microphone, which is necessary in case of a feedforward controller. However, in contrast to a feedforward controller, the feedback control approach does not require such a reference microphone. Thus, the realization of a feedback control strategy enables to save the reference microphone as well as the associated analog input circuit respectively. In order to design an efficient ANR-system, in this paper a time discrete feedback control approach is introduced. Basically, the proposed feedback control loop consists of two parts. One part is formed by the non-adaptive controller, while the other part consists of an adaptive algorithm. The non-adaptive controller guarantees steady-state noise reduction and the overall performance of the ANR-system is improved by the adaptive algorithm. The design of the control system is described and the active noise reduction performance is presented.

Graf-Langheinz, A.

• Earprotection Project Military Music
  M. Logar-Holzer, M. Emich, W. Wahler, C. Gruber, A. Graf-Langheinz and S. Spendelhofer [C]
[A]

The past experiences in the Austrian military music band with the adaption of a customized earprotectors have so far been unsatisfactory. Different aspects felt negatively by the musicians (subjectively) lay particularly in the range between the false interpretation of the sound image and/or the difficulty in the assessment of the dynamics, volume and Intonation of the sounds/tones.This observation study was conducted in three representative Austrian Military Music Bands. The results were gained out of a total of 146 persons by means of a standardised questionnaire in various subjects with regard to personal handling and the use, deployment and acceptance of the personal protection of the ears.The daily length of music performance was documented during the period of 8 weeks. Parallelly volume decibel measurements in different practical-relevant sound situations of the musicians (orchestra, single exercises,...) were gathered.

Granneman, J.

• Web-based Toolbox Noise and Provisions in Metal Working Industry
  J. Granneman and J. van Hees [I]
[A]

Employees in metal working industry are often exposed to high noise levels. European noise limits regarding occupational noise are related to the daily dose of workers. In this type of industry the daily dose is hard to obtain by measurements due to great variety in activities and equipment. A specific web-based toolbox has been developed to enable firms to determine the daily dose of their workers. Based on an estimate of activities and their duration this toolbox derives the daily dose and highlights the most dominant sound sources and activities. In the toolbox also the effect of room acoustics and contributions due to activities of co-workers are taken into account. If noise limits are exceeded, the toolbox provides specific measures for each activity to reduce the noise levels, giving detailed information about the amount of reduction of the daily dose to be expected. The web-based aspect of the toolbox makes its content easy accessible and up-to-date. The toolbox is part of a program issued by the government and branch organisations that also deals with other occupational aspects. By simple means the toolbox can also be used for other branches dealing with occupational noise problems.

• Whistling Building Objects, Origins and Solutions
  J. Granneman, M. Trooster, E. de Beer and T. van Diepen [C]
[A]

Widespread complaints often occur due to whistling sounds caused by building objects, such as (high rise) buildings and viaducts, under specific meteorological circumstances, especially wind speed and direction. In many cases this is related to the application of specific grating. Much is already known about the relevant parameters for this phenomenon from measurements in practice and in laboratories. To obtain more understanding into this subject and to be able to prognosticate the possible occurrence and intensity of it in the design phase of building objects, systematic research has been done under laboratory conditions. The sound power level and frequency characteristics of whistling sounds dependent of wind speed, angle of incidence and grid configurations were studied. Also the sound reducing effect of different and practically feasible modifications of gratings was part of the study. The paper presents the results of this study.

Granström, B.

• Noise reduction for media streams
  T. Bisitz, T. Herzke, M. Zokoll, A.-M. Öster, S. Al Moubayed, B. Granström, E. Ormel, N. van Son and R. Tanke [C]
[A]

In the Hearing at Home project the central Home Information and Communication platform is developed to support hearing-impaired persons with easy to configure ``Supportive Audio Signal Processing'' and visual support on a TV screen in home environments. Amongst other, individually configurable algorithms, several algorithms can be used for noise reduction of the media streams without individual fitting to enhance the accessibility of the presented audio materials using the Master Hearing Aid as a real-time audio-signal-processing framework: (1) Single-channel noise reduction according to Ephraim-Malah together with a speech pause detector, (2) stereo noise reduction additionally removing the central signal in a control path, (3) frequency shaping noise reduction by emphasising a frequency range important for speech intelligibility, (4) time-varying attenuation and (5) noise reduction by selection of channels. It will be possible to classify the acoustic situation in the media stream automatically in real-time using six classes allowing to activate the appropriate noise reduction strategies. In user tests the performance of algorithm 1-3 is evaluated: Results of sound quality assessment, listening effort estimation and speech intelligibility measurements with hearing impaired subjects from groups with different types of hearing impairment will be presented.

Greßkowski, J.

• Concepts of Global Noise Control in Cabins
  J. Greßkowski, D. Sachau and S. Böhme [C]
[A]

Nowadays people have to be protected for high sound pressure levels at low frequencies. An example is a propeller driven aircraft, where the propellers induce periodic noise inside the cabin. Therefore active noise reduction systems are applied to reduce these unwanted disturbances. To reach an acceptable noise level in the whole cabin actuators and sensors have to be allocated global in the distributed area. Also large numbers of actuators and sensors are required for attenuation of unwanted sound in a large volume with a high density of acoustic modes. This results in a high computational load for the controller. In order to reduce the number of components and consequently the effort of the controller unwanted sound transmission in cabin can be reduced in the propeller plane by controlling the radiated sound power. This innovative concept is investigated using an A380 structure as primary source which radiated noise due to excitation. The control sources are allocated in front of the panel. Also the microphones were placed in the vicinity of the structure. The possibility and first experimental results is shown in this research.

Grigoras, C.

• Speaker Verification Based on Formants Using Gaussian Mixture Models
  T. Becker, M. Jessen and C. Grigoras [C]
[A]

In a new method, the well known UBM-GMM framework for automatic speaker verification is applied to formant parameters replacing the common cepstral parameters (e.g. Mel frequency cepstral coefficients (MFCC)). By using full covariance matrices, both the static and variable vocal tract configurations are reflected in the speakers' models. Speaker comparisons are achieved by relating speaker similarity to speaker typicality, expressed as the likelihood ratio. The method reveals a low equal error rate while providing easy interpretability and low dimensionality.

Grosche, P.

• Combination of Onset-Features with Applications to High-Resolution Music Synchronization
  P. Grosche, M. Müller and S. Ewert [I]
[A]

Many different methods for the detection of note onsets in music recordings have been proposed and applied to tasks such as music transcription, beat tracking, tempo estimation, and music synchronization. Most of the proposed onset detectors rely on the fact that note onsets often go along with a sudden increase of the signal's energy, which particularly holds for instruments such as piano, guitar, or percussive instruments. Much more difficult is the detection of onsets in the case of more fluent note transitions, which is often the case for classical music dominated by string instruments. Here, one has to develop methods that allow for detecting smooth temporal and spectral changes in the audio signal. In this paper, we present various methods for onset detection including energy-based and phase-based techniques, which account for different properties of the audio signal. In particular, we present a recent audio feature that combines the high temporal accuracy of onset features with the robustness of chroma features. Finally, we show how a combination of various onset features can be used to significantly improve the temporal resolution in music synchronization tasks.

Groß, L.

• Analysis of the Flow Field of Rotating Milling Cutters in Idle Mode
  L. Groß and U. Heisel [C]
[A]

Rotating tools are generally assumed to be one of the major sources of noise in machines. Vortices shedding from the rotating tool may cause turbulent flow fields. This article concentrates on the flow field of milling cutters in idle mode. At first, measurements obtained by a scanning laser Doppler vibrometer are analyzed. The test object is a milling cutter for T-slots. The test reading shows turbulences caused by vortices only in the chip space of the milling cutter and the near field of the cog. Subsequently, the flow field of the same milling cutter is simulated by ANSYS CFX. The simulation confirms the test reading obtained by the scanning laser Doppler vibrometer. Additionally, a different type of milling cutter is tested in water. The turbulence caused by this milling cutter occurs again only in the chip space. Therefore different geometries are simulated to detect relevant parameters causing turbulences.

Groß-Thebing, A.

• Numerical Method for the Prediction and the Assessment of Rolling and Curving Noise
  A. Groß-Thebing and H. Zimmer [I]
[A]

The program SFE AKUSRAIL aims for the simulation of the whole chain leading to noise in railway operation. The tool generates models predicting the high frequency interaction between vehicle and track. It consists of modules regarding track and wheelset dynamics, W/R contact mechanics, rolling noise, curve squeal and interior noise prediction. The combination of track and wheelset dynamics, contact mechanics and vehicle structure dynamics represents their interaction taking into account physical properties of geometry, material and operational conditions. For the prediction of curve squeal two independent prediction methods filter the curve squeal frequency from all the possible wheel resonances. The methods show the sensitivity for variations of wheel geometry, for application of wheel absorbers and for material properties of the rubber in resilient wheels. An automatic parameter variation performs instability regimes for the linearized differential equations of the wheelset - track dynamics considering a contact point shift on wheel and rail tread. Wheel and rail amplitudes are calculated in dependency of the frequency and of the position in curved track using time step integration of the complete nonlinear model. The prediction method is validated in a research project funded by the Bundesministerium für Wirtschaft und Technologie (19U5001 B).

Groth, S.

• Adaptive Adjustment of the “Sweet Spot” to the Listener’s Position in a Stereophonic Play Back System – Part 1
  S. Merchel and S. Groth [C]
[A]

The spatial reproduction of sound in a conventional stereo system works in a small area which is located on the symmetry axis between the loudspeakers - the so called ''sweet spot''. Beyond this area, the spatial perception collapses and the stereo image moves to the nearer loudspeaker since the signal arrives both louder and sooner. Finally, the stereo image is completely located in the nearer loudspeaker due to the precedence effect. However, in desktop applications or virtual environments, image and sound have to correlate even though the listener is out of the ''sweet spot''. Although stereophony has a long history, the listener has not yet been released from its static hearing position. Most methods try to broaden the area of stereophonic perception by the adjustment of loudspeaker characteristics. This leads to contradicting localization cues. This paper presents a playback system which adjusts both loudspeaker signals depending on the listener's position in real-time. In this study, the x-y position of the listener is tracked by a stereo camera and a face recognition system. This information is used to calculate the amount of delay and amplification in the corresponding audio channels. Thereby, a correct localization over the whole off-center listening area is achieved.

• Adaptive Adjustment of the “Sweet Spot” to the Listener’s Position in a Stereophonic Play Back System – Part 2
  S. Groth and S. Merchel [C]
[A]

One major disadvantage of stereophonic play back systems is the narrow ''sweet spot'' in which correct audio localization is possible. Due to this limitation, the listener's freedom of movement is drastically restricted. The present article focuses on a system designed to adaptively adjust the ''sweet spot'' to the listener's position. The primary concern is the analysis of sound localization during the adjustment of the ''sweet spot'' via time delay and level differences. Therefore, different models of binaural hearing are used to validate the utility of the system. As a first approximation, an analytical approach by Lipshitz (1986) is used. In a more advanced approach, a binaural model after Braasch (2005) is applied. Both models lead to the same conclusion: They show that an adjustment of stereo signals allows a good localization over the whole off-center listening area. The results of several simulated hearing and play back situations are plotted in quality and vector maps to clearly show the effectiveness of a system with combined time delay and level adjustment. Furthermore, different stereo recording techniques are evaluated regarding their suitability for play back systems with signal adjustment. Thereby, intensity stereo turns out to be most favorable.

Grothe, T.

• Effect of the structural dynamics of the bocal on the sound spectrum of a bassoon
  T. Grothe, J. Baumgart and R. Grundmann [I]
[A]

The bocal of a bassoon is a thin curved metal tube to which a double-reed mouthpiece is attached. When the instrument is played, the reed acts as a pressure controlled valve and due to the pulsating flow the bocal starts to vibrate. The outer contour of the bocal has a major influence on the modal shapes of the structural modes and the relation between the eigenfrequencies. The matter of interest in this investigation is the interaction of the structural dynamics of the bocal and the reed. We assume that modifications of the bocals geometry and mass distribution will lead to measurable changes in the sound spectrum. Therefore the modes of a bocal are computed by FEM and the frequencies are checked experimentally with a LDV. In a playing experiment mass was added locally on the bocal. Sound pressure measurements were performed, while a professional bassoonist played sustained notes on a modifiable bocal. Very small changes, less than 4 dB SPL, occur in the overtones above 1 kHz and can be shown using principal components analysis of the sound spectra.

• Modal analysis of a recorder: Experiment and simulation
  F. Möbius, J. Gier, T. Grothe and S. Marburg [I]
[A]

Herein, the authors present the investigation of an alto recorder. In a first step, a modal analysis of the structure of a recorder tunes the simulation model. It is possible to adjust material parameters to set up an eigenspectrum close to the one which has been measured in the experiment. However, some deviations remain. In a second step, the first resonance frequencies - which correspond to the played tone - are compared with the eigenfrequencies of a simulation model. Even without adjustment of parameters, the resonance frequencies and first eigenfrequencies of 16 tones are close together. Some unexpected effects of the tones of the particular recorder will be shown. Simulation reveals that the pressure magnitude in the duct does not form a half-wave as expected. It is one result of this project that it can be clearly shown that structural modes will hardly affect the tonal spectrum of the fluid.

Gruber, C.

• Earprotection Project Military Music
  M. Logar-Holzer, M. Emich, W. Wahler, C. Gruber, A. Graf-Langheinz and S. Spendelhofer [C]
[A]

The past experiences in the Austrian military music band with the adaption of a customized earprotectors have so far been unsatisfactory. Different aspects felt negatively by the musicians (subjectively) lay particularly in the range between the false interpretation of the sound image and/or the difficulty in the assessment of the dynamics, volume and Intonation of the sounds/tones.This observation study was conducted in three representative Austrian Military Music Bands. The results were gained out of a total of 146 persons by means of a standardised questionnaire in various subjects with regard to personal handling and the use, deployment and acceptance of the personal protection of the ears.The daily length of music performance was documented during the period of 8 weeks. Parallelly volume decibel measurements in different practical-relevant sound situations of the musicians (orchestra, single exercises,...) were gathered.

Grubeša, S.

• Evaluation Of Acoustical Suitability Of Lecture Rooms With Respect To The Quality Of The Sound Reinforcement System And The Level Of Background Noise
  M. Horvat, H. Domitrovic and S. Grubeša [C]
[A]

In order to evaluate acoustical suitability of several lecture rooms located on the Faculty of EE and Computing in Zagreb, Croatia, a questionnaire was given to the students who take their classes in those particular rooms. The first part of the investigation is focused on the quality of sound reinforcement system located in two largest lecture halls, to be replaced in the near future with a new one of significantly better quality. The goal is to examine whether this change is reflected in the grades of acoustical suitability given by the students. The second part of the examination deals with two smaller lecture rooms of identical size and acoustic treatment, but with significantly different levels of background noise due to positions of these rooms relative to the nearby street as the source of traffic noise. The overall level and spectral content of noise are measured in these lecture rooms. In addition, speech intelligibility is examined in all four rooms. The results of these measurements are then interpreted with respect to the subjective grades given by the students.

Grüner, R.

• Identification of a Low-Frequency Sound Source in an Aircraft Engine Test Stand
  W. von Heesen, A. Péus and R. Grüner [C]
[A]

The tenants of an office building complained about strong low-frequency noise from a nearby aircraft engine test stand. The test stand resembles an open-circuit wind tunnel with the aircraft engine located at the end of a horizontal inlet duct. The engine discharges into a horizontal pipe with a conical insert at the inlet. The cross sectional area of the discharge pipe is substantially smaller than that of the inlet duct. Twelve microphones were installed inside the test stand, at the inlet and the outlet of the test stand, and at the office building. By analyzing the phase and the cross-correlation between the microphones the direction of sound propagation could be determined and the point of origin of the low-frequency sound could be identified. It turned out that the low-frequency sound is generated by the turbulent jet impinging on the conical inlet of the outlet pipe and propagates against the flow into the inlet duct. By decreasing the outer diameter of the conical inlet at the discharge pipe the sound level could be reduced. The main focus of the paper is the description of the data analysis which lead to the identification of the sound source.

Gruhler, G.

• Introduction of a Laboratory Technical Acoustics at Heilbronn University (Germany)
  G. Landstorfer, A.S. Treiber and G. Gruhler [C]
[A]

Currently, the education in acoustics at Heilbronn University is only representing by a lecture technical acoustics (2 ECTS) that gives each winter semester an overview of acoustic issues. Under the aspect of expanding the acoustic course offer, a laboratory technical acoustics (3 ECTS) will be introduced for prospective summer semesters.

The most important requirements for the laboratory concept are to achieve the following aims:

a) The students' comprehension of basic acoustical phenomena shall be a focus, b) The experiments shall deal with actual problems close to our areas of research, as well as general tasks given from surrounding industrial companies, c) The expansion of knowledge in our core competence fields with student support in projects, bachelor or master thesis';

According to the number of expecting participants in conjunction with available locations (anechoic measuring room, jury test room, echo corridor, auditorium, laboratory electronic circuit technology), groups of two or three students may work for their own of a total of six different experiments.

It will make sense to correlate the topics of the laboratory with the topics of the lecture. Therefore, the practical tasks will be located in psychoacoustics, room acoustics, electro-acoustics, acoustic quality control, multi-channel playback/recording and passive/active noise control.

• Acoustic Optimization of Rotary Switches
  A.S. Treiber and G. Gruhler [C]
[A]

Due to the complexity of modern car's on-board systems car manufacturers tend to use menu based user interfaces which offer a relatively low number of control elements for a large number of functions.

Since cars nowadays are no longer sold simply as technical but as lifestyle products it is crucial for the success that the potential customer perceives every single aspect of the car to be valuable. Since the user interface of the car can be judged even before a test drive the feeling of buttons and switches is a key aspect. This feeling as affected by both haptical and acoustical feedback of operation.

This work describes an enhanced interactive simulator for the acoustic feedback of control elements. The simulator allows the subjective rating of recorded feedback sounds of control elements without visual or haptic cues using semantic differentials or paired comparisons as well as the manipulation of stimuli by the test subjects. Furthermore, results obtained using this device are presented.

Gruhn, R.

• Survey on common Arabic Language forms from a speech recognition point of view
  M. Elmahdy, R. Gruhn, W. Minker and S. Abdennadher [C]
[A]

Arabic language is the largest still living Semitic language based on the number of speakers which exceeds 250 million first language speakers. Arabic language has two main forms: Standard Arabic and Dialectal Arabic. Standard Arabic includes ''Classical Arabic'' and ''Modern Standard Arabic (MSA)'' while dialectal Arabic includes all forms of currently spoken Arabic in day life and it vary among countries and deviate from standard Arabic to some extent. While there are many forms of Arabic, there still many common characteristics on the acoustic level and the language level. In this paper we investigate the main differences between Arabic forms from a speech recognition point of view to focus on similar and different acoustic properties among Arabic forms as well as the language properties. For standard Arabic we have chosen both classical and MSA forms and for dialectal Arabic we have chosen Egyptian colloquial Arabic as an example since it is the most popular dialect among Arabic speakers. The main purpose of this paper is to summarize all characteristics of Arabic and previous research results in Arabic speech recognition from a speech recognition point of view in one paper so researchers in this field can use it as one reference.

• Comparing Acoustic Model Adaption Methods for Non-native Speech Recognition
  H. Lang, M. Raab, R. Gruhn and W. Minker [C]
[A]

In order to enable voice control for in-car navigation- and entertainment-systems the necessity of recognizing non-native speech arises. Examples are Germans uttering destinations in France or selecting English audio tracks.

When confronted with non-native speech, recognizers trained on native speakers do not achieve high accuracy on such a task.

The explosion of possible native and foreign language combinations and - as a consequence thereof - the lack of databases covering appropriate non-native speech, make the training of separate models unfeasible.

A common approach for non-native adaption in speech recognition is to utilize a baseline model trained on native speech of target language and to perform some kind of model reestimation - utilizing a small set of utterances in spoken by native speakers of the source language .

Due to the necessary data collection effort, most previous work was limited to evaluate the results of this reestimation on one or two different accents.

With the help of new non-native databases that have recently become available, we can fully focus on evaluating and comparing different reestimation techniques on different language pairs. Our results show how stable different techniques are when they are applied to different language combinations.

• Optimal projections between Gaussian Mixture Feature Spaces for Multilingual Speech Recognition
  M. Raab, O. Schreiner, T. Herbig, R. Gruhn and E. Nöth [C]
[A]

Multilingual speech recognition is increasingly gaining attention for in-car speech controlled applications. An example is a media player that allows selection of music by voice command, requiring speech recognition for multiple languages in order to cover the languages of artist names and music titles in a given music database.

There are two traditional approaches how a system can support the recognition of multiple languages. The first is to run a set of monolingual recognizers in parallel; the second one is to train an integrated multilingual recognizer dataset for the required set of languages.

The first approach has the disadvantage that there is no parameter sharing between the different recognizers, thus needing large amounts of processing power and memory. The second approach has the disadvantage that one multilingual recognizer has to be trained for every combination of languages.

In our paper we present a scheme to create a multilingual recognizer out of monolingual trained recognizers. A formula is given for an optimal projection of emission probabilities between Gaussians Mixture feature spaces. With this formula, we can project each HMM state of all languages to one set of Gaussians without retraining the system.

• Improving Context Modeling for Phoneme Recognition
  D. Vasquez, G. Aradilla, R. Gruhn and W. Minker [C]
[A]

In this paper we investigate a novel technique for modeling long temporal context in phoneme recognition. Conventional techniques take a global context consisting of consecutive concatenated acoustic features as input for a phoneme classifier. Thus, the classifier must learn different variations corresponding to different parts of the global context. In contrast to conventional techniques, our scheme decomposes the temporal context of each phoneme into a set of slides. Each temporal slide is input for a non-linear classifier given by a Multilayered Perceptrons (MLP). Every MLP is trained using the central label of the global context. Thus, the assumption that the current phoneme occupies the global context is preserved. Furthermore, each MLP can robustly model different temporal context inside each phoneme. The outputs of the classifiers are combined to estimate posterior probabilities of phoneme-level classes. These posteriors are employed in a hybrid HMM/MLP framework. Experiments have shown an absolute phoneme error reduction of 3.6% compared to a baseline classifier with the same context length. A further improvement is achieved when a hierarchical approach is implemented, which introduces information across different phonemes.

Gruhne, M.

• Preprocessing methods for rhythmic mid-level features
  C. Dittmar, M. Gruhne and D. Gaertner [I]
[A]

Rhythmic mid-level features are an important pre-requisite in Music Information Retrieval. They can be deployed to describe the rhythmical gist of songs for various tasks, such as genre classification and music similarity search. Although different computation strategies have been proposed in the literature, rhythmic mid-level features commonly represent the most salient rhythmic periodicities in a music signal. Major shortcomings of rhythmic mid-level features are their dependency on the actual tempo of the songs and their susceptibility to degradation caused by interference with non-rhythmic signal components, such as melodic-sustained instruments. This publication describes preprocessing strategies applicable to rhythmic mid-level features. The problem of interference with melodic instruments is addressed by emphasizing the rhythmic content via a signal decomposition related to Drum Transcription. Tempo changes are tackled by logarithmic re-sampling of the features' lag a xis. An evaluation of the proposed methods is conducted via genre classification using both artificial and real-world music signals. The evaluation results are presented and discussed accordingly.

Grundmann, R.

• Effect of the structural dynamics of the bocal on the sound spectrum of a bassoon
  T. Grothe, J. Baumgart and R. Grundmann [I]
[A]

The bocal of a bassoon is a thin curved metal tube to which a double-reed mouthpiece is attached. When the instrument is played, the reed acts as a pressure controlled valve and due to the pulsating flow the bocal starts to vibrate. The outer contour of the bocal has a major influence on the modal shapes of the structural modes and the relation between the eigenfrequencies. The matter of interest in this investigation is the interaction of the structural dynamics of the bocal and the reed. We assume that modifications of the bocals geometry and mass distribution will lead to measurable changes in the sound spectrum. Therefore the modes of a bocal are computed by FEM and the frequencies are checked experimentally with a LDV. In a playing experiment mass was added locally on the bocal. Sound pressure measurements were performed, while a professional bassoonist played sustained notes on a modifiable bocal. Very small changes, less than 4 dB SPL, occur in the overtones above 1 kHz and can be shown using principal components analysis of the sound spectra.

• Resonance frequencies and sound radiation of musical woodwind instruments
  A. Richter and R. Grundmann [I]
[A]

Decoupling the acoustic resonator from the excitation mechanism allows to analyse the influence of the resonators geometry on the acoustics of the whole woodwind instrument. Such investigations are done typically in the frequency domain. Numerical investigations in the time domain offer some advantages. They allow to track single acoustic waves which gives an insight in the influence of the resonators geometry on the oscillation behaviour. This allows to study how connections, tone holes and bore perturbations may influence the wave propagation and act as sound sources. In comparsion to methods which are formulated in the frequency domain also transient effects can be studied.

We present results from numerical investigations of the acoustic resonator of the bassoon as one example for woodwind instruments. The impulse reflectometry commonly used in experiments is modelled numerically. The results allow to analyse the resonance frequencies and also the acoustic radiation patterns.

Investigations in the time domain need to track acoustic waves over a long time period. We use a high-order discontinuous Galerkin formulation which offers the high accuracy required for this task.

Guastavino, C.

• Qualitative evaluation of Wave Field Synthesis with expert listeners
  A. Pras, E. Corteel and C. Guastavino [I]
[A]

A qualitative evaluation of Wave Field Synthesis sound quality was conducted with a panel of 5 expert listeners to investigate the perceptual effect of partial decorrelation at high frequencies and of reducing the number of active loudspeakers. The experiment was carried out on a 24 loudspeaker array. In each trial, participants were presented with different versions of the same auditory scene (using 24 or only 8 speakers, with and without partial decorrelation in high frequencies) and asked to freely describe the perceived differences. The verbal data was analyzed to derive semantic scale for sound quality evaluation.

Guerin, S.

• Extended Multi-Plane Pressure Mode Matching for CFD/CAA Coupling
  C. Weckmueller, A. Fritzsch and S. Guerin [C]
[A]

The objective is to extend the pressure mode matching method proposed by Ovenden and Rienstra in order to improve the aerodynamic/acoustic splitting at the interface between CFD and CAA computational domains. The numerical calculation and propagation of BPF tones in bypass ducts is the aimed application. CFD/CAA hybrid approaches have been developed for reducing the costs of aeroacoustic computations. Within the sound source region (rotor/stator stage), the unsteady Reynolds Averaged Navier-Stokes equations are solved, whereas in the rest of the computational domain, the sound propagation is computed with CAA methods solving the Euler equations. At the CFD/CAA interface, only the acoustic fluctuations should be transfered into the CAA domain to avoid stability issues in the CAA simulation. Ovenden and Rienstra proposed a so-called multiplane pressure mode matching method to remove the vortical part of the fluctuations. As suggested by De Roeck, this method can be extended with an aerodynamic model to improve the splitting. Here we have studied an aerodynamic model consisting of the same radial eigenfunctions as the ones taken for the acoustic modal basis, while the axial phase-velocity is the mean flow convection velocity. This is analytically investigated and applied to some numerical data.

• A method based on the ray structure of acoustic modes for predicting the liner performance in annular ducts with flow
  A. Moreau, S. Guerin and S. Busse [C]
[A]

The objective is to develop a fast method for predicting the sound attenuation by acoustic liners mounted in the inlet and bypass duct of turbofan engines. The method will be integrated into a prediction tool dedicated to design-to-noise. It works as follows: 1) The sound field generated by the fan stage is decomposed into acoustic modes (m,n) for a hard-walled annular duct with flow. 2) The ray structure of each acoustic mode is determined using an extension of an existing theory. 3) Finally the sound attenuation is calculated for each mode in function of the polar and azimuthal angles of propagation (angles at which the rays impinge on the wall) and the number of bounces; it is assumed that liners do not significantly modify the propagation angles and generate no mode scattering. The wall impedance is predicted using a model based on the geometrical parameters of practical liners. The results presented include the impact of acoustic liners on the modal distribution, the sound spectra, and the overall sound power. Both broadband and tonal components of fan noise are considered for a realistic configuration.

Guidati, S.

• Psychoacoustic Evaluation of Traffic Noise
  A. Fiebig, S. Guidati and A. Goehrke [I]
[A]

In the European research project ''Quiet City Transport'' the quantitative characterization of annoyance caused by traffic noise using psychoacoustic descriptors for a perception-related environmental noise description was a major objective. In the first project stage single pass-by noises have been evaluated by subjects. The aim was to discover components and patterns, which are responsible for specific annoyance effects. The results showed that certain noise properties such as diesel knocking are particularly annoying regardless of their absolute sound pressure levels. In the second part of the project complete traffic noise scenarios were evaluated to detect indicators which allow for an adequate assessment of environmental noise quality. Using a traffic noise synthesizer as well as artificial head recordings a wide variety of stimuli were created and evaluated by subjects. By means of a principal component analysis links between specific noise properties and subjective evaluation patterns were found and provided the basis for the development of a metric. The results of the study dealing with the evaluation of traffic noise will be presented and discussed under the perspective of their applicability in environmental noise policy and for soundscape studies.

• Source Localisation with Beamforming in Vehicle Acoustics
  S. Guidati [I]
[A]

Beamforming using microphone arrays is a common used technique for the localization of sound sources. The first applications were mainly aeroacoustic measurements in wind tunnels. Here the measurement object can not be measured in the very near field with e.g. a single microphone without producing excessive noise, whereas the array can be placed outside the airflow. A problem for beamforming is a reverberant environment. The algorithm is based on the evaluation of phase relations between the array microphones. In a highly reverberant environment e.g. a vehicle cabin, the phase relations are distorted by the reflections. With the use of measured transfer functions between the array microphones and the points of interest in the cabin the accuracy of the results can be enhanced. Recent advantages in computer technologies allow for a real time processing of the microphone array data resulting in an online visualization of the sound sources. Computer vision technologies based on multiple video cameras detect the three dimensional distribution of the potential sound sources increasing again the accuracy of the evaluation. The paper describes these techniques in detail and shows applications in vehicle acoustics.

Guilloud, G.

• Towards Industry-Standard Flow-Induced Noise Prediction
  G. Guilloud, P. Martinez, J. Golliard and C. Schram [I]
[A]

Engineers are nowadays predicting flow-induced sound in order to obtain quantitative assessment of complex devices in industries that now encompass air and ground transportation, energy sectors or domestic appliances. However, a fair amount of know-how still has to be captured to reach the decibel-accuracy sought by industrials for realistic configurations. A particular concern is the assessment of the numerical requirements of both the flow and acoustic modeling, in relation with the desired prediction accuracy over a pre-defined frequency range.

The present work is focused on the hybrid approach inspired from the acoustic analogy. A crucial issue regards the reliability of the flow data. In that respect, designing a mesh and selecting a numerical scheme for a given targeted noise frequency range can be complicated. A related diffculty lies in the data transfer that takes place from the CFD code to the acoustic tool. The problem consists in transferring data from a fine mesh towards a considerably coarser mesh without significant loss of information. Hereto, conservative interpolation schemes are applied to optimize the effciency of the process. The optimized procedure is applied to the case of a rectangular diaphragm fitted in a square duct, for realistic Reynolds and Mach numbers.

Guski, R.

• Critique of Pure Psychoacoustics
  J. Blauert and R. Guski [I]
[A]

Since product-sound quality has become a major engineering issue, psychoacoustics is widely applied in industrial R&D processes. It can. however, be observed that engineers which apply psychoacoustics sometimes have a rather naive understanding of the foundations and, thus, the potency of psychoacoustic methods. A common belief is, e.g., that psychoacoustic quantities can reliably be assessed with instrumental methods. The notion behind this belief is that, in laboratory experiments, ''unbiased'' versions of psychoacoustic quantities can be identified and assessed. It is the task of cognitive psychology to point to the risks that such an over-simplifying notion of psychoacoustics implies, and to provide knowledge and means to accordingly select and refine the psychoacoustic methods applied in engineering. It will be discussed that psychometric methods which are suitable for obtaining meaningful and valid information on product-sound quality vary substantially with the degree of abstraction of the references.

• Psychoacoustics Without Psychology?
  R. Guski and J. Blauert [C]
[A]

Guyader, J.-L.

• Energy resolution of complex vibro-acoustic problems
  N. Totaro and J.-L. Guyader [I]
[A]

The present article deals with an energy resolution of complex vibro-acoustic problems. This method is based on SmEdA approach previously published for extending Statistical Energy Analysis (SEA) to systems where modal equipartition of energy is not achieved. Modal energies are calculated from modal injected powers and modal coupling loss factors whatever the excitation even for systems with modal behaviour. From the computing time point of view two basic advantages appears, the calculation is made directly for frequency averaged energies and modal coupling loss factors are simply extracted from modal characteristics of the uncoupled systems obtained with standard Finite Element method. The approach provides also an analysis tools to detect the modal transmission path, very useful in early design stage. It is possible to easily identify couple of structural and acoustical modes most likely to transmit energy and estimate contributions of each structural mode to energy into the acoustical domain One interesting point is the possibility of calculating energy maps into systems with very small increase of computation, additional assumption is necessary that will be discussed. An application on a test case will be presented and compared with standard Finite Element method.

Haderlein, T.

• Effects of Vocal Aging on Fundamental Frequency and Formants
  S. Mwangi, W. Spiegl, F. Hönig, T. Haderlein, A. Maier and E. Nöth [C]
[A]

In this paper we study changes of the articulatory organs from young adulthood to old age. In particular, we focus on how the larynx and vocal folds age along with the rest of the body. The objective of the paper is to automatically compute the effects of vocal aging on fundamental frequency and the formants.

Based on the work of Harrington et al. we compare trends of the fundamental, 1st and 2nd formant frequencies as the age progresses. We also investigate how different phonemes, vowels in particular, are affected by vocal aging. The data consists of Queen Elizabeth's annual Christmas speech since the age of 26 to 76. The idea is to have the same speaker over a long period of time.

The results indicate that vocal aging causes the decline of the fundamental and the 1st formant frequencies while the 2nd and 3rd formants are not closely correlated to vocal aging. In general, the correlation between the fundamental frequency and age is -0.9027, while the average correlation for the 1st formant and age is -0.8071. In particular, the correlation between the 1st formant frequencies of phoneme `o' and age is -0.8946 while for the 2nd formant is 0.3351.

Häb-Umbach, R.

• On the Estimation and Use of Feature Reliability Information for Noise Robust Speech Recognition
  V. Leutnant and R. Häb-Umbach [I]
[A]

Traditionally, front-end methods for noise-robust speech recognition aim at denoising the observed feature vectors to obtain a point estimate of the clean, uncorrupted feature vector sequence. These estimates are then ''plugged'' into the Bayesian decision rule as if they were perfect estimates. In an optimal decoding rule, however, the reliability of the clean feature estimate would have to be taken into account, a method termed ''Uncertainty Decoding'' in the literature. We have shown that the optimal uncertainty decoding rule involves an integration over the posterior density of the clean feature given the observed corrupted feature sequence. This can lead to significant improvements in error rate if perfect feature reliability information is available.

In this paper we show how the clean speech feature posterior can be estimated for noise robust speech recognition in the car. We employ a state space model of speech and noise and track both of them jointly. This model can be extended by allowing for multiple observations. In doing so, the Kalman filter serves as a sensor fusion algorithm which fuses clean speech or noise estimates obtained from different enhancement methods. Experiments conducted on the AURORA databases illustrate the potential of this approach.

Haider, A.

• A Boundary Element model to calculate HRTFs. Comparison between calculated and measured data
  W. Kreuzer, P. Majdak and A. Haider [C]
[A]

Head-related transfer functions (HRTFs) play an important role in spatial sound perception. With the boundary element method (BEM) it is possible to numerically calculate these functions. As a rule of thumb six BEM elements per wavelength should be used, thus it is necessary to have very fine meshes when calculating HRTFs at high frequencies. In that case, the solution of the linear system of equations generated by the BEM needs a lot of computing time and memory. Advanced algorithms like the fast multipole method (FMM) speed up calculations and reduce memory requirements which allows a computation of HRTFs even for frequencies as high as 20 kHz in feasible time.

In this work, we give a short overview of a BEM model coupled with the FMM. We present calculated results for several persons and compare the results with measured data of these persons.

This work was supported by the Austrian Science Fond (Pr. P18401-B15)

Haider, M.

• Noise Reducing Devices: an Austrian Experience with the new European Technical Specification for Measurements of in-Situ Sound Diffraction
  M. Conter and M. Haider [C]
[A]

Traffic noise is one of the major environmental concerns within the European Union. The most important solution to noise abatement along roads and railways is the construction of noise barriers. In order to measure the acoustic performance in real life conditions and the long-term performance of these devices, the working group CEN/ TC226/ WG6 has developed the so called Adrienne method. This measurement method was introduced in 2003 with the technical specifications CEN/TS 1793-5, dealing with in-situ sound absorption and sound insulation, and CEN/TS 1793-4, dealing with in-situ sound diffraction. Added devices for sound diffraction are becoming more relevant in the last years to increase the acoustic performance of barriers. This paper describes the first Austrian research dealing with measurements of in-situ sound diffraction.This study, carried out in 2005 by arsenal research, compares the acoustic behaviour of aluminium cassettes, alternately equipped with reference barriers, partially non flat products, added devices on the top of the barrier, transparent insertions and digital and digital prints on the surface. In detail, the authors have measured and analysed the in-situ acoustic properties sound absorbtion, sound insulation and sound diffraction of each variant. The research has been carried out for the company Forster Metallbau GmbH.

Hak, C.

• A Comparison of the Variable Acoustics of Two Foley Studios for Sound Effects Recording
  R. Wenmaekers and C. Hak [C]
[A]

The sound effects and voices that are heard in motion pictures are often replacements or additions of sounds recorded in a Foley Studio. In this type of studio, named after Jack Foley the pioneer in sound effects recording, many objects like wooden floors and doors are present to record sounds from. These recordings need to sound as they are heard in the real world or even as an exaggeration of that world. Therefore it should be possible to create a lot of different acoustical conditions in a Foley Studio. The variability of acoustical properties of two existing Foley Studios have been compared: one typical studio and one with a good reputation. The impulse responses of typical Foley recording setups have been measured and analyzed with DIRAC software in both studios. The results show that in order to achieve sufficient acoustical variability in a Foley Studio it is important to make the critical distance and individual reflections controllable. It should also be possible to make the room semi-anechoic.

• Measuring Sound Insulation under Extreme Conditions using Deconvolution Techniques
  N. van Hout, C. Hak, E. Slaat and E. Gerretsen [C]
[A]

To compare the sound insulation performance of a building element with the given requirements, measurements have to be done. Generally, a broadband noise source is used, according to international standards. This method does not always work in practice due to high sound insulation values or high background noise levels. From a practical point of view it is very inconvenient or even impossible to perform an accurate sound insulation measurement for all frequency bands. A solution to this problem can be found in deconvolution techniques using MLS or sweep signals. It is possible to increase the signal to noise ratio using these techniques by averaging measurements and spreading out the spectral sound energy in time. As a result an efficient use of available sound power is possible. In a laboratory it can be investigated how to use MLS or sweep as a source signal and deconvolution as a measurement technique to obtain the sound insulation under noisy conditions. This has been investigated under moderately reverberant conditions, presented at the conference 'Acoustics'08 Paris', while this paper describes the same investigation under more extreme conditions.

• Room Acoustic Scale Model Measurements Using a “Spark Train”
  C. Hak and K. Bijsterbosch [C]
[A]

The spark gap, used as a sound source in scale model investigations, has advantages and disadvantages. One of the advantages is its limited size, which is particularly important for investigations at small scales. Also, the spark gap is omnidirectional and can be constructed such that the shape has a negligible influence on the sound field to be measured, for any scale factor. A disadvantage of an excessive discharge is the appearance of a shockwave. This leads to non-linear effects that make it impossible to acquire reliable impulse responses at small distances from the spark gap. A second disadvantage is the spread in successive separate impulse levels as a result of the micro-climate near the spark gap. Experiments were performed with a low energy spark gap (maximum electrode separation of 8 mm). In these experiments the energy was spread over a longer time interval through the use of spark trains. This allows for high INR values and reliable measurements of sound pressure levels. The period stability (jitter) of the spark train and spread in strength of the individual discharges were studied.

Haller, J.

• Metrology of HITU fields
  K.-V. Jenderka, K. Beissner and J. Haller [C]
[A]

High Intensity Therapeutic Ultrasound (HITU) is a non-invasive technique for the treatment of various types of cancer, as well as non-malignant pathologies by inducing localized necrosis of the tissue. Even though HITU is already in clinical use, this technique is still an emerging therapeutic procedure, with a lack of knowledge on reliable and stable treatment planning and monitoring procedures. A crucial step for the prediction of the desired temperature rise in a HITU-treated region is the measurement of the ultrasound output power. The evaluation of tissue thermometry methods, using MR or ultrasound itself, is important for the treatment monitoring, as well as for choosing the appropriate ultrasound power for a desired local intensity and temperature rise respectively. While ultrasound fields are well-investigated for output powers smaller than 50 W, this paper describes first results of measurements in much stronger, highly focused fields ( 300 W ultrasound power at 1.5 MHz) using an absorbing target arranged close to the transducer. The measurement uncertainties are discussed.

Hanselka, H.

• LOEWE-Zentrum AdRIA – An important step towards the commercialization of adaptive systems
  J. Bös, T. Bein and H. Hanselka [C]
[A]

The LOEWE-Zentrum AdRIA (Adaptronik - Research, Innovation, Application) is a large interdisciplinary research project funded by the government of the German federal state Hessen with a duration of 3 years and a budget of 17.7 million euros for the time being. A second phase of three years with a tentative budget of 20 million euros is planned. It involves the Fraunhofer Institute for Structural Durability and System Reliability LBF, 22 institutes from 3 different departments of the Technische Universität Darmstadt as well as one department of the University of Applied Sciences, all located in Darmstadt, Germany. The goal of this project is the creation and sustainable implementation of an internationally leading adaptronics research institution in Darmstadt. Eventually, it will lead to the foundation of a new Fraunhofer Institute for Adaptronics and of four new university institutes. The project is organized in ten technology areas (e.g., materials, sensors and actuators, systems and control, embedded systems, simulation tools, reliability), which supply the necessary technologies for three application scenarios, i.e., adaptive car, quiet office, and adaptive tuned vibration absorber. An important step towards the commercialization of adaptive systems will be the close cooperation with companies and suppliers during the second phase of the project.

Hansen, M.

• Modeling the masking of tones by Schroeder-phase harmonic tone complexes
  E. Rasumow, M. Hansen and S.D. Ewert [C]
[A]

The amount of masking from harmonic tone complexes (HTC) with Schroeder phase on a sinusoidal signal varies up to over 30 dB when the phase curvature of the masker is altered. This masking differences can be explained by the dispersion on the basilar membrane (BM) which, in interaction with the phase of the HTC components, influences the temporal ''peakiness'' of the BM excitation. When modeling the phase curvature dependent amount of masking the use of a linear BM-filter with optimized phase-curvature and magnitude response is insufficient in terms of the resulting dynamic range of the simulated masking thresholds. One important factor might be the inclusion of peripheral compression into the model. However, current studies using the dual-resonance nonlinear filter (DRNL) [Meddis et al., J. Acous. Soc. Am, 109(6):2852-2861,2001] have shown no improvement of the simulations. This study deals with the refinement of the response characteristic of the DRNL filter. The effect of modifications of the DRNL filter was evaluated in the framework of the CASP-model [Jepsen et al., J. Acoust. Soc. Am., 124(1):422-438, 2008]. Furthermore, the dependence of the model-parameters concerning the dynamic range of the internal representation was investigated.

Hardtke, H.-J.

• Validation of a measurement based Area Contribution Analysis
  H. Jakob, A. Bocquillet, S. Marburg and H.-J. Hardtke [C]
[A]

The Automotive Industry is in need of effective NVH measurement methods to optimize the design of isolation and damping concerning weight and cost constrains. In order to gain better knowledge on the contribution of each panel from the interior it is imperative to find a new and simple method of analysis in the early prototype stages of a vehicle. Using the technology of the Microflown-element it became possible to grasp the particle velocity close to any surface based on measurements. This information and the quantification of the transfer path allow for a prediction on contribution from subsystems in the cabin. The method is based on a combination of the Transfer Path Analysis principle and the quantification of the Emission of the surface. A simple example below shows the validation of such measurements as well as the comparison between a direct measurement and a simulation. This example depicts the critical range of the mid frequencies (200-3500Hz). The precision and usability of these simulated results will be discussed moving forward based on this specifically designed setup.

Hartmann, F.

• Transmission of noise through a pipe as structure-borne sound and fluid sound
  S. Buss, M. Wildemann and F. Hartmann [C]
[A]

Harwardt, C.

• Modeling the speaker-specific F0 Changes caused by raised Vocal Effort
  C. Harwardt [C]
[A]

The F0 behavior of a speaker is an important feature for forensic as well as for automatic speaker recognition tasks. Especially in forensic speaker recognition the two examined speech samples often contain speech of different degree of vocal effort. In such cases it is not recommendable to use the F0 measurements for speaker recognition because the F0 behavior changes with raising vocal effort. Hence an adequate estimation technique, that predicts a speaker's F0 changes for a given vocal effort, can improve speaker recognition tasks. The degree of raising the F0 caused by raised vocal effort is different for each speaker. To predict this speaker-specific degree of raising F0 is the goal of this investigation. Therefore we describe in this paper present estimation techniques for assigning F0 changes and introduce an estimation technique to predict the F0 changes affected by a raise of vocal effort. This prediction method and the problems in estimating speaker-specific F0 changes caused by raised vocal effort are presented in detail. Then the estimation technique is evaluated by using the Pool 2010 corpus. After discussion of the evaluation results we give an outline of future research.

Hasenpflug, H.

• Model Scale Measurements of Surface Ship Radiated Flow Noise
  C. de Jong, J. Bosschers and H. Hasenpflug [C]
[A]

Advances in weapon and sensor capabilities are driving an increased interest in the control of underwater signatures of naval platforms. The control of machinery and propeller noise is well understood but there is a shortfall of knowledge of the mechanisms that govern noise due to the flow around the hull of a surface ship. The subject has been investigated by the collaborative US-NL project Mechanisms and Prediction of Surface Ship Radiated Flow Noise. The project aimed to determine and quantify the sources of flow noise generated by surface ships in terms of ship speed and hull shape. The focus was on the underwater noise generated by turbulence excited hull plating vibration and noise resulting from breaking bow waves. A combination of full scale, large scale, model scale tests and computational fluid dynamics was used. The paper will discuss the test set-up and results of the model scale experiments which were made in a (highly reverberant) towing tank. A special silent towing carriage was developed to reduce the background noise, while an acoustic antenna was used to measure the noise due to the breaking bow waves. The pressure fluctuations due to the turbulent boundary layer were measured at several hull locations.

Haubold, J.

• Speech intelligibility prediction for normal hearing and hearing impaired subjects – MCHI-S
  L. Goenner and J. Haubold [C]
[A]

Haulick, T.

• Automatic Evaluation of In-Car Communication Systems
  A. Wolf, T. Haulick and G. Schmidt [I]
[A]

Due to high background noise and sound absorbing materials the communication between front and rear passengers in a car is often difficult. The in-car communication (intercom) system distributes the seat dedicated microphone signals via the sound system in order to improve speech intelligibility. Due to interfering signals and the closed loop operation of the intercom system, various signal processing techniques are required to reduce echoes and noise and to prevent feedback instability of such a system.

In this contribution a methodical evaluation of an intercom system and its incorporated signal processing schemes is presented. It considers different requirements for both speaking and listening passengers. The evaluation is performed with three types of systems or setups. The first one is an ''ideal'' intercom system. Here, a simulated system without any noise or feedback problem is computed in real-time and presented to listeners and to the measurement equipment. This ''ideal'' system is used to obtain suitable parameter settings like the maximum desired gain for instance. Secondly a real intercom system is evaluated, in order to analyze the achieved speech intelligibility and system quality. Finally, all measurements are performed without any intercom system, to obtain a basis for comparison.

Haumer, A.

• Modelling Sound Level and Pressure Loss in HVAC Duct Networks using a Framework in Modelica/Dymola
  H. Kühnelt, A. Haumer, T. Bäuml, U. Reisenbichler, C. Reichl and G. Karlowatz [C]
[A]

In the early design phase of large HVAC duct networks the joint calculation of the sound level and air flow rate and pressure loss as well as the prediction of the noise level in the passenger compartment gain more and more in importance. A unified framework for the one-dimensional flow acoustic simulation using Dymola/Modelica is introduced. The acoustic part is described by plane wave two-port theory, the fluid mechanic part by pressure loss models. Various approaches including analytic and (semi)empirical models based on experimental and simulation data can be implemented. Each component either represents the acoustic or the flow part of a duct element or can, itself, be composed of both types allowing the modelling of non-standard components. The object oriented approach provided by Modelica assures the extensibility of the models e.g. by transport and exchange of heat or the (de)humidification of the air. It also allows the integration into a complete system simulation of the HVAC network including the cooling circuit and the electronic control loop. The simulation is easily set up using a GUI; an external parameterization ensures persistent data management and a smooth workflow. The resulting system of equations is automatically pre-processed and solved by Dymola.

Haun, M.

• Measurements of Sound Localization Performance and Speech Quality in the Context of 3D Audio Conference Calls
  M. Hyder, M. Haun and C. Hoene [I]
[A]

The simulation of virtual acoustic environments helps to increase the quality of teleconferencing. Teleconferencing solutions supporting 3D audio have two advantages. First, the talkers can be identified easily by locating their position in the virtual acoustic environment. Second, if multiple persons talk at the same time, humans can focus on talkers of their choice by taking advantage of the cocktail party effect. However, 3D audio decrease the speech quality by decreasing loudness and adding reverberation and echoes. In this publication we study the trade off between speech quality and the ability of focusing and locating the talkers. We extend our initial work presented in the ITU Workshop in Lannion, Sep. 2008, in which we study the 3D audio rendering engine by Raine Kajastila et al. (AES 122th Convention, May 2007). We conduct subjective and objective listening-only tests using different head related transfer functions (HRTF), different algorithmic complexities, different positions of the conference call participants, and different geometries of the virtual environment. We present the statistical significant test results for sound localization performance and the speech quality. This study helps us to design our teleconference solution that we call 3D Telephony.

Haverkamp, M.

• Application of Synesthetic Design as multi-sensory Approach on Sound Quality
  M. Haverkamp [C]
[A]

Heckmann, M.

• Combining Auditory Inspirations and Hierarchical Feature Extraction for Robust Speech Recognition
  M. Heckmann, X. Domont, F. Joublin and C. Goerick [C]
[A]

Heidemann, H.

• Comodulation masking release with time-varying spectra
  H. Heidemann, J.L. Verhey and B. Epp [C]
[A]

The detectability of a tone masked by noise bands can be enhanced if noise bands with common level fluctuations are used i.e comodulated bands. This effect is commonly referred to as ''Comodulation Masking Release''(CMR). Since comodulation is a property of many natural sounds like e.g. speech, it is generally assumed, that this sensitivity is relevant in the differentiation of natural sound sources. So far, CMR has been investigated using stationary spectra. However, speech and other natural sounds are also characterized by spectral and spectro-temporal modulations. In order to test the robustness of CMR in the presence of these properties a modified CMR paradigm is used where the stimulus spectrum changes over time. In the first experiment sweeps rather than tones are used for the signal and masker band center frequencies. In the second experiment simultaneous manipulation along the spectral and the temporal axes is used in order to test the sensitivity to spectro-temporal patterns. The results of the first experiment show that CMR is still observed if the CMR is calculated relative to a condition with the same spectral properties but uncorrelated bands. The results of the second experiment indicate that changing from temporal modulation to spectro-temporal modulations abolishes CMR.

Hein, A.

• Interactive Fitting-Wizard in a Home Environment
  H. Baumgartner, A. Schulz, A. Hein, T. Herzke and I. Holube [C]
[A]

The European Hearing at Home (HaH) project aims at supporting hearing-impaired persons in their home environment. TV headsets with linear volume control allow people with hearing deficiencies good speech perception even in bad acoustic conditions. This approach attacks sensorineural hearing losses which imply a changed loudness perception combined with reduced perceivable dynamic range.

To help sensorineural users adapting the dynamic of their head phones according to individual hearing loss, three different interactive fitting procedures - based on broadcast audio material - have been evolved to fit parameters of a three band dynamic compressor performed in the master hearing aid (MHA). The MHA is the real-time audio-signal-processing framework in the background.

The fitting schemes can be used without any expertise and are intuitive and easy to use. A full walkthrough of the fitting wizard takes about 10 to 15 minutes. First results of user studies with 11 subjects (54-73 years old, 5 women, 6 men, slight to moderate hearing loss) show effectiveness of the approaches with respect to the improvement of the speech intelligibility and acceptance of sound quality. Further user studies will be performed.

Heintze, F.

• Sound Radiation of Double Reed Woodwinds
  F.B. Konkel, A. Jakob, F. Heintze and M. Möser [C]
[A]

In the recent past the verification of radiating properties by woodwinds was measured by using microphones next to the woodwind holes. Hence, it was of particular importance to use suitable microphone positions between the woodwind holes and free space. In the present contribution the radiation properties of the woodwind holes and the bell of a bassoon are investigated by using the ''acoustic camera''. For the acoustic testing the instrument is observed by three directions. The examination of the results verify the known ''cut off'' frequency of woodwinds at medium frequencies. From the experimental investigation, two distinct new frequency ranges are found for the radiation properties of the bassoon. The first located range between the medium and the high frequencies of the bassoon modes shows that the central radiation is emitted by the bell, but the last open woodwind holes are also a part of the sound radiation. The second noticed range is located at high frequencies and it becomes clear that the sound radiation is only emitted by the last open woodwind holes. However, there is no sound radiation by the bell at these frequencies.

Heise, S.J.

• Modulation perception and threshold fine structure: Experimental evaluation of the underlying mechanisms
  S.J. Heise, M. Mauermann and J.L. Verhey [C]
[A]

Amplitude modulation detection at low levels is influenced by the fine structure of the threshold in quiet. Modulation detection thresholds are higher when a tonal carrier falls in a fine-structure minimum than when it falls in an adjacent maximum. The aim of this study is to shed some light on possible underlying mechanisms, such as (1) a change in effective modulation depth due to a spectral colouring of the stimulus by threshold fine structure, (2) a reduction of the effective modulation depth in fine-structure minima by cochlear resonances such as spontaneous otoacoustic emissions (SOAEs), (3) modulation masking caused by the beating of a SOAE with the stimulus carrier. Modulation detection thresholds of sinusoidally amplitude-modulated carriers were measured for various stimulus parameters: five carrier levels (7.5-37.5 dB SL), two carrier bandwidths (0 and 30 Hz) and two carrier and five modulation frequencies which were adapted to the individual fine structure. Fifteen subjects with different degrees of threshold fine structure participated. The experimental results favour hypothesis (2). However, the data for different modulation frequencies indicate that spectral cues could be involved in modulation detection as well (hypothesis (1)), even for low modulation frequencies. Hypothesis (3) is not supported by the results.

• Modulation perception and threshold fine structure: Simulations with a nonlinear cochlea model
  B. Epp, J.L. Verhey, S.J. Heise and M. Mauermann [C]
[A]

Many natural sounds show temporal level fluctuations, i.e. amplitude modulations. Thus, a detailed knowledge about mechanisms underlying modulation perception is essential for the understanding of auditory signal processing. Most psychoacoustical studies investigated amplitude modulation processing for carrier levels well above threshold in quiet. Recently, Heise et al. (presented at DAGA2008) showed that the quasi-periodic fine structure in the threshold in quiet commonly observed in normal-hearing listeners affects the modulation detection of sinusoidally modulated tones with low carrier levels. The influence of threshold fine structure on modulation detection thresholds is not predicted by current models of modulation perception. This is essentially due to the simplification of the peripheral stages in these models. In contrast to the previous studies, the present study uses a one-dimensional active and nonlinear model of the cochlea which simulates a fine structure in the threshold in quiet. The representation of amplitude modulation in regions of fine structure at the level of the basilar membrane is analyzed and compared to the performance of subjects in the modulation detection task. The simulations can further enhance our understanding of the relative contribution of temporal and spectral mechanisms underlying modulation perception.

Heisel, U.

• Analysis of the Flow Field of Rotating Milling Cutters in Idle Mode
  L. Groß and U. Heisel [C]
[A]

Rotating tools are generally assumed to be one of the major sources of noise in machines. Vortices shedding from the rotating tool may cause turbulent flow fields. This article concentrates on the flow field of milling cutters in idle mode. At first, measurements obtained by a scanning laser Doppler vibrometer are analyzed. The test object is a milling cutter for T-slots. The test reading shows turbulences caused by vortices only in the chip space of the milling cutter and the near field of the cog. Subsequently, the flow field of the same milling cutter is simulated by ANSYS CFX. The simulation confirms the test reading obtained by the scanning laser Doppler vibrometer. Additionally, a different type of milling cutter is tested in water. The turbulence caused by this milling cutter occurs again only in the chip space. Therefore different geometries are simulated to detect relevant parameters causing turbulences.

Helfer, M.

• Sound Source Localisation with Acoustic Mirrors
  M. Helfer [I]
[A]

Acoustic mirrors, also known as sound mirrors, has been used for sound source detection and localisation for many years even though they were not very common. Probably their first application took place on the military field as early warning systems for adversary aircraft. In recent years many of them were used for sound source determination in rail and road transport. Especially in aeroacoustic wind tunnels sound mirrors have become one of the preferred measurement techniques for the investigation of extrior sound radiation of vehicles and planes. One of their favourable properties is the simple setup which allows easy identification and estimation of the main sources without requiring extensive data processing. On the other hand a complete documentation of the radiated sound pattern may be quite time-consuming.

Hellmuth, O.

• Efficient Parametric Audio Coding for Interactive Rendering: The Upcoming ISO/MPEG Standard on Spatial Audio Object Coding (SAOC)
  L. Terentiev, C. Falch, O. Hellmuth and J. Herre [C]
[A]

Spatial Audio Object Coding (SAOC) is one of the recent standardization activities in the ISO/MPEG audio group. SAOC is a technique for bitrate efficient parametric coding and flexible, user-controllable rendering of multiple audio objects. An extremely high audio compression efficiency is reached due to transmission of only a backward compatible mono (or stereo) downmix signal and a small amount of additional side information describing the perceptual properties of the sound objects. At the SAOC decoder, all sound objects can be rendered interactively to different spatial reproduction set-ups of choice. The new technology extends the MPEG Surround standard in a natural way and inherits its advantages and spatial rendering capabilities. Both bitrate efficiency and rendering interactivity makes SAOC attractive for broad adoption in a variety of interactive applications, for example music 2.0 (personal remixing), next generation teleconferencing and broadcasting. The publication provides an overview of the SAOC reference model architecture and its basic underlying concepts. It describes modes of operation, shows typical application scenarios and reports the current status of the ISO/MPEG standardization process.

Hemmer, D.

• Methods for the identification of background noises and noise events
  D. Hemmer, C. Pörschmann and P. Korte [C]
[A]

Noise is one of the main reasons for an increasing number of public complaints. This is why sound immissions caused by aircraft are supervised by noise long-term measurements. Yet, in many cases one is confronted with the problem that at the place of immission several noise sources, like noise from aircrafts and noise from street traffic, are mixed. In this case the contribution of aircraft noise needs to be detected from the recorded noise immissions. Within this presentation, methods are presented, which make it possible to separate the singular noise components by digital audio signal processing. Therefore, spectral and temporal characteristics of the audio signal are analysed and assigned to miscellaneous noise categories. One approach for this is to build a feature vector from several parameters and to compare it to vectors of reference signals. Different noise categories can thus be separated from one another. In a next step the noise contribution is automatically assigned to several emission sources. However, certain boundary conditions need to be considered, in which noise from different noise sources occurs at the same time. Furthermore, a high identification rate needs to be assured.

Hemmert, W.

• Coding of speech into nerve-action potentials
  M. Holmberg, H. Wang, M. Isik and W. Hemmert [C]
[A]

One of the most critical processing steps during encoding of sound signals for neuronal processing is when the analog pressure wave is coded into discrete nerve-action potentials. As any information lost during this process is no longer available for neuronal processing, it is important to understand and quantitatively model the underlying principles. We have developed a detailed model of auditory processing, which codes sound signals into spike-trains of the auditory nerve. We have also developed Hodgkin-Huxley models of cochlear nucleus neurons, which are driven by auditory nerve spike-trains. We analyze the quality of coding with the framework of automatic speech recognition and the temporal information processing capabilities with the methods of information theory. Our latest improvements in speech coding by introducing the effect of offset-adaptation together with an improved matching of neuronal features to the speech recognizer using an artificial neuronal network has lead to significant improvements of recognition scores, now reaching the values of successful technical feature extraction methods. Offset adaptation is also required to drive onset neurons in the cochlear nucleus, which are able to code temporal information with sub-millisecond precision ( 0.02 ms). Our results provide quantitative insight into temporal processing strategies of neuronal processing and are highly relevant for cochlear implants.

Hengst, K.

• Methods to characterize the acoustic properties of periodic surfaces
  K. Hengst, H. Drotleff, R. Wack and M. Blau [C]
[A]

Sound striking a periodic surface is absorbed and scattered depending on the surface properties. In comparison to non-periodic surfaces, periodic surfaces cause a higher sound absorbtion, due to their nearfield: If the nearfield is large, the scattered soundfield will be small and the amount of the absorbed sound high. If the scattered sound field is large, the nearfield is becomming smaller and the periodic surface is less absorbing because more sound energy is reflected. The mechanisms of sound absorption and acoustic scattering by periodic surfaces consisting of bands of alternating impedance are demonstrated first by means of a numerical model. Subsequently, measuring methods are presented to determine the acoustic properties of periodic surfaces at normal, diffuse and directed sound incidence. These measuring methods are assessed with regard to accuracy, informational value and practicability and compared to the theoretically calculated values. It is shown that it is possible (with varying accuracy) to determine the structural absorption at normal, diffuse and directional incidence. The geometrical reflection can be separated from the scattered field at diffuse and directional sound incidence. None of the measuring methods presented allows the separate examination of the near and farfield. The measurements essentially confirm the calculated values.

Henning, B.

• Generation of short ultrasonic pulses using active damping
  J. Rautenberg, S. Olfert and B. Henning [C]
[A]

In Non Destructive Testing with ultrasound the emission of short pulses is essentially if the echo signal should be separated from the emitted one. In this contribution a new approach is shown how to generate a short ultrasonic pulse using a 3-1 composite transducer. Therefore the ceramic bars, which are embedded in a polymer matrix, will be regarded as a kind of engine for the entire structure. After the pulse excitation the same bars can also be used for active damping to increase the bandwidth. The Mason model serves as a basis to the description of the ultrasonic transducer. This model is extended to the composite structure. The excitation signal is adaptively modified by a suitable signal preprocessing. Therefore the changes of boundary conditions, like the ambient temperature or acoustic impedance of adjacent media are considered. The modeling of the composite transducer, the experimental setup and first results will be presented.

Hensel, J.

• Investigation on Auditory Cortex Activation by Low-Frequency Sound and Infrasound Using Magnetic Resonance Imaging: Stimulus Generation and Control, and Noise Assessment
  J. Hensel, E. Dommes and G. Scholz [C]
[A]

Following the demonstration of the impact of infrasound in the cochlea [Hensel et. al. 2007, Hear. Res. 233, 67-76], in the present study the stimulation of the auditory cortex down to 12 Hz was investigated by means of functional magnetic resonance imaging (MRI).

The neuroradiological evidence of cortical excitation will be reported elsewhere. In this report, details of stimulus application and noise assessment are described.

In order to exclude the other sensory paths, and to avoid the difficulties of free-field stimulation, the stimuli were delivered directly into the ear canal.

This would ideally have required a metal-free earphone, which was not available. 12 meters of silicon tube and an ear-plug were used instead. To control the influence of standing waves, the sound pressure at the ear was measured in-situ by a metal-free optical microphone, embedded at the interface of tube and ear-plug. At very low frequencies, where high sound pressures were needed even near the hearing threshold, it turned out to be a demanding task to keep the harmonics small.

The microphone position near the ear canal entrance allowed to measure the MRI noise entering the system up to that point. Spectral analysis showed that this noise already contained some infrasound.

Hensel, K.

• Noniterative second harmonic ultrasound filed simulations: an axisymmetric approach
  N. Testoni, K. Hensel, M. Siepmann, N. Speciale and G. Schmitz [C]
[A]

Nonlinear numerical modeling of finite amplitude acoustic beams has a key role in the design of state-of-art ultrasonic systems. The Khokhlov-Zabolotskaya-Kuznetsov (KZK) equation is one of the most accurate model describing diffraction, absorbtion and nonlinearity combined effects in ultrasound wave propagation.

KZK solutions usually follow two different approaches: the spectral and the time domain method; the former is well suited for periodic ultrasound excitation, while the latter is more efficient in propagation of short pulses. As finite difference methods and stepping techiques are usually employed in most realizations, prediction accuracy and computational burden of ultrasound field simulations are strongly related to the depth.

The work described in this paper points to develop a model for simulating second harmonic ultrasound fields by featuring equations that can be solved without the use of iterative techniques. Using a domain change and addressing the integration along the axial direction with a proper approximation, robust estimations of nonlinear sound fields at arbitrary depth are produced for both focused and unfocused axisymmetrical sources. Predicted profiles for nonlinear propagation in water from real transducers will be presented and compared both visually and numerically with measurements from water tank experiments, showing the good agreement and artifact rejection obtained.

Hepberger, A.

• Wave Based Prediction Technique for Sound Radiation Analysis
  J. Rejlek, T. Mocsai, P. Silar, A. Hepberger and H.-H. Priebsch [I]
[A]

Recently, the wave based prediction technique (WBT) has been developed as an alternative method for solving steady-state acoustic problems. The WBT has proven to be a robust tool for interior acoustics. However, a considerable class of real-life acoustic applications involves the analysis of problems in unbounded spatial domains, such as sound scattering and sound radiation problems. Besides the methods based on boundary integral formulation of the governing differential equation various strategies employing the standard finite element schemes were developed in order to tackle unbounded problems. Although based on different approaches, all these concepts have the same basic idea in common, namely introducing an artificial truncation boundary that divides the infinite domain into bounded and unbounded region. This paper outlines the recent developments of the novel wave based approach and its application for three-dimensional sound radiation analysis by adopting similar strategy of dividing the model into two parts, as described above. The conventional interior formulation is used in the bounded part, while some new additional functions, which inherently satisfy the Sommerfeld radiation condition are adopted in the unbounded part of the wave model. Application to industry-sized problem demonstrates the enhanced computational efficiency, which allows the practical computational limitation to be shifted towards higher frequencies.

Herbig, T.

• Joint Speaker Identification and Speech Recognition for Speech Controlled Applications in an Automotive Environment
  T. Herbig and F. Gerl [C]
[A]

• Optimal projections between Gaussian Mixture Feature Spaces for Multilingual Speech Recognition
  M. Raab, O. Schreiner, T. Herbig, R. Gruhn and E. Nöth [C]
[A]

Multilingual speech recognition is increasingly gaining attention for in-car speech controlled applications. An example is a media player that allows selection of music by voice command, requiring speech recognition for multiple languages in order to cover the languages of artist names and music titles in a given music database.

There are two traditional approaches how a system can support the recognition of multiple languages. The first is to run a set of monolingual recognizers in parallel; the second one is to train an integrated multilingual recognizer dataset for the required set of languages.

The first approach has the disadvantage that there is no parameter sharing between the different recognizers, thus needing large amounts of processing power and memory. The second approach has the disadvantage that one multilingual recognizer has to be trained for every combination of languages.

In our paper we present a scheme to create a multilingual recognizer out of monolingual trained recognizers. A formula is given for an optimal projection of emission probabilities between Gaussians Mixture feature spaces. With this formula, we can project each HMM state of all languages to one set of Gaussians without retraining the system.

Hering, T.

• Vibration analysis by means of structure-borne sound intensity (SSI) simulations
  T. Hering and K. Wolf [C]
[A]

Measuring the airborne sound intensity (SI) is a well known and widely used method for the analysis of sound sources. By measuring sound intensities one can identify and locate sound sources and then compute the radiated acoustic power. In the case of real machine structures, the SI indicates the location of sound radiation but does not provide any information about the paths of the energy flow within the structure itself. Therefore the structure-borne sound intensity (SSI) is used to analyze energy flows and to evaluate the effects of structural modifications on the vibro-acoustic behavior.

Studies are carried out concerning the computation and measurement of the SSI to visualize the energy flow throughout test structures and real machine components that can be defined as plates and shells. The SSI is computed on FEM simulation data. Once the paths of energy flow are known it may be of interest to redirect energy flow into less critical areas instead of a mere global reduction of structure-borne sound. Based upon these SSI analyses, the effects of structural modifications, such as ribs and dampers, on the energy flow in plates or shells are demonstrated.

Heringa, P.

• Variability and Adaptability of the Acoustics in the new Conservatory of Amsterdam
  P. Heringa, D. Isbrücker and M. Valk [I]
[A]

Last summer the newly built conservatory for music of Amsterdam was delivered. The conservatory has 5 larger halls. The primary use of these halls is rehearsal and exams. They are also used for musical performances. The halls are quite different in order to learn the students to cope with different acoustics. They range from a hall suitable for rehearsal of an symphonic orchestra to a recital hall, a hall for jazz and pop and a hall for opera and musical. The acoustics of three halls can, to a certain extend, be adapted to the use of the halls. The acoustics aimed and achieved and the variation possibilities will be subject of the invited paper. Furthermore the conservatory has more then 100 education rooms and about 70 study rooms. The acoustics of these rooms can, by means of ''cushions'', be adapted to the lessons scheduled. The needed adaptability was predicted by means of experiments together with students and teachers in a mock-up of ''new rooms'' in the former conservatory building. They had to express their preference and give their evaluation of different subjective parameters. The adaptability and the evaluation of it will be shown.

Hernández, L.

• Reworking spoken dialogue systems with context awareness and information prioritisation to reduce driver workload
  J.L. Blanco, Á. Sigüenza, D. Díaz, M. Sendra and L. Hernández [C]
[A]

An entire host of new information, communication and entertainment applications and security systems are assaulting our cars. Interaction with onboard systems is moving beyond the classic switches, knobs, sticks, pedals and dashboard lights and sounds, to incorporate speech, touch-screens and a variety of other displays. While some of these systems are intended to help with the task of driving (e.g., navigation aids, lane departure and vehicle-in-blind-spot warnings), a constellation of displays, messages and requests threaten to overload the driver with information that is too rich and may distract him. Speech being eyes- and hands-free, it is increasingly being considered as a safe communication channel. However, interactive dialogue can increase the driver's mental workload, especially when robustness problems occur (e.g., recognition errors). Here we present our research results on dialogue as a secondary task, to gain understanding on how it may affect driving, on what dialogue strategies foster safer driving and on which applications, messages and interaction modalities should be prioritised over which others in each driving context. The proposed contribution is related to our current research activities within MARTA (Mobility for Advanced Transport Networks; www.cenitmarta.org), a Spanish, publicly-funded project. Experimental results from tests with a simulated driving platform are presented.

Herold, S.

• Active control of the sound transmission through a double-glazing window
  J. Bös, S. Herold, O. Heuss, M. Kauba and D. Mayer [C]
[A]

Within the framework of the European Integrated Project InMAR (Intelligent Materials for Active noise Reduction) a system for the active control of the sound transmission through a double-glazing window was developed. The window has a size of 1.2 m x 0.6 m and consists of two glass panes (6 mm and 4 mm thick) with a gas-filled gap of 20 mm in between. The vibrational behavior of the window panes was analyzed by means of finite element simulations and experimental modal analyses using a scanning laser vibrometer. The results of these analyses were used to determine suitable locations for the application of piezoelectric patch transducers. If these patches are used as sensors they can serve as error sensors or reference sensors for an active control system, if they are used as actuators they can suppress the vibrations of the window panes and, therefore, reduce the sound transmission through the window. Several transducer configurations and control strategies were tested and compared with each other. The sound radiation of a broad-band shaker excitation was reduced by up to 20 dB at various frequencies, and the sound transmission through the window was reduced by up to 12 dB at various frequencies.

Herre, J.

• Efficient Parametric Audio Coding for Interactive Rendering: The Upcoming ISO/MPEG Standard on Spatial Audio Object Coding (SAOC)
  L. Terentiev, C. Falch, O. Hellmuth and J. Herre [C]
[A]

Spatial Audio Object Coding (SAOC) is one of the recent standardization activities in the ISO/MPEG audio group. SAOC is a technique for bitrate efficient parametric coding and flexible, user-controllable rendering of multiple audio objects. An extremely high audio compression efficiency is reached due to transmission of only a backward compatible mono (or stereo) downmix signal and a small amount of additional side information describing the perceptual properties of the sound objects. At the SAOC decoder, all sound objects can be rendered interactively to different spatial reproduction set-ups of choice. The new technology extends the MPEG Surround standard in a natural way and inherits its advantages and spatial rendering capabilities. Both bitrate efficiency and rendering interactivity makes SAOC attractive for broad adoption in a variety of interactive applications, for example music 2.0 (personal remixing), next generation teleconferencing and broadcasting. The publication provides an overview of the SAOC reference model architecture and its basic underlying concepts. It describes modes of operation, shows typical application scenarios and reports the current status of the ISO/MPEG standardization process.

Herrenkind, M.

• About the Measurement of the Hands-Free Transmission Quality – An Experience Report
  H. Pastillé and M. Herrenkind [I]
[A]

The method of evaluation of the transmission quality by car-integrated hands-free telephones, presented on DAGA 2006, was tested in practice. First results from investigations will be shown. The performance of the perception is illustrated in a multidimensional judgement.
To produce a decision memo for the management and definition of the ''state-of-the-art'' in car-integrated phones, a pool of references and benchmarks are helpful. The reduction on a singular value will be discussed.

Herrmann, J.

• Fast FE-Analysis and Measurement of the Hydraulic Transfer Function of Pipes with Non-Uniform Cross Section
  J. Herrmann, M. Spitznagel and L. Gaul [C]
[A]

Automotive piping systems are often characterized by a non-uniform cross-section. Orifices in fuel charge assemblies are a typical example. The goal of this research is to quantify the influence of diameter changes on the hydraulic transfer function between the dynamic pressure at the beginning and at the end of the pipe. A finite element based substructuring and component mode synthesis technique is used to assemble the spatial piping system including the orifice. The adaptation of the Craig-Bampton method on fluid-structure coupled piping systems with an additional reduction of the remaining interface degrees of freedom leads to a considerable model order reduction. A local, frequency dependend fluid damping model is integrated in the finite element code and the result of the harmonic analysis is compared to measurements. A hydraulic test bench with a dynamic pressure source is used to estimate the hydraulic transfer function for different orifice diameters.

Herzel, H.

• Biomechanics and control of vocalisation in a non-songbird
  C. Elemans, R. Zaccarelli and H. Herzel [I]
[A]

The neuromuscular control of vocalisation in birds requires complicated and precisely coordinated motor control of the vocal organ (i.e. the syrinx), the respiratory system and upper vocal tract. The biomechanics of the syrinx is not well understood. Here, we aim to unravel the contribution of different control parameters in the coo of the ring dove (Streptopelia risoria) at the syrinx level. We designed and implemented a quantitative biomechanical syrinx model that is driven by physiological control parameters and includes a muscle model. Our simple nonlinear model reproduces the coo, including the inspiratory note, with remarkable accuracy and suggests that harmonic content of song can be controlled by the geometry and rest position of the syrinx. Furthermore, by systematically switching off control parameters, we demonstrate how they affect amplitude and frequency modulation and we generate new experimentally testable hypotheses. Our model suggests that independent control of amplitude and frequency seems not possible with the simple syringeal morphology of the ring dove. We speculate that songbirds evolved a syrinx design that uncouples the control of different sound parameters and allows for independent control. This evolutionary key innovation provides an additional explanation for the rapid diversification and speciation of the songbirds.

• Biomechanical modeling of chest-falsetto registers and their transition
  I. Tokuda, M. Zemke, M. Kob and H. Herzel [I]
[A]

The exact definition of registers in the human voice is still under debate. Especially the quantitative analysis of transitions between the registers has not been investigated in much detail yet. Excised larynx experiments show different kinds of voice instabilities that appear close to the transition from chest to falsetto register. Similar phenomena are also observed in vocalization of untrained singers. Towards comprehensive understanding of such register transitions, this paper carries out a biomechanical modeling of the human voice. Our model has a body-cover structure, which is in total composed of fourth masses. A smooth geometry is realized by introducing a polygon shape to the vocal fold model. Acoustical coupling with the vocal tract as well as the trachea is also considered. The simulation study shows that the model can reproduce many complex phenomena such as register jumps, hysteresis, subharmonics, and chaos, observed in excised larynx experiment as well as in vocalization of untrained singers. We study in detail the influence of the subglottal and supraglottal resonances on the transition point of the registers.

Herzke, T.

• Noise reduction for media streams
  T. Bisitz, T. Herzke, M. Zokoll, A.-M. Öster, S. Al Moubayed, B. Granström, E. Ormel, N. van Son and R. Tanke [C]
[A]

In the Hearing at Home project the central Home Information and Communication platform is developed to support hearing-impaired persons with easy to configure ``Supportive Audio Signal Processing'' and visual support on a TV screen in home environments. Amongst other, individually configurable algorithms, several algorithms can be used for noise reduction of the media streams without individual fitting to enhance the accessibility of the presented audio materials using the Master Hearing Aid as a real-time audio-signal-processing framework: (1) Single-channel noise reduction according to Ephraim-Malah together with a speech pause detector, (2) stereo noise reduction additionally removing the central signal in a control path, (3) frequency shaping noise reduction by emphasising a frequency range important for speech intelligibility, (4) time-varying attenuation and (5) noise reduction by selection of channels. It will be possible to classify the acoustic situation in the media stream automatically in real-time using six classes allowing to activate the appropriate noise reduction strategies. In user tests the performance of algorithm 1-3 is evaluated: Results of sound quality assessment, listening effort estimation and speech intelligibility measurements with hearing impaired subjects from groups with different types of hearing impairment will be presented.

• Interactive Fitting-Wizard in a Home Environment
  H. Baumgartner, A. Schulz, A. Hein, T. Herzke and I. Holube [C]
[A]

The European Hearing at Home (HaH) project aims at supporting hearing-impaired persons in their home environment. TV headsets with linear volume control allow people with hearing deficiencies good speech perception even in bad acoustic conditions. This approach attacks sensorineural hearing losses which imply a changed loudness perception combined with reduced perceivable dynamic range.

To help sensorineural users adapting the dynamic of their head phones according to individual hearing loss, three different interactive fitting procedures - based on broadcast audio material - have been evolved to fit parameters of a three band dynamic compressor performed in the master hearing aid (MHA). The MHA is the real-time audio-signal-processing framework in the background.

The fitting schemes can be used without any expertise and are intuitive and easy to use. A full walkthrough of the fitting wizard takes about 10 to 15 minutes. First results of user studies with 11 subjects (54-73 years old, 5 women, 6 men, slight to moderate hearing loss) show effectiveness of the approaches with respect to the improvement of the speech intelligibility and acceptance of sound quality. Further user studies will be performed.

Hessinger, J.

• Optimization of lightweight floors in the low frequency range with a FEM based prediction model
  A. Rabold, A. Düster, J. Hessinger and E. Rank [I]
[A]

The impact noise transmission at low frequencies is a well known problem of lightweight floors, which is treated in many publications. A satisfying solution, considering the different construction principles of lightweight floors, could not be found so far. To overcome this problem a FEM based prediction model for the optimization of the floor construction and the improvement of the impact sound insulation has been developed and applied in a current research project at the TU München. The overall approach of the prediction model consists of the three-dimensional modelling of the structure and the excitation source (standard tapping machine), the subsequent modal- and spectral analysis and the computation of radiated sound from the ceiling. The validation of the prediction model has been carried out by comparing the evaluated impact sound pressure levels with results from measurements on 25 different floor constructions. In the next step the prediction model was used for the improvement of established lightweight timber floors. Finally these constructions were tested in a laboratory test stand according to ISO 140-6 at ift Rosenheim centre for acoustics and the ibp Stuttgart. This contribution shows the results of the computations and the construction rules developed for optimized lightweight floors.

Heuss, O.

• Active control of the sound transmission through a double-glazing window
  J. Bös, S. Herold, O. Heuss, M. Kauba and D. Mayer [C]
[A]

Within the framework of the European Integrated Project InMAR (Intelligent Materials for Active noise Reduction) a system for the active control of the sound transmission through a double-glazing window was developed. The window has a size of 1.2 m x 0.6 m and consists of two glass panes (6 mm and 4 mm thick) with a gas-filled gap of 20 mm in between. The vibrational behavior of the window panes was analyzed by means of finite element simulations and experimental modal analyses using a scanning laser vibrometer. The results of these analyses were used to determine suitable locations for the application of piezoelectric patch transducers. If these patches are used as sensors they can serve as error sensors or reference sensors for an active control system, if they are used as actuators they can suppress the vibrations of the window panes and, therefore, reduce the sound transmission through the window. Several transducer configurations and control strategies were tested and compared with each other. The sound radiation of a broad-band shaker excitation was reduced by up to 20 dB at various frequencies, and the sound transmission through the window was reduced by up to 12 dB at various frequencies.

Heute, U.

• Attribute-based Instrumental Assessment for Speech- Transmission Quality
  L. Huo and U. Heute [C]
[A]

• Is the Fujisaki model a suitable (prosodic) model for the voice-conversion task?
  J. Schwarz, M. Tran and U. Heute [C]
[A]

Heutschi, K.

• Footprint of a vehicle - the Swiss Contribution for noise of heavy road traffic
  I. Schlachter, K. Heutschi and L.D. Poulikakos [C]
[A]

Footprint is a European collaborative research project within the pan-European EUREKA framework for development and exploitation of technologies crucial tho global competitiveness and better life quality. The aim of the project is to relate the environmental footprint of a heavy road or rail vehicle to the lifetime cost of maintaining the infrastructure and environment. As a project partner Switzerland installed in 2005 the first Footprint Measuring station in Europe for road traffic on the North-South-transit motorway A1. Objectives include developing techniques to measure and characterise impacts including noise. Methods of analysing the data have been established to develop strategies to minimise these impacts.

For noise, methods to determine the acoustical footprint of a vehicle passing have been developed in Switzerland by evaluation of the in situ measured maximum sound pressure level. The impact of various vehicle types and traffic flows can also be recorded. Successfully evaluated pass-by events reach 95% for dense traffic. The Greening Transport Package announced 8 July 2008 promotes measures to reduce the environmental impact of transport, including reallocating external costs from society to users. A further stage of Footprint is to monitor the effectiveness of noise reduction strategies and to cost an environmental index for different vehicle types.

Hillenbrand, J.

• Ferroelectret-film accelerometers with high sensitivities
  M. Kodejska, J. Hillenbrand and G.M. Sessler [C]
[A]

Ferroelectrets are piezoelectric films of charged cellular or porous polymeric electrets and have larger d33-coefficients than lead zirconate titanate (PZT). Compared to PZT and other piezoelectric materials, ferroelectrets are flexible, lightweight, and cheap. These properties allow the construction of light, flat, and sensitive accelerometers of various shapes. Such accelerometers were built with polypropylene (PP) ferroelectrets in single- and in multi-layer arrangements and seismic masses from 1 to 20 g. Film area, number of ferroelectret films, and seismic mass determine sensitivity and resonance frequency of the accelerometer and these parameters were varied for the production of different accelerometers. The frequency responses of the sensitivity of the accelerometers were measured using an electrodynamic vibration exciter. A typical sensitivity of about 3 pC s2/m was obtained for an accelerometer consisting of a single ferroelectret film with d33 500 pC/N in the audio frequency range, a seismic mass of 9 g, and an area of 3 cm2. A theoretical description of the various accelerometers has been achieved using simple models. Experimental and theoretical data were compared and showed good agreement. The results demonstrate that ferroelectret-film accelerometers have advantageous properties, in particular when adapted to specific applications.

Hils, T.

• About Speech Intelligibility depending on different Sound Insulations (part 1)
  T. Hils and H. Alphei [I]
[A]

In the course of the development of the DAGA recommendation ''Schallschutz im Wohungsbau - Schallschutzausweis'' (sound protection in residential construction - certificate of noise protection), in a former study theoretical considerations were made on the basis of the calculated loudness of transferring noises that served, among other things, as a basis for the predefinition of the degree levels in the different classes of noise protection.

The recommendation also includes a list of verbal descriptions of the subjective perception of (typical) sounds in housing space and speech from neighbouring living areas. In another list the perception descriptions are linked with the levels of noise protection for giving an orienting evaluation of sound transfer in the seperate classes of noise protection. However, the application for the planning in practice is usually difficult, as these experimental data are only correct under special conditions, which have not been sufficiently investigated so far.

The research in hand starts here. The perception of different situations of sound insulation shall be experimentally investigated under controlled conditions. Thereby, the speech intelligibility is first investigated with sample tests for different sound insulations.

• About Speech Intelligibility depending on different Sound Insulations
  H. Alphei and T. Hils [I]
[A]

In the first part (see previous abstract Hils/Alphei, 16:00), the performed hearing tests were described, and first results were shown. The aim is the verification of the expressions of the list based on the measured intelligibility for as real as possible situations.

In this second part it is tried to create a correlation between intelligibility and sound insulation at a specific initial level in the emitting room.

It is looked at the loudness more closely that, on the other hand, verifies the relation beween loudness and the levels of sound protection.

Hirsch, K.-W.

• Noise Management at Civil Shooting Ranges
  D. Knauss and K.-W. Hirsch [C]
[A]

DIN ISO 17201-2 proposes a scheme to set up a daily noise management on civil shooting ranges. The scheme relies on rating levels based either on calculations or measurements of the LAFmax at relevant receptor points. In order to employ such a management system, detailed information have to be known. For civil shooting ranges however, it is sometimes difficult to gather all the details in advance and to put everything together in an appropriate computer based management system.

This paper reports on a system that follows the basic ideas of the DIN ISO 17201-5 but relies on the measurement of each single shot at given positions at close distance during normal operation. Using these single shot levels and the met. data, the system applies a transfer matrix to predict the contribution of each shot to the rating level at each relevant receiver point. The coefficients of the matrix are determined during a preceding measurement campaign and accounts for the wind speed and direction.

• General classification scheme for outdoor sound propagation situations
  K.-W. Hirsch and B. Vogelsang [C]
[A]

Noise assessment normally considers a long term energy equivalent sound pressure level and a maximum level to indicate the noise for correlation with noise impairment, e. g. annoyance. This concept is found throughout many national and local regulations. The information on noise situations from these two indicators is not sufficient to make reliable assessments and, in particular, to make progress in the field of environmental noise control. There are several reasons to go beyond. For instance, in complex noise situation the perception of noise depends on audibility and background noise. And modern noise management wants to take advantage of weather conditions on a daily base. Progress is possible because the knowledge on sound propagation outdoors has strongly improved. Also the increasing computer power and the facilities of modern measurement systems add new options to overcome the constraints of a two-level-indicator concept.

This paper suggests the prediction of a level distribution and its indicators to broaden the information on noise exposure. It recommends a general scheme of classification of all decisive input parameters of the source, of the sound propagation situation and of the receiver site in order to obtain this distribution through a variation of these parameters within their classes.

Höfker, G.

• In-situ measurements of road barriers made of natural stones
  A. Isele, G. Höfker and C. Nocke [C]
[A]

With an increasing disturbance of traffic noise the length of barriers along highways and railways grows year by year. Besides the well-known materials like concrete, metal, glass or synthetic stones also natural stones are more and more in use because of its appearence.

A minimum value for sound absorption coefficient and transmission loss is necessary for the acoustic efficiency of the barrier. Measurements of the acoustic properties of gabions in the reverberation chamber and laboratory are often impossible because of the construction or the weight respectively. To avoid these problems the reflection index and the sound insulation index were determined by in-situ measurements according to DIN EN ISO 1793, part 5. Results of the reflection index for several materials, grain sizes and layer thicknesses of the charge of the wire mesh are presented.

Earlier research have shown that a minimum value of the transmission loss could be reached with additional elements inside the gabion only. Different constructions of concrete with and without slits between the elements have been tested. A mathematical description of the problem is given, which shows the essential parameters of sound insulation and allows approaches to optimisation.

Höldrich, R.

• Recruiting and Evaluation Process of an Expert Listening Panel
  A. Sontacchi, H. Pomberger and R. Höldrich [C]
[A]

During the last few years an increased desire to assess and quantify acoustical properties of technical products can be recognized. Utilizing objective evaluation methods, which are commonly based on modelling the human sound perception and underlying auditory hearing process, can deliver qualitative and quantitative acoustical description of the investigated product. However, only a limited number of application areas with specific restricted requirements can be treated reliable within tolerable assessment results. The predominant crucial aspect is based on the complex interaction of acoustical stimulus and auditory perception. In general, overlapping and competitive acoustical stimuli can be instructively assessed only by a special group of listeners. To derivate the obtained answers to the majority of potential people how are involved with the regarded device under test specific demands result to select this special group of listeners. Beside the listeners selection process and trainings phases the design and accomplishment of the listening test has to guarantee that the subjective results are within defined limits and can be reproduced anytime. Within this work the constitution and rehearsal of an expert listening panel will be described and discussed on the basis of already established and published selection procedures and new presented methods.

Höller, C.

• Prediction of the sound radiation from a plate excited by a structure-borne sound source
  M. Lievens, P. Dietrich and C. Höller [I]
[A]

The radiation of structure-borne sound into a medium is generally predicted by linear system theory. In characterising a structure-borne sound source and its radiation, the goal is to predict the sound pressure at an arbitrary point in the medium using only mobilities and frequency responses of the individual system components. This seems straight forward but in practice the prediction often fails due to one or more of three causes: an incompletely measured coupling situation, a high measurement uncertainty of the mobilities, or the non-linear source behaviour.

To investigate the validity of the multipole theory a small one-dimensional measurement setup was built. It consists of a small electrical motor connected to a plate that radiates into a receiving room of one cubic metre. In this way many positions, plates and motor excitations can be analysed with limited effort.

The analyses will compare power measured in situ to power predicted using in situ measured mobilities and source excitations. The sound pressure in the receiving room is used to verify the assumption of a one dimensional excitation process and to monitor the error in the prediction.

Hoene, C.

• Measurements of Sound Localization Performance and Speech Quality in the Context of 3D Audio Conference Calls
  M. Hyder, M. Haun and C. Hoene [I]
[A]

The simulation of virtual acoustic environments helps to increase the quality of teleconferencing. Teleconferencing solutions supporting 3D audio have two advantages. First, the talkers can be identified easily by locating their position in the virtual acoustic environment. Second, if multiple persons talk at the same time, humans can focus on talkers of their choice by taking advantage of the cocktail party effect. However, 3D audio decrease the speech quality by decreasing loudness and adding reverberation and echoes. In this publication we study the trade off between speech quality and the ability of focusing and locating the talkers. We extend our initial work presented in the ITU Workshop in Lannion, Sep. 2008, in which we study the 3D audio rendering engine by Raine Kajastila et al. (AES 122th Convention, May 2007). We conduct subjective and objective listening-only tests using different head related transfer functions (HRTF), different algorithmic complexities, different positions of the conference call participants, and different geometries of the virtual environment. We present the statistical significant test results for sound localization performance and the speech quality. This study helps us to design our teleconference solution that we call 3D Telephony.

Hönig, F.

• Effects of Vocal Aging on Fundamental Frequency and Formants
  S. Mwangi, W. Spiegl, F. Hönig, T. Haderlein, A. Maier and E. Nöth [C]
[A]

In this paper we study changes of the articulatory organs from young adulthood to old age. In particular, we focus on how the larynx and vocal folds age along with the rest of the body. The objective of the paper is to automatically compute the effects of vocal aging on fundamental frequency and the formants.

Based on the work of Harrington et al. we compare trends of the fundamental, 1st and 2nd formant frequencies as the age progresses. We also investigate how different phonemes, vowels in particular, are affected by vocal aging. The data consists of Queen Elizabeth's annual Christmas speech since the age of 26 to 76. The idea is to have the same speaker over a long period of time.

The results indicate that vocal aging causes the decline of the fundamental and the 1st formant frequencies while the 2nd and 3rd formants are not closely correlated to vocal aging. In general, the correlation between the fundamental frequency and age is -0.9027, while the average correlation for the 1st formant and age is -0.8071. In particular, the correlation between the 1st formant frequencies of phoneme `o' and age is -0.8946 while for the 2nd formant is 0.3351.

Hörchens, L.

• Plane wave decomposition: an inverse problem in array processing
  L. Hörchens and D. de Vries [I]
[A]

The decomposition of wave fields into plane waves is a standard tool for the analysis of impulse responses measured with an array of microphones. If the geometry is appropriately chosen, the decomposition can be obtained by spatio-temporal deconvolution. Several measures can be taken to tackle the ill-posedness of the associated mathematical problem. Typical examples are given by beamforming techniques or plane wave decomposition using circular or spherical harmonics. These approaches are compared based on the presented framework. The stability of the procedure for plane wave decomposition is strongly influenced by the choice of the array setup, which is extensively used for current approaches. On the processing side, regularisation techniques and prior knowledge can be employed to overcome instabilities and improve the achievable resolution. Exemplary results based on this approach are shown, pointing out the potential of further research in this field.

Hoevelmann, N.

• ANC Concept study using a 1-D-Model
  N. Hoevelmann, D. Sachau and K. Kochan [C]
[A]

In an advanced - especially propeller driven - aircraft the sound pressure level may reach up to high values caused by propeller blades producing disturbances when passing the fuselage. This excitation induces high tonal noise levels in the low frequency range in the aircraft interior. Due to mass reduction as a major aspect in aircraft design research has been focused on active noise control (ANC). In order to test the performance of ANC-systems test facilities are needed for different technology readiness levels. To analyze the expected acoustic behaviour of future aircrafts under realistic conditions a test rig for acoustic ground test is needed. The present paper will outline the design process of a mechatronic mock-up including both airborne and structure borne noise sources as well as the associated transfer paths. It is planned to use data taken from numerical simulations to adjust a specific noise field inside the mock-up by a proper control scheme. To simulate relevant transfer paths it is planned to integrate modern light weight structures - such as CFK - into the test rig.

Hoever, C.

• Active Structural Control of Impacts with a View Towards Noise Control
  C. Hoever, W. Kropp and B.A.T. Petersson [C]
[A]

Impacts between bodies of any sort often constitute an important source of noise and vibration. These impacts can be characterised as very short contacts between bodies leading to a sudden release of energy as audible noise and to vibrations of the involved structures. Resulting sound pressure levels often pose a serious health risk and structural vibrations lead to further noise emission, material fatigue and breakdown.

Due to a variety of source mechanisms and multiple possible ways of transmission, propagation and radiation, classical methods of noise and vibration control are often difficult to implement.

In this study it is theoretically investigated how noise and vibration generated by a sphere-plate impact can be affected by application of an active force acting on the plate at the impact location.

Two different numerical models are derived as a basis for simulations. A parameter study is conducted to investigate the possibilities of active structural control of impacts. With the obtained data, an active control configuration is developed which is applicable to a wide variety of plate-sphere impact situations and which leads to promising noise and vibration reduction.

Hoffmann, R.

• Measuring Frequencies with Historic Resonators from SCHAEFER
  R. Hoffmann, D. Mehnert, R. Dietzel and G. Fuder [I]
[A]

Measuring frequencies of acoustic processes within the audible range required large technical expense in the early times of experimental phonetics and acoustics. Indirect methods like KUNDT's tube, the monochord, the interpretation of kymographic recordings, or the interference tube from QUINCKE allowed such kind of measurement, but had been too complicated for practical purposes. Therefore simpler measurement methods have been looked for in early times. E. g., it is possible to apply one-sided closed pipes of variable length as resonators. The acoustic perception of the experimentator was used as indicator.

This paper investigates this measuring method at first within the historic arrangement. Then, the transfer behaviour of single tube resonators from SCHAEFER will be measured in the free sound field.

These results as well as the ability of judgement of the listener are used to estimate the measuring uncertainty of such historic methods.

Hofmann, M.R.

• Comparison of Coding Techniques for Photoacoustic Coded Excitation
  M. Mienkina, A. Eder, C.-S. Friedrich, N. Gerhardt, M.R. Hofmann and G. Schmitz [C]
[A]

In photoacoustics (PA) ultrasound is generated by laser irradiation of tissue. Usually, PA signals are generated by Nd:YAG lasers but it would be attractive to use less expensive laser diodes instead. Although laser diodes exhibit low pulse energy, the pulse repetition frequency (PRF) of pulsed laser diodes is much higher than the PRF of Nd:YAG lasers thus averaging can be used to increase the SNR. The applied PRF for averaging is limited by the acoustical time-of-flight. To further increase the SNR we propose to use coded excitation for PA imaging. In optical-time-domain-reflectometry, which is similar to PA imaging, Simplex Codes (SC) are widely used for coding. In ultrasound imaging Golay Codes (GC) are successfully employed. Therefore, we compare the performance of SC and GC for PA coded excitation (PACE) by a simulation study. For both codes, the coding gain normalized to equivalent averaging procedures is computed as a function of code length and PRF of the laser diode. Assuming a time-of-flight of 60 s and e.g. a PRF of 250 kHz the coding gain of GC is 1.6 dB higher than the coding gain of SC. Consequently, GC is more suitable for PACE than SC.

Hohmann, V.

• Subband instantaneous-frequency analysis to determine masking with high temporal resolution for use in audio codecs
  N. Koppaetzky, S.D. Ewert, B. Kollmeier and V. Hohmann [C]
[A]

Algorithms for perceptual audio coding use psychoacoustical masking models to compress audio data with minimal impact on the perceived audio quality. Latest psychoacoustical research emphasizes the excellent spectro-temporal resolution of the auditory system. Most common audio codecs, however, apply classical spectral masking models which generally reach the required high spectral resolution at the expense of an insufficient temporal resolution. An application of a current psychoacoustical model which simulates the high spectro-temporal resolution of the auditory system, might offer an increase of the coding efficiency or of the perceptual coding quality, respectively. To analyze the general effect of an enhanced time resolution, a current nonlinear auditory filterbank model with instantaneaous-frequency conrol was tested in this constribution. The applicability of the model for the classification of the tonality of the signal and for forming an estimate of the spectro-temporal masking pattern was examined. The masking model as well as the resulting perceptual coding quality which can be achieved by a combination of the model with a simple perceptual coder are presented and discussed.

• Evalution of a Nonlinear Auditory Filterbank with Instantaneous Frequency Control
  R. Eilers, S.D. Ewert, J.L. Verhey and V. Hohmann [C]
[A]

Psychoacoustical studies suggest that the increase in suppression as a function of suppressor level is greater for a suppressor spectrally below than above the signal frequency. This pattern was investigated by measuring suppression for a tonal signal with two different psychoacoustical methods in the same subjects. The first method was the pulsation threshold technique that was used in a two-tone suppression paradigm with suppressor frequencies below and above the signal frequency. The second method was a forward-masked paradigm where the masker was a bandpass noise and the parameter the masker bandwidths. The noise was either centred at the signal frequency or was asymmetric with a larger portion of the spectrum on either the low or the high frequency side. The results of the two experiments were consistent with the hypothesis of different suppression growth functions for the different frequency regions. The measured data was modelled with a nonlinear filter approach. The model uses a modified dual-resonance filter by including a control of the nonlinear filter gain by the sub-band instantaneous frequency. It is investigated, if the model is able to quantitatively predict suppression and the effect of suppressor frequency.

• Classification of Reverberant Acoustic Situations
  J. Schröder, T. Rohdenburg, V. Hohmann and S.D. Ewert [C]
[A]

In daily communication, speech intelligibility depends on the acoustic surrounding or acoustic situation. Particularly for hearing impaired persons, speech understanding is often problematic if speech is distorted by (room) reverb, noise or competing talkers. Acoustic situations are characterized by different dominating types of distortion. Hearing aids might provide appropriate algorithms to enhance speech intelligibility in the different acoustic situations. A robust and fast automatic classification of the acoustic situation should therefore select the appropriate hearing aid algorithm without requiring an action of the hearing aid wearer. This study is concerned with the automatic estimation of the reverberation time (T60) in natural situations and with unknown excitation signal. Acoustic test situations were generated by convolving speech signals with artificial and real room impulse responses with T60-times ranging from 0.1 to 4 s. Three different features derived from the cepstral mean, the autocorrelation function and from the distribution of modulation energy were used to train and classify multiple groups of different reverb times with Gaussian mixture models (GMM). Additionally, information about the unknown excitation signal was gained by estimation of its (audio-) spectral content and used to improve the classification. The classifiers, features and results are presented and discussed.

• The Effect of Envelope Waveform on Lateralization
  M. Klein-Hennig, M. Dietz, S.D. Ewert and V. Hohmann [C]
[A]

Interaural timing disparities are important cues for lateralization in the human auditory system. The auditory system is sensitive to timing disparities in the fine-structure and the envelope of sounds. At high frequencies, only envelope disparities can be exploited due to the lack of phase-locking to the fine-structure. In previously published data [e.g. Bernstein and Trahiotis in ''Hearing - from basic research to applications'', Springer 2007] parameters like the modulation frequency, or the power of a modulating sinusoid have been altered in order to investigate the role of the envelope waveform on lateralization. However, these ''classical'' parameters simultaneously change several ''secondary'' envelope parameters like attack and decay times, as well as pause and hold durations, which might conceal the fundamental cues exploited by the auditory system. In this study, psychophysical measurements were conducted with customized envelope waveforms in order to investigate the effect of isolated secondary parameters on lateralization. For high-frequency tones centred at 4 kHz with systematic envelope modifications, the just noticeable differences of interaural envelope time differences were measured. The results indicate that attack times and pause durations prior to the attack are the most important envelope features. The results are compared to predictions by a binaural auditory model.

• Azimuthal Localization of Concurrent Speakers Employing Interaural Coherence
  M. Dietz, S.D. Ewert and V. Hohmann [I]
[A]

A computational model for the localization of concurrent speakers is presented that extends the binaural model presented by Dietz et al. (Fortschritte der Akustik - DAGA 2007, 107-108). The processing is motivated by mammalian physiology and human psychoacoustic data. A wide range of interaural parameters is exploited, namely the interaural level difference (ILD), the interaural phase difference (IPD) of the fine structure, the IPD of the fundamental modulation frequency and the interaural coherence (IC). The parameters are determined for each auditory filter which allows a simultaneous estimation of several sound directions. By focusing on instants with high interaural coherence the robustness of the estimation could be greatly improved (Faller and Merimaa, 2004, J. Acoust. Soc. Am., vol. 116, 3075-3089). For two well-separated speakers in quiet the estimation error is generally less or equal 10. In noisy environments a single speaker can be localized with an error of less than 20, even for a signal-to-noise ratio of 0 dB.

Holmberg, M.

• Coding of speech into nerve-action potentials
  M. Holmberg, H. Wang, M. Isik and W. Hemmert [C]
[A]

One of the most critical processing steps during encoding of sound signals for neuronal processing is when the analog pressure wave is coded into discrete nerve-action potentials. As any information lost during this process is no longer available for neuronal processing, it is important to understand and quantitatively model the underlying principles. We have developed a detailed model of auditory processing, which codes sound signals into spike-trains of the auditory nerve. We have also developed Hodgkin-Huxley models of cochlear nucleus neurons, which are driven by auditory nerve spike-trains. We analyze the quality of coding with the framework of automatic speech recognition and the temporal information processing capabilities with the methods of information theory. Our latest improvements in speech coding by introducing the effect of offset-adaptation together with an improved matching of neuronal features to the speech recognizer using an artificial neuronal network has lead to significant improvements of recognition scores, now reaching the values of successful technical feature extraction methods. Offset adaptation is also required to drive onset neurons in the cochlear nucleus, which are able to code temporal information with sub-millisecond precision ( 0.02 ms). Our results provide quantitative insight into temporal processing strategies of neuronal processing and are highly relevant for cochlear implants.

Holsteyns, F.

• Comparison of cavitation bubble arrays at different frequencies
  A. Otto, T. Nowak, R. Mettin, F. Holsteyns and A. Lippert [I]
[A]

Arrays of bubbles are formed in gassy water under sonication at different frequencies from 20 kHz to 1 MHz. We investigate details of the structures with respect to similarities and frequency scaling. The experimental methods include high-speed imaging, acoustic pressure and luminescence measurements. Visually the structures appear rather similar: bigger ''degassing'' bubbles gather in arrays close to the pressure nodes, and smaller bubbles show a fast motion inbetween. Bubble sizes, velocities, and densities are measured, and we try to highlight also differences between the structures at different acoustic frequencies from an experimental and theoretical point of view. One goal of the study is to reveal and understand qualitative transitions in acoustic cavitation bubble structure formation for more and more increased driving frequency.

Holube, I.

• Interactive Fitting-Wizard in a Home Environment
  H. Baumgartner, A. Schulz, A. Hein, T. Herzke and I. Holube [C]
[A]

The European Hearing at Home (HaH) project aims at supporting hearing-impaired persons in their home environment. TV headsets with linear volume control allow people with hearing deficiencies good speech perception even in bad acoustic conditions. This approach attacks sensorineural hearing losses which imply a changed loudness perception combined with reduced perceivable dynamic range.

To help sensorineural users adapting the dynamic of their head phones according to individual hearing loss, three different interactive fitting procedures - based on broadcast audio material - have been evolved to fit parameters of a three band dynamic compressor performed in the master hearing aid (MHA). The MHA is the real-time audio-signal-processing framework in the background.

The fitting schemes can be used without any expertise and are intuitive and easy to use. A full walkthrough of the fitting wizard takes about 10 to 15 minutes. First results of user studies with 11 subjects (54-73 years old, 5 women, 6 men, slight to moderate hearing loss) show effectiveness of the approaches with respect to the improvement of the speech intelligibility and acceptance of sound quality. Further user studies will be performed.

Hooghwerff, J.

• Road Traffic Noise of Expansion Joints – Get a Grip on it
  W.-J. van Vliet, J. Hooghwerff, N. Booij and R. van Loon [C]
[A]

Expansion joints used in road traffic bridges can be a serious source of noise annoyance. This is caused by the impulsiveness of the sound of vehicles passing over the joints and sometimes the presence of low frequencies in the sound. In the Netherlands large-scale application of low noise road surfaces creates a need for low noise expansion joints. Therefore Rijkswaterstaat has created a guideline to deal with noise of expansion joints. The goal is to avoid excessive noise problems by setting noise limits in contracts. The approach in the guideline is based on existing calculation and measurement methods with which there is already considerable experience in the field of low-noise road surfaces. The application of these methods to the noise caused by expansion joints is new.

Horacek, J.

• PIV Measurement of Flow-Patterns in a Human Vocal Tract Model
  J. Horacek, P. Sidlof, V. Uruba, J. Vesely, V. Radolf and V. Bula [I]
[A]

The contribution describes developed complex physical model of the voice production that consists of the trachea, the self-oscillating vocal folds and the vocal tract with acoustical spaces corresponding to the vowel /a:/. The vocal tract model was developed from the 3D FE models designed from MR images The vocal folds were made of a latex thin cover layer filled by a polyurethane rubber, joined to the subglottal spaces. The time-resolved PIV method was used for instantaneous velocity field evaluation. Preliminary results will be presented for measurements performed within a physiologically range of mean air flow rates and fundamental frequencies. The vibrating vocal folds during one oscillation period were recorded by the high-speed camera in the same instants as the airflow velocity patterns. Flow structures resembling large vortices with dimensions comparable with the channel cross-section is possible to see above the ventricular folds. The vortices disappear in the narrowest epilaryngeal part of the vocal tract where the flow is uniform. Large eddies of a size comparable with the channel height were observed in the model of the mouth cavity where the airflow was attached to one wall of the channel resembling a Coanda effect.

Horvat, M.

• Influence of Short Term Noise on Concentration and Human Performance
  M. Horvat, K. Jambrosic and H. Domitrovic [C]
[A]

It is well known that short-term exposure to noise causes fatigue, insomnia, headache, loss of concentration and the occurrence of negative emotions. In order to examine certain aspects of this problem, a subjective testing has been made with the emphasis on the ability of human beings to maintain concentration and to perform well in a noisy environment. The test itself included the evaluation of the subjects' psychophysical state before and after completing the required task in order to examine the changes in psychophysical state caused by exposure to noise. In the central part of the test the subjects were asked to solve a simple task that did not require any specific knowledge, only a certain amount of concentration. Simultaneously, the subjects were exposed to four types of artificially generated noise. The type of noise varied from group to group and the loudness of the signal varied from one test séance to the next. The results of the tests reveal that the subjects were able to maintain their concentration and a certain level of performance when exposed to short-term noise, although they experienced significant changes in their psychophysical state.

• The design of an in-situ absorption measuring system using the Adrienne method
  K. Jambrosic, M. Horvat and M. Suhanek [C]
[A]

The Adrienne method, standardized in the CEN/TS 1793-5, is a frequently used measurement technique developed for outdoor in-situ measurements of sound reflection and airborne sound insulation. The data given in the standard was used as the basis for the design of a measuring system to be used for determining the absorption coefficient and sound insulation of noise barriers and built-in roofs and wall systems. In this paper the design process is described, specifically, solving of encountered construction problems and system optimization in order to reduce the measurement errors and to eliminate elements that introduce uncertainty. As stipulated in the standard, a fixed source-microphone combination has been achieved, in our case by mounting the microphone holder at the end of a bar kept tightly in position by cables attached to the loudspeaker cabinet. The unit is mounted on a swivel that allows variations of both azimuth and elevation in the range of 40 degrees from the reference axis in 10-degree steps. The excitation signals used in the measurements included the MLS and the sine sweep. The post processing of the measured impulse responses for various rotation angles and source positions was done in Matlab, developed especially for the Adrienne method.

• Evaluation Of Acoustical Suitability Of Lecture Rooms With Respect To The Quality Of The Sound Reinforcement System And The Level Of Background Noise
  M. Horvat, H. Domitrovic and S. Grubeša [C]
[A]

In order to evaluate acoustical suitability of several lecture rooms located on the Faculty of EE and Computing in Zagreb, Croatia, a questionnaire was given to the students who take their classes in those particular rooms. The first part of the investigation is focused on the quality of sound reinforcement system located in two largest lecture halls, to be replaced in the near future with a new one of significantly better quality. The goal is to examine whether this change is reflected in the grades of acoustical suitability given by the students. The second part of the examination deals with two smaller lecture rooms of identical size and acoustic treatment, but with significantly different levels of background noise due to positions of these rooms relative to the nearby street as the source of traffic noise. The overall level and spectral content of noise are measured in these lecture rooms. In addition, speech intelligibility is examined in all four rooms. The results of these measurements are then interpreted with respect to the subjective grades given by the students.

Houben, D.

• Investigation of voice production in death metal singers
  C. Eckers, D. Hütz, M. Kob, P. Murphy, D. Houben and B. Lehnert [I]
[A]

Singers in ''metal'' style often use a unique form of voice production called ''growling''. The underlying mechanisms of voice production as well as possible potentially harmful influences on the singer's voice have not yet been intensively studied. Supra glottal laryngeal constriction appears to play a major role in metal singing. This constriction results in difficulties to apply standard laryngeal endoscopy for visualization of vocal fold function during this particular phonation style. In this study we present an examination method using a questionnaire including the voice handicap index as well as a combined simultaneous endoscopy, electroglottography and acoustic analysis to investigate in particular the production of death metal singing in seven subjects. None of the subjects complained about persistent voice problems and none had signs of secondary organic voice disorders like vocal fold nodules or inflammation. However, most showed minor irregularities on stroboscopic examination. Further findings reveal that at least two forms of supra glottal laryngeal constriction are used to obtain the typical death metal sound which differ in use of vocal folds and ventricular folds as well as in supra-glottal adduction effort. The study shall help to answer the question whether death metal singing can be performed such that the risk of voice damage can be reduced or avoided.

Houtgast, T.

• The acoustical engineer as a researcher in speech and hearing
  T. Houtgast [I]
[A]

There is a long tradition of acoustical engineers working in the field of speech and hearing. Although it is well recognized that speech reception, and hearing in general, involves cognitive processes beyond the reach of the engineer, this engineering approach has led to interesting results which will be reviewed briefly. For the engineer, speech is often described in acoustical terms as a succession of sounds with specific properties in terms of spectra and decibels. It has been possible to identify the acoustical cues in a speech signal which are essential for speech intelligibility. These cues are related to the dynamics, or the pattern of fluctuations, in the speech signal. By applying a specific type of speech analysis, the strength of these cues in an ongoing speech signal can be quantified. It will be shown that the effect of speech transmission and room acoustics on speech intelligibility can be estimated successfully by measuring the degree of preservation of these cues in the speech signal received by the listener.

• Cognitive Factors in Speech Perception
  A. Zekveld, S.E. Kramer and T. Houtgast [I]
[A]

Authors: Adriana Zekveld, Sophia Kramer, Tammo Houtgast

Both auditory, modality-specific and cognitive, modality-aspecific abilities influence the ability to perceive and comprehend speech in noisy listening situations. Important cognitive factors contributing are information processing speed and working memory processes that allow the use of linguistic context to fill in incomprehensible 'gaps' in the speech. When listening conditions are challenging (due to noise or hearing loss), listeners rely to a greater extent on effortful top-down processes to improve speech comprehension. Therefore, it is important to measure both auditory and cognitive abilities when examining the causes of speech comprehension problems. To prevent the performance on cognitive tests from being confounded by hearing loss, visual tests of cognitive functions need to be considered when testing hearing impaired participants. Therefore, our group recently developed the Text Reception Threshold (TRT) test, a visual analogue of the Speech Reception Threshold test. Several studies showed that speech comprehension in noise is associated with the TRT, indicating that both abilities rely on similar modality-aspecific cognitive processes. Future studies should focus on further unravelling the auditory and cognitive functions in speech comprehension and on the development of a clinically applicable test of relevant cognitive functions in speech comprehension in noise.

Hudde, H.

• Fundamental sound fields in the core region of curved ear canals
  H. Hudde and S. Schmidt [C]
[A]

The core region of ear canals denotes the section between the ''entrance'' and the beginning of the ''drum coupling region'' (DCR). The extension of the core region is specified by the absence of higher order modes. Within the core region usually pure fundamental modes are assumed which follow a middle axis of the ear canal. Such modes are characterized by regular surfaces of constant sound pressure (''isosurfaces''). Regular isosurfaces completely cover the aperture of the ear canal, and the middle axis is normal to the surface. However, Finite Element computations reveal that in the core region irregular surfaces of constant sound pressure occur. Such ''one-sided isosurfaces'' are not oriented to the middle axis, but form dome-shaped surfaces oriented to a point on one side of the ear canal wall. In the talk the origin of ''one-sided isosurfaces'' will be elucidated. Although one-sided isosurfaces cannot be described as fundamental modes, it is reasonable to consider these local disturbances as being part of a ''fundamental sound field''. It can be shown that one-sided isosurfaces are widely governed by the principle of energy minimization which also applies to regular fundamental modes.

• Features of sound fields within and outside the ear canal
  S. Schmidt and H. Hudde [C]
[A]

In the sound field between pinna and eardrum, three regions with different characteristics can be distinguished: (a) The external field is determined by the pinna geometry and the source. Surfaces of equal pressure (''isosurfaces'') have distinctly three-dimensional structure. (b) In the core region, the isosurface shape does not depend on the source and is fairly regular. (c) Near the drum, higher order modes are excited by its vibration, thus the local field does not belong to the core region. A finite element model of a natural pinna and ear canal terminated with a vibrating eardrum was used to study the sound field features in detail and to check several conditions usually assumed when using one-dimensional models. It turned out that an entrance surface of the core region can actually be specified although it depends to some extent on the wave incidence direction and on frequency. Isosurfaces of pressure magnitude and phase are not always coinciding. Close to the drum, the field is composed of the expected higher order modes and of isosurfaces that are oriented orthogonal to the membrane.

Hüppe, A.

• Limits of Finite Element Methods in the Mid-Frequency Range
  A. Hüppe and M. Kaltenbacher [I]
[A]

Field simulation methods like the Finite Element Method (FEM) in computational acoustics offer some advantages in comparison to energy based methods. For example, they can be applied in the frequency as well as in the time domain and the solution for acoustic pressure and/or velocity is available at every point inside the simulation domain. On the other hand, especially when considering higher frequencies, a FEM simulation requires a huge amount of computational time and memory, thus limiting its range of application to smaller geometries or lower frequencies.

In our contribution we focus on the simulation of the acoustic field in the time domain as the solution provides the whole frequency range of interest in one simulation. Therefore, we consider h-FEM and p-FEM applied to the wave equation and the Spectral Finite Element Method (s-FEM) applied to the governing equations of acoustics. We perform simulations in the mid-frequency range to show the limits of the methods in terms of accuracy of the solution and computational costs. As a practical application, the methods are used to simulate the sound field in the environment of a noise barrier.

• Computational Acoustics based on Conservation Equations using Taylor Hood Finite Elements
  M. Kaltenbacher and A. Hüppe [I]
[A]

Within this contribution we concentrate on the efficient solution of enhanced wave problems, which are mathematically modeled by their conservation equations, namely the mass and momentum conservation. Such formulations arise in computational aeroacoustics or in nonlinear acoustics. Even for acoustic wave propagations, where an acoustic wave equation exists, formulations based on the conservation equations have the advantage, that the numerical solution provides directly the two main engineering quantities, the acoustic particle velocity and the acoustic pressure.

The application of the standard Finite Element (FE) method to the acoustic conservation equations will result in an instable formulation. In order to obtain stable finite elements, we need a mixed FE formulation. The way to obtain such a formulation is to use H(div) functions for the approximation of the particle velocity and L2 functions for the acoustic pressure. However, it can be proofed, that approximating the acoustic particle velocity by Lagrangian finite elements one order higher than those for the acoustic pressure will also provide us a stable formulation.

We have applied this so-called Taylor-Hood elements to the acoustic conservation equations and will present their applicability to general wave problems. In detail, we will present a study of the accuracy and computational efficiency.

Hütz, D.

• Investigation of voice production in death metal singers
  C. Eckers, D. Hütz, M. Kob, P. Murphy, D. Houben and B. Lehnert [I]
[A]

Singers in ''metal'' style often use a unique form of voice production called ''growling''. The underlying mechanisms of voice production as well as possible potentially harmful influences on the singer's voice have not yet been intensively studied. Supra glottal laryngeal constriction appears to play a major role in metal singing. This constriction results in difficulties to apply standard laryngeal endoscopy for visualization of vocal fold function during this particular phonation style. In this study we present an examination method using a questionnaire including the voice handicap index as well as a combined simultaneous endoscopy, electroglottography and acoustic analysis to investigate in particular the production of death metal singing in seven subjects. None of the subjects complained about persistent voice problems and none had signs of secondary organic voice disorders like vocal fold nodules or inflammation. However, most showed minor irregularities on stroboscopic examination. Further findings reveal that at least two forms of supra glottal laryngeal constriction are used to obtain the typical death metal sound which differ in use of vocal folds and ventricular folds as well as in supra-glottal adduction effort. The study shall help to answer the question whether death metal singing can be performed such that the risk of voice damage can be reduced or avoided.

Hufenbach, W.

• Design and Tests of Thermoplastic Textile-reinforced Composite Trays for Vibro-acoustic Relevant Applications
  W. Hufenbach, M. Dannemann, S. Friebe, F. Kolbe and M. Krahl [C]
[A]

Driven by increasing customer demands, environmental and economical requirements, lightweight engineering gains in importance for various industrial and transport applications. Here, composite materials based on thermoplastic hybrid yarns offer new possibilities for multifunctional design due to its adjustable properties. Requirements concerning high stiffness and damping as well as low weight and short cycle times during manufacture can be met by an adapted composite lay-up and manufacturing process. In the presented work, the vibro-acoustic behaviour of selected textile-reinforced thermoplastics was characterized dependent on the kind of reinforcement, number of layers and its fibre orientation. Therefor, models based on Finite Element Method and Boundary Element Method for the calculation of modal and acoustic parameters of textile-reinforced structures were developed. For the verification of this simulation models as an example composite trays were manufactured by a novel hot pressing technology. The modal parameters and transmission loss of the produced trays were measured. Subsequently, the experimental results were used for the verification of the numerical simulations. The performed investigations illustrate that the vibro-acoustic behaviour of composite trays is influenced by the complex interaction of geometry and material parameters. Thus, an optimal sound radiation behaviour can only be achieved by a comprehensive material-adapted vibro-acoustic design.

• Identification of the Optimal Lay-up for Fibre-reinforced Composite Plates with Minimized Sound Emission
  M. Dannemann and W. Hufenbach [I]
[A]

Increasing customers requirements and environmentally demands, the shortened development times as well as the necessary weight reduction and function-integration are pressing influence factors for the design of future lightweight structures. In the course of these challenging demands, composites with fibre- or textile-reinforcements offer new possibilities for a function-integrated design due to their versatile property profile. The big amount of adjustable parameters like layerwise fibre orientation, thickness and material requires the utilization of efficient optimization methods, which take into account not only the stiffness but also the vibration and sound behaviour. For this, an optimization tool based on a semi-analytical model for the calculation of the sound radiation of composite plates was developed. The design tool enables the calculation of optimal parameters to given requirements. Within this paper the results of selected optimization scenarios are presented including the influence of frequency/stiffness limits and mass restrictions on the optimal composite lay-up with minimized sound radiation. The developed tool and the performed studies allow the derivation of design guidelines giving the engineer the possibility of fully exploit the lightweight and vibro-acoustic potential of composite materials.

Huijssen, J.

• Modeling Pulsed Nonlinear Ultrasound for Medical Applications: the INCS Method
  M. Verweij and J. Huijssen [I]
[A]

The higher harmonics in nonlinear ultrasound fields gain importance for improving the image quality of medical echoscopy. Consequently, the optimization of medical imaging techniques has become highly dependent on the accurate simulation of pulsed nonlinear ultrasound fields in tissue. In this context, a successful simulation method should be able to compute the pulsed ultrasound field of a phased array transducer in a nonlinear and attenuating medium. Moreover, the method should cope with a large-scale domain (approximately 100 wavelengths in three spatial directions and 100 periods in time) and wide-angle propagation. Current methods are either based on approximations that favor a specific propagation direction or require prohibitive large grids. This paper presents the Iterative Nonlinear Contrast Source (INCS) method, which meets all above requirements and provides a full-wave solution of the Westervelt equation. This method treats the nonlinear term from the Westervelt equation as a distributed contrast source. The field is iteratively updated by convolution of an estimate of this nonlinear contrast source with the Green's function of the linear wave equation. Moreover, the performance of the INCS method is discussed by showing the iterative generation of the higher harmonics, the directional independence, and the simulation of a highly steered beam.

Huo, L.

• Attribute-based Instrumental Assessment for Speech- Transmission Quality
  L. Huo and U. Heute [C]
[A]

Hyder, M.

• Measurements of Sound Localization Performance and Speech Quality in the Context of 3D Audio Conference Calls
  M. Hyder, M. Haun and C. Hoene [I]
[A]

The simulation of virtual acoustic environments helps to increase the quality of teleconferencing. Teleconferencing solutions supporting 3D audio have two advantages. First, the talkers can be identified easily by locating their position in the virtual acoustic environment. Second, if multiple persons talk at the same time, humans can focus on talkers of their choice by taking advantage of the cocktail party effect. However, 3D audio decrease the speech quality by decreasing loudness and adding reverberation and echoes. In this publication we study the trade off between speech quality and the ability of focusing and locating the talkers. We extend our initial work presented in the ITU Workshop in Lannion, Sep. 2008, in which we study the 3D audio rendering engine by Raine Kajastila et al. (AES 122th Convention, May 2007). We conduct subjective and objective listening-only tests using different head related transfer functions (HRTF), different algorithmic complexities, different positions of the conference call participants, and different geometries of the virtual environment. We present the statistical significant test results for sound localization performance and the speech quality. This study helps us to design our teleconference solution that we call 3D Telephony.

Imagawa, H.

• Estimating physical properties of vocal fold paralysis from high-speed filimg data
  I. Tokuda, M. Kimura, H. Imagawa, K.-I. Sakakibara and N. Tayama [C]
[A]

Pathological voice originates from complex nonlinear vibrations of the disordered vocal folds. Deep understanding of the voice disorders based on mathematical modeling of the vocal folds may provide an important insight into the abnormal vocal fold vibration as well as a good hint for improving the voice surgery. Up to date, various models have been developed to simulate complex vibratory patterns of the vocal folds, ranging from a simple two-mass model to complex multi-mass models. However, only a few researches aim at construction of the vocal fold model, which realizes quantitative characteristics of the individual voice. Recently, Doellinger et al. proposed a method for estimating the parameters of asymmetric two-mass model for high-speed filming data recorded from human subjects. The present paper applies this technique to pathological data recorded from patients with vocal fold paralysis. Right and left vocal fold tensions, subglottal pressure and glottal area have been estimated. Whether the effect of the voice surgery is reflected in the estimated parameters is judged by using the data before and after the voice surgery. On the basis of the estimation results, applicability of the mathematical model as a simulator for the voice surgery is discussed.

Imbert, C.A.C.

• The Aubrapan: Revival of a lost invention
  S. Maloney, C.A.C. Imbert and A.G. Bryan [C]
[A]

The Aubrapan is a new design of the melody (or soprano) pan developed in 1978 by Guyanese pan maker and panist Aubrey Bryan in the United Kingdom. In the Aubrapan the higher frequency notes are placed nearest the rim of the pan in contrast to the center of the pan where most high notes are generally located. The lower octaves are placed directly opposite each other in order to facilitate the pendulum-like action of the arms - these ascend in whole tones. This concept facilitates easier and rapid playing of the chromatic scale in single and double note form. In this paper we describe the development of the Aubrapan and an attempt is made to discover why the Aubrapan has not yet found a place among the steelpan family. To answer this and other questions, the experiences of the Aubrapan's inventor are noted and a comparison of the Aubrapan's playability is made with that of the conventional Trinidad and Tobago 4th's and 5th's soprano pan.

Isbrücker, D.

• Variability and Adaptability of the Acoustics in the new Conservatory of Amsterdam
  P. Heringa, D. Isbrücker and M. Valk [I]
[A]

Last summer the newly built conservatory for music of Amsterdam was delivered. The conservatory has 5 larger halls. The primary use of these halls is rehearsal and exams. They are also used for musical performances. The halls are quite different in order to learn the students to cope with different acoustics. They range from a hall suitable for rehearsal of an symphonic orchestra to a recital hall, a hall for jazz and pop and a hall for opera and musical. The acoustics of three halls can, to a certain extend, be adapted to the use of the halls. The acoustics aimed and achieved and the variation possibilities will be subject of the invited paper. Furthermore the conservatory has more then 100 education rooms and about 70 study rooms. The acoustics of these rooms can, by means of ''cushions'', be adapted to the lessons scheduled. The needed adaptability was predicted by means of experiments together with students and teachers in a mock-up of ''new rooms'' in the former conservatory building. They had to express their preference and give their evaluation of different subjective parameters. The adaptability and the evaluation of it will be shown.

Isele, A.

• In-situ measurements of road barriers made of natural stones
  A. Isele, G. Höfker and C. Nocke [C]
[A]

With an increasing disturbance of traffic noise the length of barriers along highways and railways grows year by year. Besides the well-known materials like concrete, metal, glass or synthetic stones also natural stones are more and more in use because of its appearence.

A minimum value for sound absorption coefficient and transmission loss is necessary for the acoustic efficiency of the barrier. Measurements of the acoustic properties of gabions in the reverberation chamber and laboratory are often impossible because of the construction or the weight respectively. To avoid these problems the reflection index and the sound insulation index were determined by in-situ measurements according to DIN EN ISO 1793, part 5. Results of the reflection index for several materials, grain sizes and layer thicknesses of the charge of the wire mesh are presented.

Earlier research have shown that a minimum value of the transmission loss could be reached with additional elements inside the gabion only. Different constructions of concrete with and without slits between the elements have been tested. A mathematical description of the problem is given, which shows the essential parameters of sound insulation and allows approaches to optimisation.

Isik, M.

• Coding of speech into nerve-action potentials
  M. Holmberg, H. Wang, M. Isik and W. Hemmert [C]
[A]

One of the most critical processing steps during encoding of sound signals for neuronal processing is when the analog pressure wave is coded into discrete nerve-action potentials. As any information lost during this process is no longer available for neuronal processing, it is important to understand and quantitatively model the underlying principles. We have developed a detailed model of auditory processing, which codes sound signals into spike-trains of the auditory nerve. We have also developed Hodgkin-Huxley models of cochlear nucleus neurons, which are driven by auditory nerve spike-trains. We analyze the quality of coding with the framework of automatic speech recognition and the temporal information processing capabilities with the methods of information theory. Our latest improvements in speech coding by introducing the effect of offset-adaptation together with an improved matching of neuronal features to the speech recognizer using an artificial neuronal network has lead to significant improvements of recognition scores, now reaching the values of successful technical feature extraction methods. Offset adaptation is also required to drive onset neurons in the cochlear nucleus, which are able to code temporal information with sub-millisecond precision ( 0.02 ms). Our results provide quantitative insight into temporal processing strategies of neuronal processing and are highly relevant for cochlear implants.

Jabben, J.

• Background noise: an increasing environmental problem
  J. Jabben, D. Bergmans, E. Schreurs and T. Koeman [C]
[A]

Noise quality in most studies is related to the average noise level (Lden) only, although this indicator merely expresses the average intensity of the soundscape. A complementary feature of increasing importance is the background noise level, which is the noise level that is exceeded during 95 % of the observation time. Inhabitants of areas with a high background noise level may be more vulnerable to annoyance or health effects, as recuperation times are narrowed. Due to strong growth of road, railway and air traffic, background noise levels have steadily increased in the Netherlands. In order to assess the impact of environmental noise levels from transport sources, RIVM in cooperation with the Dutch National Aerospace Laboratory have started a research project in 2007 with the aim of setting up background noise maps for road-, railway- and air traffic. This paper discusses the mapping and characteristic behavior of background noise. The implications of the background noise levels on health and annoyance will be discussed. The results indicate that a further increase of background noise levels in the near future is to be expected that may affect the Dutch population. Hence an intensified reduction policy of noise emission is most urgent.

• Breaking the Sound Barrier: Why Environmental Noise is Still Not Reduced Properly, and What To Do About It
  E. Verheijen and J. Jabben [C]
[A]

Various international cost-benefit studies show that traffic noise can be reduced cost-effectively by noise measures, especially by measures taken at the source (i.e. silent tyres or silent road pavement). The benefits are obvious: a quieter environment leads to less annoyance and sleep disturbance, resulting in reduced health damage and production loss. Previous research pointed out that 7 billion euros can be spent cost-effectively on noise measures in the Netherlands. This does not only justify the present national noise policy (which aims at avoiding noise growth), but also suggests that much more ambitious goals are within reach: an overall noise reduction of 5 dB. How to convince decision makers and politicians that such a noise reduction is both urgent and cost-effective? Why do they still hesitate while environmental noise is annoying so many people and threatens their well-being and health? The present article analyzes the doubts and it provides actions that should be considered to take the noise issue higher on the political agenda.

Jackson, P.

• A model of jet modulation in voiced fricatives
  A. Barney and P. Jackson [I]
[A]

The radiated sound during the production of voiced fricatives has harmonic and random components. The harmonic components are associated with the vibration of the larynx and the noise components are associated with the time-varying flow through a constricted region of the oral cavity. The two sources interact non-linearly to produce a noise component with an amplitude envelope modulated at the fundamental frequency of voicing. This paper describes pressure measurements made in a dynamic physical model of the larynx and vocal tract, which were used to study the aerodynamic and acoustic behaviour in a flow regime corresponding to that found during voiced fricative production. An orifice plate and a sharp-edged obstacle downstream model the geometry of the oral cavity, representing the tongue constriction and teeth respectively. The depth of modulation of the frication noise was found to be linearly related to the pressure upstream of the orifice plate. The phase of the noise bursts relative to the upstream pressure was found to vary systematically with the constriction-obstacle distance. These findings form the basis of a parametric model of voiced fricative production and give some insight into the likely aeroacoustic mechanism responsible for sound production.

Jaeckel, O.

• Comparision of various Array Geometries with Respect to their Depth of Field for acoustic Mappings
  O. Jaeckel, R. Schröder and D. Döbler [I]
[A]

The wide-spread use of planar microphone arrays for acoustic 2D-beamforming mappings does not consider the influence of actual threedimensional object surface structures on the calculated maximum sound pressure levels. To quantify the effects of the very complex interactions between array- and objectgeometry in more detail, several investigations of common two- and threedimensional arrays with varying focus distances have been performed. All simulations were done using the correct spherical wave model for point sources for the distance calculations. In the simulations, ring- and spiral arrays as well as spherical 3D-arrays will be compared. Measurement examples for some of the arrays will also be shown and will be related to the simulations. The results show that most arrays have a remarkably strong dependence on the focal distance actually used for the beamforming calculation. Especially for the case of beamforming within 3D-interiours, which is very difficult to handle due to strong individual distance variations, it is therefore inevitable to take the actual object-geometry into account to get useful results at all. So the use of 3D-CAD-models for acoustic 3D-mappings is a well-founded decision, even if it makes software-implementations much more complex than the easy but oversimplifying use of just planar beamforming maps.

• Threedimensional Beamforming considering the Effects of covered Microphones
  O. Jaeckel, D. Döbler, R. Schröder and A. Meyer [I]
[A]

Beamforming onto complex geometric structures in three dimensions must take into account surface areas from which the emitted soundwaves can not reach the microphone array directly. Similar to optics, the beamformer is faced with the problems of ''full shade'' (no microphone has a direct path to the acoustic source) and ''half shade'', i.e. the acoustic source is only visible partly by a certain subset of all microphones. The article presents two methods to geometrically take those areas into consideration in 3D-beamforming. The first method is a relatively simple one that can only determine either full shading or full visibility of some parts of the object geometry. It gives good results in most measurement scenarios when the microphone array is compact or has a comparatively small diameter, e.g. with a small spherical array in car interiours. The second solution presented is computationally far more intensive but also more general. This method does also consider the ''half shade'' baffle effects and therefore it also allows measurements with distributed microphone arrays onto 3D-structures from the outside. The necessary modifications of the beamforming-algorithm for both cases are presented and the resulting effects will be discussed using practical measurement examples.

Jaillet, F.

• On the Structure of the Phase around the Zeros of the Short-Time Fourier Transform
  F. Jaillet, P. Balazs, M. Dörfler and N. Engelputzeder [C]
[A]

The short-time Fourier transform is a time-frequency representation widely used in audio signal processing. At first sight, the phase of the transform may appear to be difficult to interpret, a fact that led to the development of many processing schemes in audio processing relying on the modulus of the transform only. However, it has been shown, that the phase information can be successfully exploited for improved analysis and processing. This is illustrated for example by the reassignment method or the phase vocoder algorithm.

In order to better exploit the phase information in audio processing, a better understanding of its structure is desirable. This study describes a particular behavior of the phase which can be observed when looking at the phase derivative over time or frequency around the points where the transform equals zero. This recurrent behavior is illustrated using numerical simulations on synthesized and real audio signals. The performed simulations show the influence of the parameters of the transform on the observed structures. Some consequences of this phase structure for audio processing are also presented.

• Removing Components from a Time-Frequency Representation
  M. Dörfler, P. Balazs and F. Jaillet [C]
[A]

Time-frequency representations such as the spectrogram or short-time Fourier transform seem to be well suited to the task of removing certain components with approximately disjoint support in the time-frequency domain. For example, one might be interested in suppressing a certain instrument's contribution from a music signal. Such approaches are used in Computational Auditory Scene Analysis by the name of Time-Frequency masks. However, the trivial approach of just deleting the corresponding component in the time-frequency representation leads to artefacts such as ''ghost-tones'', i.e. a modified version of the deleted component is still audible, or to the perception of unnatural sound coloration comparable to a ''phasing effect''. We compare different approaches to tackle this problem: a soft-thresholding procedure is applied and compared to a method using statistical models. On the other hand, we draw conclusions from considering the optimal separating mask for known signal components. In this situation, a model promoting sparsity in the representation can lead to favourable results by yielding a feasible approximation of the optimal mask. We perform informal listening experiments to evaluate the resulting removal of certain components in a complex audio-signal.

Jakob, A.

• Sound Radiation of Double Reed Woodwinds
  F.B. Konkel, A. Jakob, F. Heintze and M. Möser [C]
[A]

In the recent past the verification of radiating properties by woodwinds was measured by using microphones next to the woodwind holes. Hence, it was of particular importance to use suitable microphone positions between the woodwind holes and free space. In the present contribution the radiation properties of the woodwind holes and the bell of a bassoon are investigated by using the ''acoustic camera''. For the acoustic testing the instrument is observed by three directions. The examination of the results verify the known ''cut off'' frequency of woodwinds at medium frequencies. From the experimental investigation, two distinct new frequency ranges are found for the radiation properties of the bassoon. The first located range between the medium and the high frequencies of the bassoon modes shows that the central radiation is emitted by the bell, but the last open woodwind holes are also a part of the sound radiation. The second noticed range is located at high frequencies and it becomes clear that the sound radiation is only emitted by the last open woodwind holes. However, there is no sound radiation by the bell at these frequencies.

• Active absorption systems: Study and implementation of an adaptive control procedure.
  M. Norambuena, A. Jakob and M. Möser [C]
[A]

The present work follows the intensity measurement technique developed by Fahy to conduct a study into active absorption systems and its practical implementation. The method separates the total sound field into two parts; an incident () and a reflected () part. This is valid in simple cases, such as a source radiating sound over a rigid surface which in response generates a reflected wave. For cases in which the wave propagates in one dimension the separation is realized efficiently, however for cases in which the wave propagates in 2 or 3 dimensions the efficiency diminishes. Measurements of absorption coefficients for pure tones as well as broad band noise were made to evaluate the efficiency of the system. The practical implementation of the system was performed using an adaptive FX-NLMS filter programmed inside a DSP chip.

Jakob, H.

• Validation of a measurement based Area Contribution Analysis
  H. Jakob, A. Bocquillet, S. Marburg and H.-J. Hardtke [C]
[A]

The Automotive Industry is in need of effective NVH measurement methods to optimize the design of isolation and damping concerning weight and cost constrains. In order to gain better knowledge on the contribution of each panel from the interior it is imperative to find a new and simple method of analysis in the early prototype stages of a vehicle. Using the technology of the Microflown-element it became possible to grasp the particle velocity close to any surface based on measurements. This information and the quantification of the transfer path allow for a prediction on contribution from subsystems in the cabin. The method is based on a combination of the Transfer Path Analysis principle and the quantification of the Emission of the surface. A simple example below shows the validation of such measurements as well as the comparison between a direct measurement and a simulation. This example depicts the critical range of the mid frequencies (200-3500Hz). The precision and usability of these simulated results will be discussed moving forward based on this specifically designed setup.

Jambrosic, K.

• Influence of Short Term Noise on Concentration and Human Performance
  M. Horvat, K. Jambrosic and H. Domitrovic [C]
[A]

It is well known that short-term exposure to noise causes fatigue, insomnia, headache, loss of concentration and the occurrence of negative emotions. In order to examine certain aspects of this problem, a subjective testing has been made with the emphasis on the ability of human beings to maintain concentration and to perform well in a noisy environment. The test itself included the evaluation of the subjects' psychophysical state before and after completing the required task in order to examine the changes in psychophysical state caused by exposure to noise. In the central part of the test the subjects were asked to solve a simple task that did not require any specific knowledge, only a certain amount of concentration. Simultaneously, the subjects were exposed to four types of artificially generated noise. The type of noise varied from group to group and the loudness of the signal varied from one test séance to the next. The results of the tests reveal that the subjects were able to maintain their concentration and a certain level of performance when exposed to short-term noise, although they experienced significant changes in their psychophysical state.

• The design of an in-situ absorption measuring system using the Adrienne method
  K. Jambrosic, M. Horvat and M. Suhanek [C]
[A]

The Adrienne method, standardized in the CEN/TS 1793-5, is a frequently used measurement technique developed for outdoor in-situ measurements of sound reflection and airborne sound insulation. The data given in the standard was used as the basis for the design of a measuring system to be used for determining the absorption coefficient and sound insulation of noise barriers and built-in roofs and wall systems. In this paper the design process is described, specifically, solving of encountered construction problems and system optimization in order to reduce the measurement errors and to eliminate elements that introduce uncertainty. As stipulated in the standard, a fixed source-microphone combination has been achieved, in our case by mounting the microphone holder at the end of a bar kept tightly in position by cables attached to the loudspeaker cabinet. The unit is mounted on a swivel that allows variations of both azimuth and elevation in the range of 40 degrees from the reference axis in 10-degree steps. The excitation signals used in the measurements included the MLS and the sine sweep. The post processing of the measured impulse responses for various rotation angles and source positions was done in Matlab, developed especially for the Adrienne method.

Janse, E.

• Hearing and Cognitive Measures Predict Elderly Listeners' Difficulty Ignoring Competing Speech
  E. Janse [I]
[A]

Age-related hearing and cognitive factors may play a role in the difficulty elderly listeners have listening to one talker in the presence of a competing speaker, but it is not always easy to establish direct relations between speech performance and cognitive measures. The present study investigated how well (detection accuracy) and how fast (detection time) elderly listeners detected pre-assigned target sounds in the speech of a target talker, both in a single-talker condition and in a condition with a competing talker. If general cognitive decline contributes to problems with competing speech over and above hearing loss, a measure of visual selective attention might be related to performance in the present study. 39 Elderly listeners with varying degrees of hearing loss participated in the study. Apart from the effects of hearing loss, overall detection accuracy correlated with the selective attention measure: the more difficulty one had ignoring irrelevant visual information, the poorer one's speech processing performance. The response time analysis showed that individual hearing acuity and the attention measure predicted how fast listeners detected the pre-assigned sounds in both listening conditions. These results confirm that age-related cognitive factors contribute to the typical problems elderly adults report in daily speech communication.

Janssen, A.J.E.M.

• Estimating the velocity profile and acoustical quantities of a harmonically vibrating membrane from on-axis pressure data
  R. Aarts and A.J.E.M. Janssen [I]
[A]

Power series expansions are derived for acoustical quantities of a harmonically excited resilient, flat, circular radiator in an infinite baffle. These quantities are the sound pressure: on-axis, far-field, and at the edge of a radiator, as well as the reaction force on the radiator and the total radiated power. These expansions are obtained by expanding the velocity distributions in terms of orthogonal polynomials. For the rigid, simply supported and clamped radiators, this results in explicit, finite-series expressions involving (spherical) Bessel functions for both the on-axis and far-field pressure. In the reverse direction, a method of estimating velocity distributions from (measured) on-axis pressures by matching in terms of expansion coefficients is proposed. Together with the forward far-field computation scheme, this yields a method for assessment of loudspeakers in the far-field and of the total radiated power from (relatively near-field) on-axis data (generalized Keele schema). The relation between the radiated power as obtained from near-field measured sound pressure data via the theory on one hand and from a far-field measurement in a reverberation room on the other is highlighted.

Janssen, B.

• Early decay development in inside and outside reverberating spaces
  B. Janssen [C]
[A]

Reverberation is a complex phenomenon, which can be approached from various different physical approaches. Inside and outside spaces alike show distinct reverberant behaviour, which is analyzed and categorized cognitively to distinguish spaces. Studies investigating which aspects of reverberation are cognitively most relevant exist, but mostly restrict themselves to inside spaces. The question whether there is a cognitive continuity in the handling of inside and outside reverberating spaces led to a listening experiment using impulse responses from various spatial settings. In order to achieve a presentation of the perception space, 49 subjects were presented with 36 pairs of impulse responses, stemming from nine inside and outside spaces. They were asked to state the difference between the stimuli, and these judgements were processed by multidimensional scaling. In the resulting perception space, the first dimension was found to be most important, and related to the early decay development, which is very different for inside and outside situations. This implies that an investigation of the early decay process is a rewarding direction to follow, which may indicate new aspects to consider in the evaluation, but also simulation of reverberation.

Janssen, N.

• Acoustics of Muziekkwartier Enschede (NL)
  N. Janssen and L. van Luxemburg [I]
[A]

November 22, 2008 the new building ''Muziekkwartier'' (Music Quarter) in the town of Enschede (NL) is inaugurated. The building will be the new home of the Opera (Nationale Reisopera), the Theatre (Podium Twente), the School of Music (Muziekschool Twente) and the Stage for Popmusic (Atak).

The main auditorium is to be used for drama, opera and as a special: for opera production. The auditorium will be the second hall specific for opera in the Netherlands, after Amsterdam. The capacity of the hall is 1000 seats, the volume of the hall is 9500 m3. Varying the acoustics in the hall is solved mechanically. Movable sound reflectors moved in height, ground plan and vertical angle are the ingredients for variable acoustics meant for speech or opera. The variation in reverberation time is T = 1.3 to 1.6 sec. in combination with strong support of early reflections.

The combination of functions, performing simultaneously without mutual disturbance, leads to a high aim on the sound insulation in the building.

The acoustics of both the halls for popmusic were not only well-damped, but also pinpointed to performance in low frequencies and neutralizing the absence of visitors during sound checking.

Janssens, K.

• Innovative Approaches to Fast Transfer Path Analysis
  H. van der Auweraer, P. Gajdatsy, K. Janssens, P. Mas and L. Gielen [I]
[A]

Over the last decennia, Transfer Path Analysis (TPA) has become an accepted tool for NVH troubleshooting and internal load estimation. A major bottleneck preventing its even more widespread use in the actual vehicle development process is the test time to build the full data model, requiring extensive Frequency Response Function testing next to the obvious in-operation tests. Recent innovations on instrumentation and new procedures such as Fast and Multilevel TPA and Operational TPA address this limitation, however often introducing new constraints. For example, OTPA only requires operational response data at reference and target locations. But, being basically a transmissibility method, its application requires special care regarding the inherent ill-conditioning and reference-completeness problems. It essentially assesses a co-existence relationship between targets and references and does allow causality and contribution interpretations. Hence a new approach is proposed, using a parametric model for the estimation of loads. This makes the method scalable, enabling the engineer to use a simpler model based on a small amount of measurement data for quick troubleshooting or increase accuracy by using a more complex model together additional measurements. The method is applied to a customer problem and compared with the existing approaches, showing the advantages in real life.

Janssens, M.

• Helicopter detection: a fast computational method for long range sound propagation
  M. Janssens, E. Salomons and F. van der Eerden [C]
[A]

A person can detect the presence of a nearby flying helicopter in various ways. One of these is aural detection. This paper, in combination with an accompanying paper, presents an approach to assess the probability of detection in varying circumstances of for example the terrain and the meteorology. Main parameters are: 1. the helicopter noise source strength and directivity, 2. the long range sound propagation as a function of terrain level and meteorology, 3. the presence of masking sound sources near the observer, 4. the probability of observer response to the exposed sound field. The focus of this paper is on long range sound propagation. A computational method is presented that allows a quick assessment (typically within a few seconds) of the long range sound propagation (typically up to 20 km) around a source or a receiver. The method allows for arbitrary terrain altitudes, so that shielding by (multiple) hills and mountains is taken into account. Moreover, the effects of windspeed and temperature on sound ray curvature is included in a parametric way. Results of this fast computational method are compared with results from a numerical Parabolic Equation approach. Finally, results of the full approach (all four steps) are illustrated.

• Helicopter detection - Part 2 - Approach for the determination of the probability of detection
  B. Quesson, M. Janssens and F. van der Eerden [C]
[A]

A flying helicopter can be detected, without any high tech material, by an observer hearing the characteristic sound emitted by the rotor of any helicopter. This paper, in combination with an accompanying paper, presents an approach to assess the probability of aural detection in various circumstances. The main parameters that influence this probability are: 1. the helicopter's noise source strength and directivity, 2. the long range (several km) sound propagation which is heavily influenced by e.g. terrain level and meteorological conditions, 3. the presence of masking sound sources near the observer, 4. the probability of the observer's response to the exposed sound field. This paper focuses on the two last points. Given a certain sound level (based on points 1. and 2.), it presents an approach to quantify the probability of aural detection. The calculation is based on studies found in literature on human aural response to helicopter sound and on an estimate of the masking sound level at the observer's location. The masking sound will vary from rural to urban areas and estimates for these are given. The paper presents the calculation of an auditory threshold but will address especially the estimate of the ambient wind induced noise levels.

Jebasinski, R.

• Active Exhaust Silencers for Internal Combustion Engines
  J. Krüger, M. Pommerer and R. Jebasinski [I]
[A]

In the past years Eberspaecher has installed Active Exhaust Silencers on several personal vehicles with different engines on a prototype level. Meanwhile, a substantial reduction of the exhaust noise is regularly achieved in a broad frequency range covering all relevant engine orders. Results from roller test benches focus on the acoustic performance but as well demonstrate a marked improvement in backpressure which allows better engine efficiency and fuel economy. Further progress was made in the development of the durability and industrialization of all relevant components of the system. But instead of silencing also the shaping of different types of sound was implemented in some sports cars: Here, a variety of sounds are available from one and the same engine. Unlike with conventional passive exhaust systems the active silencer can thus be specifically adapted to the vehicle, to the driving situation, and on request even to the driver's preferences. Furthermore, in this paper current design trends and possible fields of application will be discussed.

Jenderka, K.-V.

• Metrology of HITU fields
  K.-V. Jenderka, K. Beissner and J. Haller [C]
[A]

High Intensity Therapeutic Ultrasound (HITU) is a non-invasive technique for the treatment of various types of cancer, as well as non-malignant pathologies by inducing localized necrosis of the tissue. Even though HITU is already in clinical use, this technique is still an emerging therapeutic procedure, with a lack of knowledge on reliable and stable treatment planning and monitoring procedures. A crucial step for the prediction of the desired temperature rise in a HITU-treated region is the measurement of the ultrasound output power. The evaluation of tissue thermometry methods, using MR or ultrasound itself, is important for the treatment monitoring, as well as for choosing the appropriate ultrasound power for a desired local intensity and temperature rise respectively. While ultrasound fields are well-investigated for output powers smaller than 50 W, this paper describes first results of measurements in much stronger, highly focused fields ( 300 W ultrasound power at 1.5 MHz) using an absorbing target arranged close to the transducer. The measurement uncertainties are discussed.

• Space-resolved sonochemistry in cleaning vessels in comparison to mechanical effects
  C. Kling, C. Koch and K.-V. Jenderka [I]
[A]

There are many applications of ultrasound in medical science and engineering which are based on the physical effect of cavitation in liquid media such as cleaning and sonochemistry. Unfortunately, cavitation is a stochastic process and is strongly influenced by environmental conditions. This makes it difficult to control or even optimise a certain application process. Up to now, empirical assessment methods are used to account for the influence of parameters like temperature, frequency, vessel geometry, surfactants, input power and others because it lacks of an objective description of the different cavitation processes. Therefore, a project at PTB aims to relate different cavitation effects to each other and to measurement parameters obtained from the underlying sound field.

In this context, sonochemical reaction in a cleaning vessel was studied. Simple oxidation of iodide ions was used to investigate the influence of different parameters. The reaction was restricted to small closed volumes inside the filled vessel to obtain spatial resolution of the cavitation action. The chemical efficiency of ultrasonic irradiation was analysed spectrometrically. The spatially resolved data were compared to quantitative results from mechanical cavitation effects.

Jensen, S.H.

• Complex Wavelet Based Modulation Analysis of Speech
  J.-M. Luneau, J. Lebrun and S.H. Jensen [C]
[A]

Low-frequency modulations of sound carry essential syllabic information for speech. The complex modulation spectrum is commonly obtained by spectral analysis of the sole temporal envelopes of the sub-bands out of a time/frequency analysis. Amplitudes and tones of speech or music tend to vary slowly over time. Thus the temporal envelopes are mostly of polynomial type. Filtering in this domain usually creates undesirable distortions because only the magnitudes of the temporal envelopes are processed and the phase data is often neglected. We remedy this problem with the use of a complex wavelet transform as a more appropriate envelope and phase processing tool. Complex wavelets carry both magnitude and phase explicitly with great sparsity and preserve well polynomials. Furthermore an analytic Hilbert-like transform is possible with complex wavelets implemented as an orthogonal filter bank. As a result, the transform delivers replicas of the time-frequency analysis at different modulation scales, namely ''Modulation Sub-Spaces''. By working in this alternative domain the transform shows very promising denoising capabilities and suggests new approaches for joint spectrotemporal analytic processing of slow frequency and phase varying signals.

Jeon, J.Y.

• Investigation of sound diffusion characteristics using scale models in concert halls
  J.Y. Jeon and Y.H. Kim [I]
[A]

The diffuse characteristics of the surfaces in lateral walls were investigated to find out the appropriate diffuser profiles in concert halls. The design factors of diffusers in auditoria are geometrical shape, structural height and surface coverage. The hemisphere diffusers were evaluated by measuring the diffusion and scattering coefficients in a 1:10 scale model reverberation chamber, and then applied to scale model concert halls, where the acoustical parameters such as RT, EDT, C80, and SPL were calculated from the impulse responses. In addition, the numbers and amplitudes of the reflection rays within -20 dB after the direct sound were calculated to evaluate the diffuseness in the halls. It was found that the diffusive lateral surfaces close to the stage area are more effective in increasing both RN and RE.

Jessen, M.

• Speaker Verification Based on Formants Using Gaussian Mixture Models
  T. Becker, M. Jessen and C. Grigoras [C]
[A]

In a new method, the well known UBM-GMM framework for automatic speaker verification is applied to formant parameters replacing the common cepstral parameters (e.g. Mel frequency cepstral coefficients (MFCC)). By using full covariance matrices, both the static and variable vocal tract configurations are reflected in the speakers' models. Speaker comparisons are achieved by relating speaker similarity to speaker typicality, expressed as the likelihood ratio. The method reveals a low equal error rate while providing easy interpretability and low dimensionality.

Jeub, M.

• Performance Analysis of Wavelet-based Voice Activity Detection
  M. Jeub, D. Kolossa, R. Fernandez Astudillo and R. Orglmeister [I]
[A]

The objective of this paper is to analyse the performance of wavelet-based voice activity detection algorithms (VAD) and to contrast it with that of the VAD standardized for the AMR-WB (Adaptive Multi-Rate Wideband) codec. Experimental results show that wavelet approaches lead to good results with respect to speech clipping and offer a much lower computational complexity. Integration of these algorithms into a Hidden Markov model (HMM) speech recognizer shows that the recognition performance using the AMR VAD can also be obtained or improved upon by wavelet based approaches, again at a notably reduced computational effort.

Joublin, F.

• Combining Auditory Inspirations and Hierarchical Feature Extraction for Robust Speech Recognition
  M. Heckmann, X. Domont, F. Joublin and C. Goerick [C]
[A]

Jürgens, T.

• Consonant Recognition of Listeners with Hearing Impairment and Comparison to Predictions Using an Auditory Model
  T. Jürgens, T. Brand and B. Kollmeier [C]
[A]

Contrary to listeners with normal hearing, persons with high-frequency sensorineural hearing loss suffer especially from a poor consonant recognition. This is mainly attributed to the loss of audibility within high-frequency regions. But still it is not clear how suprathreshold factors like the loss of dynamic range and thus the modification of dynamic compression in the impaired auditory system contribute to the poor consonant recognition. In a former study we showed that consonant recognition for normal-hearing listeners can be predicted well using an auditory model combined to a speech recognizer. In this study we address the following question by modifying the auditory model to model the signal processing of listeners with hearing-impairment: Does a detailed modelling of hearing-impairment by including suprathreshold factors like a modified dynamic compression enhance the prediction of consonant recognition compared to a model using audibility only? Therefore, observed consonant recognition rates are compared to model predictions using different model approaches to involve hearing-impairment.

Juhl, P.

• Calculation of Head Related Transfer Functions of bats using the Boundary Element Method
  P. Juhl, V. Cutanda Henriquez and D. Vanderelst [I]
[A]

ChiRoPing (Chiroptera, Robots, and Sonar) is an EU-funded research project aimed at understanding how bats use their echolocation perception ability and apply this knowledge to the design of new robotic senses. Four species of bats are selected for the study and models of their heads including minute details of ears, mouth and nose are obtained through CT scans. The project involves, among other things, the use of numerical methods on such scanned models to study the role of their features in the bat sensorial performance. As the bats operate at very high frequencies and as their ears and noses are very complicated and involve thin surfaces, the numerical modelling is challenging. The paper reports preliminary results of the numerical study and investigates the possibilities of reducing the computational complexity of the models.

• Numerical design and testing of a sound source for secondary calibration of microphones using the Boundary Element Method
  V. Cutanda Henriquez, P. Juhl and S. Barrera Figueroa [I]
[A]

Secondary calibration of microphones in free field is performed by placing the microphone under calibration in an anechoic chamber with a sound source, and exposing it to a controlled sound field. A calibrated microphone is also measured as a reference. While the two measurements are usually made consecutively, a variation of this procedure, where the microphones are measured simultaneously, is considered more advantageous from the metrological point of view. However, it must be guaranteed that the two microphones receive the same excitation from the source, although their positions are some distance apart to avoid acoustic interaction.

As a part of the project Euromet-792, aiming to investigate and improve methods for secondary free-field calibration of microphones, a sound source suitable for simultaneous secondary free-field calibration has been designed using the Boundary Element Method. The source has a central plug with the effect of reducing the spatial variations of the sound pressure at the microphone positions within the frequency range of the calibration. After some preliminary measurements, the source has been tested thoroughly in a calibration setup at the acoustic laboratory of the Danish Fundamental Metrology Institute (DFM). The design and verification of the source are presented in this communication.

Jung, C.

• About the Threshold of the Transient Ultrasonic Cavitation in the Cavitational Noise at Different Frequencies
  C. Jung and R. Sobotta [I]
[A]

Kabel, M.

• Optimization of Multilayered Porous Acoustic Absorbers
  H. Andrä, M. Kabel, M. Spies and H. Rieder [I]
[A]

A new method for the optimal material and layer design of acoustic trims without the usage of prototypes is presented. As an example the acoustic behaviour of a two-layered ceiling panel is optimized without increasing its weight or its thickness.

The optimization method consists of two steps: (1) The possible varieties of porous materials are analysed by computer simulation on the microscale, i.e. a certain set of boundary value problems is solved on the fully resolved microstructure. In a subsequent homogenization procedure a characteristic set of effective acoustic material parameters is computed together with their dependency on microstructural parameters like porosity. (2) A data base of these pre-computed characteristic material parameters is used in the actual optimization algorithm, which determines both the optimal thickness of the layers and their optimal microstructure with respect to a desired acoustic absorption behaviour.

A wide variety of materials can be taken into consideration: fibrous absorbers, open and closed cell foams, viscoelastic solids, porous or viscoelastic composites.

Since the main difficulty in this procedure is the computation of the characteristic acoustic parameters, we present the verification of the numerical scheme used for the computation of these parameters for the case of cylindrical pores.

Kalivoda, M.

• Assessment of track related noise mitigation measures
  M. Kalivoda [I]
[A]

Within the last decade methodologies for the measurement of railway noise generation have been strongly improved. This knowledge is used for type testing of railbound vehicles. However, a similar standard such as ISO 3095 is not available for track components, yet. There is a strong need for standardised assessment methods for track related noise abatement. This paper will present results from a recent measurement campaign in Austria dealing with changed track components. The discussion shows, that we have to control all the factors that might influence the results very carefully. Often we talk about effects of 1 - 2 dB(A) from the changed component while adverse influences can be of the order of 5 - 10 dB(A). Therefore it is necessary to collect enough data and thus get a sound statistical validation. It has been seen that only an unmanned automatic measuring system is able to collect sufficient acoustic and Meta information at recent costs.

Kallinger, M.

• Estimation of the Optimum System Delay for Speech Dereverberation by Inverse Filtering
  S. Goetze, M. Kallinger, A. Mertins and K.-D. Kammeyer [C]
[A]

Equalization of room impulse responses is an straightforward approach for dereverberation of speech signals in a hands-free scenario. The well-known least-squares equalization filter minimizes the euclidian distance between the concatenated system of room impulse response (RIR) and equalizer and a given target system. This target system usually is chosen as a delayed discrete pulse, a delayed band-pass or a delayed high-pass filter. In this contribution we address the choice of the this delay which has to be introduced w.r.t. maximum dereverberation performance. Since designing one equalizer (EQ) for each possible delay and choosing the best one is computationally inefficient we investigate the dependence of the optimum EQ delay on various parameters describing RIRs. A high correlation was found between the so-called central time of the room impulse response and the optimum EQ delay. Since the central time can be determined based on estimates of the initial peak of the RIR and the room reverberation time, we propose to use a very short filter for system identification and an estimate of the room reverberation time to identify the optimum equalizer delay.

Kaltenbacher, M.

• Limits of Finite Element Methods in the Mid-Frequency Range
  A. Hüppe and M. Kaltenbacher [I]
[A]

Field simulation methods like the Finite Element Method (FEM) in computational acoustics offer some advantages in comparison to energy based methods. For example, they can be applied in the frequency as well as in the time domain and the solution for acoustic pressure and/or velocity is available at every point inside the simulation domain. On the other hand, especially when considering higher frequencies, a FEM simulation requires a huge amount of computational time and memory, thus limiting its range of application to smaller geometries or lower frequencies.

In our contribution we focus on the simulation of the acoustic field in the time domain as the solution provides the whole frequency range of interest in one simulation. Therefore, we consider h-FEM and p-FEM applied to the wave equation and the Spectral Finite Element Method (s-FEM) applied to the governing equations of acoustics. We perform simulations in the mid-frequency range to show the limits of the methods in terms of accuracy of the solution and computational costs. As a practical application, the methods are used to simulate the sound field in the environment of a noise barrier.

• Computational Acoustics based on Conservation Equations using Taylor Hood Finite Elements
  M. Kaltenbacher and A. Hüppe [I]
[A]

Within this contribution we concentrate on the efficient solution of enhanced wave problems, which are mathematically modeled by their conservation equations, namely the mass and momentum conservation. Such formulations arise in computational aeroacoustics or in nonlinear acoustics. Even for acoustic wave propagations, where an acoustic wave equation exists, formulations based on the conservation equations have the advantage, that the numerical solution provides directly the two main engineering quantities, the acoustic particle velocity and the acoustic pressure.

The application of the standard Finite Element (FE) method to the acoustic conservation equations will result in an instable formulation. In order to obtain stable finite elements, we need a mixed FE formulation. The way to obtain such a formulation is to use H(div) functions for the approximation of the particle velocity and L2 functions for the acoustic pressure. However, it can be proofed, that approximating the acoustic particle velocity by Lagrangian finite elements one order higher than those for the acoustic pressure will also provide us a stable formulation.

We have applied this so-called Taylor-Hood elements to the acoustic conservation equations and will present their applicability to general wave problems. In detail, we will present a study of the accuracy and computational efficiency.

• Aeroacoustics taking Fluid-Structure Interaction into Account
  M. Kaltenbacher, S. Becker, J. Grabinger and F. Schäfer [I]
[A]

In the last years, manufacturers have started to consider the aerodynamic noise level in many industrial applications as a relevant design parameter (e.g. airplanes, cars, air conditioning systems, etc.). Because of the growing demand for reducing noise levels and for fulfilling noise regulations, there is a great motivation to investigate the basic aeroacoustic phenomena and mechanisms of sound generation and propagation. In many cases, the occurring noise is not only generated by the turbulent flow but also due to vibrations resulting from the interaction between the flow and the mechanical structure.

In the present work, we investigate the three field coupling for a flow past a thin flexible structure. Our developed numerical algorithms are capable to simulate the complex interaction between fluid, structure and acoustics. The scheme is based on a partitioned approach employing a finite-volume flow solver and a finite-element structural mechanics and acoustics solver. An important feature of the computational scheme is that it allows for a separate prediction of flow-induced and vibration sound. We will present detailed numerical studies of the interaction of the fluid flow with the flexible structure and the generated noise. Moreover, a comparison of computational results to our experimental investigations will be provided.

• Numerical study of the human phonation process by the Finite Element Method
  S. Zörner, M. Kaltenbacher, G. Link, R. Lerch and M. Döllinger [I]
[A]

The basis of the human phonation process is given by the complex interaction of air and flow in the larynx together with the structural mechanics of the vocal folds. Our goal is to study the influence of the geometric shapes of vocal folds on the quality of the generated sound. Since direct measurements on such variations are not feasible, we developed an enhanced numerical scheme based on 2d Finite Element Methods, capable to compute this coupled field problem. The geometric variations on the vocal folds change the main frequency component and, furthermore, influence the occurrence of the Coanda effect (the jet is attaching randomly to either side of the trachea). With our simulations, we can demonstrate, that this effect causes a broadband acoustic signal. Currently, we perform numerical simulations with asymmetric vocal folds to investigate different types of dysphonia.

• Human phonation analysis by 3d aero-acoustic computation
  S. Zörner, M. Kaltenbacher, W. Mattheus and C. Brücker [I]
[A]

A numerical model of the phonation process, which takes place inside the larynx, is presented. The air flow through the glottis has been replaced by water and the physical dimensions have been scaled accordingly. Therewith, the characteristic frequencies and velocities are reduced. The movement of the vocal folds are prescribed by geometric changes of the computational domain.

In utilising numerical simulations, the goal is to study the complex interaction of the modulated flow, due to the prescribed geometric variations. The fluid flow is simulated by solving the incompressible Navier- Stokes equation with the help of a direct numerical simulation (DNS) based on the software OpenFOAM. The results are scaled to comply with a simulation in air. This data is then taken to calculate the acoustic sources inside the model by using Lighthill's analogy and the acoustic sound propagation by solving the inhomogeneous wave equation. Currently we prescribe a periodical flow on the inflow boundary and compute the flow field for fixed vocal fold geometries. The computed acoustic field inside the glottis shows the same dominant frequency as the prescribed fluid inflow.

• Optimizing Material Properties and Geometry of a Physical Multi-Layered Vocal Fold Model
  B. Schmidt, G. Leugering, M. Döllinger, R. Lerch, M. Kaltenbacher and M. Stingl [I]
[A]

Understanding the physical fundamentals of human voice generation is a challenging and interdisciplinary task. The present work takes a closer look at a self-sustained physical model of the human larynx, in which the vocal folds are modeled by multi-layered elastic bodies, i.e. silicone layers. We search for optimal material and geometry parameters in the sense that the deformation of the multi-layer model at its surface is as similar to given deformation patterns arising from hemilarynx experiments as possible. For this purpose, we employ an optimization concept based on a 2D finite element simulation in which the vocal fold is modeled as linear elastic body. Keeping the layer geometry fixed, we use the material optimization approach in order to identify optimal material parameters for each layer by assuming isotropic or transversal-isotropic material properties. Additionally, we employ techniques from shape optimization, in order to optimize geometry of the boundaries between the different layers parameterized as spline curves, while keeping the material tensors fixed. We consider the simultaneous solution of both problems, discussing optimization of material and geometry concurrently. We conclude this presentation by giving some numerical results using a 2D vocal fold model.

Kammeyer, K.-D.

• Estimation of the Optimum System Delay for Speech Dereverberation by Inverse Filtering
  S. Goetze, M. Kallinger, A. Mertins and K.-D. Kammeyer [C]
[A]

Equalization of room impulse responses is an straightforward approach for dereverberation of speech signals in a hands-free scenario. The well-known least-squares equalization filter minimizes the euclidian distance between the concatenated system of room impulse response (RIR) and equalizer and a given target system. This target system usually is chosen as a delayed discrete pulse, a delayed band-pass or a delayed high-pass filter. In this contribution we address the choice of the this delay which has to be introduced w.r.t. maximum dereverberation performance. Since designing one equalizer (EQ) for each possible delay and choosing the best one is computationally inefficient we investigate the dependence of the optimum EQ delay on various parameters describing RIRs. A high correlation was found between the so-called central time of the room impulse response and the optimum EQ delay. Since the central time can be determined based on estimates of the initial peak of the RIR and the room reverberation time, we propose to use a very short filter for system identification and an estimate of the room reverberation time to identify the optimum equalizer delay.

Karlowatz, G.

• Modelling Sound Level and Pressure Loss in HVAC Duct Networks using a Framework in Modelica/Dymola
  H. Kühnelt, A. Haumer, T. Bäuml, U. Reisenbichler, C. Reichl and G. Karlowatz [C]
[A]

In the early design phase of large HVAC duct networks the joint calculation of the sound level and air flow rate and pressure loss as well as the prediction of the noise level in the passenger compartment gain more and more in importance. A unified framework for the one-dimensional flow acoustic simulation using Dymola/Modelica is introduced. The acoustic part is described by plane wave two-port theory, the fluid mechanic part by pressure loss models. Various approaches including analytic and (semi)empirical models based on experimental and simulation data can be implemented. Each component either represents the acoustic or the flow part of a duct element or can, itself, be composed of both types allowing the modelling of non-standard components. The object oriented approach provided by Modelica assures the extensibility of the models e.g. by transport and exchange of heat or the (de)humidification of the air. It also allows the integration into a complete system simulation of the HVAC network including the cooling circuit and the electronic control loop. The simulation is easily set up using a GUI; an external parameterization ensures persistent data management and a smooth workflow. The resulting system of equations is automatically pre-processed and solved by Dymola.

Kauba, M.

• Active control of the sound transmission through a double-glazing window
  J. Bös, S. Herold, O. Heuss, M. Kauba and D. Mayer [C]
[A]

Within the framework of the European Integrated Project InMAR (Intelligent Materials for Active noise Reduction) a system for the active control of the sound transmission through a double-glazing window was developed. The window has a size of 1.2 m x 0.6 m and consists of two glass panes (6 mm and 4 mm thick) with a gas-filled gap of 20 mm in between. The vibrational behavior of the window panes was analyzed by means of finite element simulations and experimental modal analyses using a scanning laser vibrometer. The results of these analyses were used to determine suitable locations for the application of piezoelectric patch transducers. If these patches are used as sensors they can serve as error sensors or reference sensors for an active control system, if they are used as actuators they can suppress the vibrations of the window panes and, therefore, reduce the sound transmission through the window. Several transducer configurations and control strategies were tested and compared with each other. The sound radiation of a broad-band shaker excitation was reduced by up to 20 dB at various frequencies, and the sound transmission through the window was reduced by up to 12 dB at various frequencies.

Kauffman, J.

• Scalable Microphone Array System Optimized for Streaming Inverse Acoustics
  J. Kauffman [I]
[A]

Kaynar, S.

• Simulation of differences between male and female vocal fold configuration during phonation
  M. Kob, P. Dejonckere, E. Calderón and S. Kaynar [I]
[A]

Statistically, some vocal fold lesions such as vocal fold nodules - considered as a localized tissue reaction to voice loading - are much more frequent in female ( 95%) than in male ( 5%) adult humans. This can only partially be explained by the about two times higher oscillation frequency during voice production in females. In previous studies, gender-related differences in the shape of the vocal fold edge have been observed that can be related to different geometries of the laryngeal framework. With a time-domain model of vocal fold function the oscillation patterns and resulting forces between and inside the vocal folds have been simulated, and the conditions for eliciting a hourglass-shaped vibration pattern (necessary for determining the localized tissue reaction of the nodules) could be clearly defined. Our simulations and measurements support the assumption that collision and tearing forces on the vocal fold edges during phonation are higher and more specifically located in females compared to males.

Kellermann, W.

• Robust Superdirectional Beamforming for Hands-Free Speech Capture in Cars
  E. Mabande, A. Schad and W. Kellermann [I]
[A]

The application of multichannel signal processing algorithms in the front-ends of hands-free systems facilitates a better extraction of a desired source signal and allows for a suppression of unwanted noise and interference. In this paper, a novel superdirective beamforming design method called the Robust Least-Squares Frequency-Invariant Beamformer (RLSFIB) is investigated for microphone arrays in cars. The RLSFIB design offers significant spatial selectivity at low frequencies even for small array apertures and some variability in microphone positioning. To cope with the high sensitivity inherent in superdirectional beamformers, this method directly incorporates a white noise gain (WNG) constraint for controlling the beamformer sensitivity. The resulting problem is formulated as a convex optimization problem which can be solved by conventional iterative solvers and the convexity guarantees a globally optimal solution. As the WNG constraint can be chosen freely, the method can match desired robustness levels, and therefore can be adapted to given prior knowledge on microphone mismatch, positioning errors, and microphone self-noise. Design examples for typical car requirements will be presented.

Kemp, J.

• Reconstructing the Lituus: A Reassessment of Impedance, Harmonicity, and Playability
  A. Apostoli, S. Logie, A. Myers, J. Kemp, J. Chick and A. Braden [C]
[A]

Bach's motet `O Jesu Christ, meins Lebens Licht' (BWV no. 118), first performed in 1736-7, calls for a Lituus, believed to be a long trumpet-like instrument, with playing characteristics somewhere between a trumpet and an alphorn. Unfortunately, there are no surviving examples of Bach's instrument. Using software recently developed at the University of Edinburgh for brass instrument optimisation, a new bore profile with theoretically calculated input impedance curves has been proposed and submitted to a collaborating instrument maker with the aim of providing a natural instrument, with a distinct timbre, that responds well to the player and is in tune over a wide range of natural notes.

This paper describes the process, computational modelling, and experimental work undertaken in an attempt to ensure a satisfactory final instrument design. An initial validation of this process was achieved, in part, by investigating the playing properties of an instrument with a notionally similar bore profile, a 2.4m long Fiscorn in C (a type of Catalonian flugelhorn). This work compared the measured bore profile, input impedance, playability and intonation with those derived theoretically.

Kerber, S.

• Localisation dominance for long lead-lag stimuli in background noise
  S. Kerber and B. Seeber [C]
[A]

• Algorithmic modelling of the disturbance impact of background sounds
  S. Schlittmeier, T. Weissgerber, S. Kerber, H. Fastl and J. Hellbrück [I]
[A]

Mentally demanding tasks very rarely take place during silence. However background sound can reduce cognitive performance, even if it is irrelevant to the task and is intended to be ignored. This so-called Irrelevant Sound Effect (ISE) has been verified in a multitude of behavioral experiments for verbal short-term memory performance.

Although a multitude of cognitive psychological experiments have explored the ISE, no psychoacoustically based instrumental procedure existed to predict its occurrence. This poses problems in an applied context: In office environments, for example, the potential beneficial - or less advantageous - effects of noise abatement on cognitive performance could not be calculated before their realization.

The talk presents an algorithm which models performance data in ISE experiments on the basis of instrumental measurements of the hearing sensation fluctuation strength. It was verified with a database consisting of about 50 background sounds and corresponding performance data that have been collected in cognitive psychological experiments at the KU Eichstätt-Ingolstadt. The algorithm is able to reproduce the performance results in about 90 % of cases within the interquartile ranges. It will be discussed within the scope of cognitive short-term memory models, which claim to explain the ISE and with respect to practical implications.

Kettler, F.

• New developments in VoIP testing
  N. Rohrer, S. Poschen and F. Kettler [I]
[A]

The market for IP based speech transmission increases rapidly. Integrated access devices (IAD) are offered by network providers for end users. VoIP gateways provide the interface between NGN (Next Generation Networks) and traditional PSTN (Public Switched Telephone Networks). Speech quality measurements are also further developed in order to cover more realistic test scenarios and more efficient test conduction. This contribution discusses new tests and test results on different devices. Information are provided about ongoing international testing activities e.g. the ETSI Speech Quality Test Events (SQTE).

• Evaluation of aurally-adequate analyses for echo assessment
  F. Kettler, M. Lepage and M. Pawig [I]
[A]

The migration towards NGN (Next Generation Networks) is supposed to introduce higher propagation delay in telecommunication. This emphasizes the need for reliable echo control. Wideband telephony will further change speech perception - and echo perception. New investigations on echo perception demonstrated the necessity to renew tolerances for wideband echo attenuation. These trends also motivate new echo assessment methods. A new approach based a hearing model analysis is introduced and discussed in this presentation suitable to extend existing echo analysis methods.

Kim, Y.H.

• Investigation of sound diffusion characteristics using scale models in concert halls
  J.Y. Jeon and Y.H. Kim [I]
[A]

The diffuse characteristics of the surfaces in lateral walls were investigated to find out the appropriate diffuser profiles in concert halls. The design factors of diffusers in auditoria are geometrical shape, structural height and surface coverage. The hemisphere diffusers were evaluated by measuring the diffusion and scattering coefficients in a 1:10 scale model reverberation chamber, and then applied to scale model concert halls, where the acoustical parameters such as RT, EDT, C80, and SPL were calculated from the impulse responses. In addition, the numbers and amplitudes of the reflection rays within -20 dB after the direct sound were calculated to evaluate the diffuseness in the halls. It was found that the diffusive lateral surfaces close to the stage area are more effective in increasing both RN and RE.

Kimura, M.

• Estimating physical properties of vocal fold paralysis from high-speed filimg data
  I. Tokuda, M. Kimura, H. Imagawa, K.-I. Sakakibara and N. Tayama [C]
[A]

Pathological voice originates from complex nonlinear vibrations of the disordered vocal folds. Deep understanding of the voice disorders based on mathematical modeling of the vocal folds may provide an important insight into the abnormal vocal fold vibration as well as a good hint for improving the voice surgery. Up to date, various models have been developed to simulate complex vibratory patterns of the vocal folds, ranging from a simple two-mass model to complex multi-mass models. However, only a few researches aim at construction of the vocal fold model, which realizes quantitative characteristics of the individual voice. Recently, Doellinger et al. proposed a method for estimating the parameters of asymmetric two-mass model for high-speed filming data recorded from human subjects. The present paper applies this technique to pathological data recorded from patients with vocal fold paralysis. Right and left vocal fold tensions, subglottal pressure and glottal area have been estimated. Whether the effect of the voice surgery is reflected in the estimated parameters is judged by using the data before and after the voice surgery. On the basis of the estimation results, applicability of the mathematical model as a simulator for the voice surgery is discussed.

Kindt, P.

• A Ring-based Tyre Model for the Prediction of Structure-borne Interior Tyre/Road Noise
  P. Kindt, P. Sas and W. Desmet [I]
[A]

The tyre/road interaction causes a structural excitation of the tyre. This vibrational energy is transmitted through the suspension towards the vehicle body. The resulting body panel and window vibrations cause noise radiation in the vehicle passenger compartment. In order to predict this interior noise during the vehicle development phase, an accurate tyre model is required. This paper presents a structural tyre model, based on a three-dimensional flexible ring on an elastic foundation. The ring represents the belt and the elastic foundation represents the tyre sidewall. The model is valid till 300 Hz and includes a submodel of the wheel and the air cavity. Unlike most ring models, which only consider in-plane modes, the presented model also predicts modes that involve torsion of the belt in circumferential direction. The parameterization of the model, which does not require detailed knowledge of the tyre construction, is based on the main geometrical properties of the tyre and a limited modal test. The unloaded tyre response is calculated and validated with test data. As the presented model is physical, it can be applied to describe other operational conditions such as loading and rotation. In this paper, the response of the loaded tyre is also considered.

Klein-Hennig, M.

• The Effect of Envelope Waveform on Lateralization
  M. Klein-Hennig, M. Dietz, S.D. Ewert and V. Hohmann [C]
[A]

Interaural timing disparities are important cues for lateralization in the human auditory system. The auditory system is sensitive to timing disparities in the fine-structure and the envelope of sounds. At high frequencies, only envelope disparities can be exploited due to the lack of phase-locking to the fine-structure. In previously published data [e.g. Bernstein and Trahiotis in ''Hearing - from basic research to applications'', Springer 2007] parameters like the modulation frequency, or the power of a modulating sinusoid have been altered in order to investigate the role of the envelope waveform on lateralization. However, these ''classical'' parameters simultaneously change several ''secondary'' envelope parameters like attack and decay times, as well as pause and hold durations, which might conceal the fundamental cues exploited by the auditory system. In this study, psychophysical measurements were conducted with customized envelope waveforms in order to investigate the effect of isolated secondary parameters on lateralization. For high-frequency tones centred at 4 kHz with systematic envelope modifications, the just noticeable differences of interaural envelope time differences were measured. The results indicate that attack times and pause durations prior to the attack are the most important envelope features. The results are compared to predictions by a binaural auditory model.

Kleinhenrich, C.

• A Three-Dimensional Acoustic Simulation for the Development and Evaluation of Active Noise Control Systems using the FDTD Method
  C. Kleinhenrich, A. Niepenberg, T. Weigler and D. Krahé [C]
[A]

The development of Active Noise Control Systems to reduce noise in certain two- or three-dimensional areas or spaces, respectively, requires good knowledge of the desired installation location as well as the understanding of the systems properties and behavior. In this paper simulation software is presented which allows the user to build virtual rooms or open air test sites and to place noise sources and active systems in an arbitrary manner to reduce the noise. A FDTD (Finite Differences Time Domain) algorithm was implemented to compute sound field behavior, including reflection, absorption, transmission and diffraction in different media. The active systems can consist of an arbitrary number of sensors (microphones) and actuators (loudspeakers) which are combined and connected via different signal paths and signal tool boxes. This allows the creation of various system setups to find the best solution and to predict the effects of noise cancellation for a given problem.

Kletschkowski, T.

• Broadband Active Noise Control Around Human's Head - Determination and Achievement of Physical Limits
  T. Kletschkowski and D. Sachau [C]
[A]

In order to prevent serious diseases caused by continuous noise pollution and to guarantee a restorative sleep the concept of active noise control was applied to bedrooms. To determine the physical limits of this concept in a particular situation a transmission test rig was used to simulate real world conditions. A conventional bed - placed in the reverberation room - with a dummy head microphone was used to model the sleeping person. Two error microphones - placed inside the pillow - were used for multi-channel feedforward control. The reference microphone was placed in front of a tilted window that was mounted into the transmission path. At first the physical limits of the control profit were determined by coherence-analysis. Then a proper filter length for two adaptive controller was calculated by quadratic optimization in time domain. This procedure was based on measured impulse response functions. Finally the noise reductions at the error microphones were determined for broadband disturbances 80Hzf480Hz. It turned out that the control profit of up to 18dB was equivalent to the noise reduction predicted by coherence-analysis.

• Noise Source Identification in a Cross-Section of a Long-Range Airliner by Means of the Inverse Finite Element Method
  M. Weber, T. Kletschkowski and D. Sachau [C]
[A]

An inverse finite element method (IFEM) for noise source identification in closed cavities is presented. The procedure requires sound pressure measurements in the cavity first. If all sound sources are located on the boundary, the equation system resulting from a matching finite element (FE) model can be resorted in such a way that computation of the unknown boundary pressure and particle velocity is possible by solving the emerging inverse problem. Regularization techniques must be applied to account for the ill-posedness of the problem.

The method was first validated using a simplified three-dimensional FE model. The numerically calculated data inside an inner sub-domain were impinged with a stochastic error and used as simulated measurement data to re-calculate the boundary data.

In a next step, the IFEM was applied to an aircraft mock-up, consisting of a cross-section of a long-range airliner. The sound field attuning in the cavity with an interior noise source was mapped with a custom-built microphone array. A matching FE model was developed, and the sound pressure and particle velocity prevailing on the boundary were calculated from the measured data.

• Acoustic Ground Tests in a Cross-Section of a Long-Range Airliner for Validation of the Inverse Finite Element Method
  K. Simanowski, T. Kletschkowski, D. Sachau and B. Samtleben [C]
[A]

Identification of noise sources in aircraft cabins proves to be difficult particularly at low frequencies. A new approach, based on the Inverse Finite Element Method (IFEM), reconstructs the spatial distribution of sound pressure and particle velocity at the interior wall. This procedure requires measurements in the cavity first. If all sound sources are located on the boundary, the equation system resulting from a matching FE model can be resorted in such a way that computation of the unknown boundary impedance is possible. The paper presents the results of acoustic ground tests that were performed to prepare a validation of this noise source localization technique. An aircraft mock-up (cross section of a long-range airliner) was excited by both interior and exterior noise sources for tonal as well as broadband disturbances. In both cases the resulting SPL and the frequency response were mapped with a custom-built microphone array at 7000 measurement points. The measured data was used to analyze the operational modes and the acoustic properties (such as reverberation time) of the cavity. Furthermore, the impedance at selected boundaries was calculated. The experimental results were then compared with numerical results of forward (FEM) and backward (IFEM) calculations.

Kling, C.

• Space-resolved sonochemistry in cleaning vessels in comparison to mechanical effects
  C. Kling, C. Koch and K.-V. Jenderka [I]
[A]

There are many applications of ultrasound in medical science and engineering which are based on the physical effect of cavitation in liquid media such as cleaning and sonochemistry. Unfortunately, cavitation is a stochastic process and is strongly influenced by environmental conditions. This makes it difficult to control or even optimise a certain application process. Up to now, empirical assessment methods are used to account for the influence of parameters like temperature, frequency, vessel geometry, surfactants, input power and others because it lacks of an objective description of the different cavitation processes. Therefore, a project at PTB aims to relate different cavitation effects to each other and to measurement parameters obtained from the underlying sound field.

In this context, sonochemical reaction in a cleaning vessel was studied. Simple oxidation of iodide ions was used to investigate the influence of different parameters. The reaction was restricted to small closed volumes inside the filled vessel to obtain spatial resolution of the cavitation action. The chemical efficiency of ultrasonic irradiation was analysed spectrometrically. The spatially resolved data were compared to quantitative results from mechanical cavitation effects.

Klosak, A.K.

• Relationship between room shape and Early Lateral Energy Fraction in rectangular concert halls
  A.K. Klosak and A.C. Gade [C]
[A]

The acoustic properties of 24 rectangular rooms representing ''shoe-box'' type concert halls with volumes between 8000m3 and 16000m3 have been analysed using Odeon (v.8) computer simulation software. Among several objective measures calculated in numerous receiver positions (between 300 and 850) in each hall, this paper concentrates on the behaviour of Early Lateral Energy Fraction (LF80). The dependency of LF80 on source-receiver distance, receiver location within the room and room dimensions will be discussed. It is found that the distribution of LF80 in rectangular rooms can be described as a three-dimensional ''tongue-shape''. The simulation results have also been compared with regression models based on measurements in real halls. In rectangular halls, the variation in LF80 values can be described by the influence of the overall room dimensions and their mutual ratios. Formulas for prediction of LF80 in rectangular halls have been proposed, which takes into account both the width and the length of the hall.

Knauss, D.

• Noise Management at Civil Shooting Ranges
  D. Knauss and K.-W. Hirsch [C]
[A]

DIN ISO 17201-2 proposes a scheme to set up a daily noise management on civil shooting ranges. The scheme relies on rating levels based either on calculations or measurements of the LAFmax at relevant receptor points. In order to employ such a management system, detailed information have to be known. For civil shooting ranges however, it is sometimes difficult to gather all the details in advance and to put everything together in an appropriate computer based management system.

This paper reports on a system that follows the basic ideas of the DIN ISO 17201-5 but relies on the measurement of each single shot at given positions at close distance during normal operation. Using these single shot levels and the met. data, the system applies a transfer matrix to predict the contribution of each shot to the rating level at each relevant receiver point. The coefficients of the matrix are determined during a preceding measurement campaign and accounts for the wind speed and direction.

Kniesburges, S.

• Sound Generation in a Flow-Induced Vibrating Human Vocal Folds Model
  S. Becker, S. Kniesburges, S. Müller and A. Delgado [I]
[A]

For investigating the fluid-structure-acoustic coupled (FSA) process of human phonation, synthetic models of the vocal folds (VF) were designed which showed flow-induced oscillations and produced a tonal sound. Two different kinds of models were developed: one for observing the flow field vibrating at constant frequency and a more realistic which could be stretched in longitudinal direction to simulate the prestress within human VFs. Besides the investigation of the flow field using PIV the major aim was to observe the FSA process to determine the acoustic sources. Therefore two approaches were applied: Firstly the vibration of the VFs was recorded using a high-speed camera triggered by the subglottal pressure to determine phase shift between the two signals. Secondly synchronous measurements of the subglottal pressure, the flow velocity behind the VFs and the acoustic pressure were done to determine their coherence. The results support the existence of the Coanda-effect. Furthermore the vibration frequency and the phase shift between the subglottal pressure and the glottal width depended on the stiffness and the tension in the VFs. The synchronous measurements determined the pulsating flow rate as main acoustic source.

Knospe, H.

• Spectral Analysis of Audio Signals for the Identification of Spam Over IP Telephony
  C. Pörschmann and H. Knospe [C]
[A]

With modern computer and telecommunication systems voice calls can be automatically set up and prerecorded speech messages can then be played. Especially in IP-based networks the costs for voice calls are quite low, Spam over IP Telephony (SPIT) can become a serious problem in the near future. Most activities ongoing in order to identify SPIT and to protect telephone networks from being flooded with SPIT make use of black or white lists. In this talk, a method adapted from the area of music identification is proposed which can be used to identify SPIT and to automatically establish black-lists. From the audio speech data of voice calls an ''acoustic fingerprint'' is calculated. The ''acoustic fingerprints'' of the voice calls are compared to each other and those are identified which have a high degree of similarity. The method, which is resistant to various modifications of the audio signal, can be used to detect SPIT which is typically characterized by similar or identical voice data in a large number of calls. Privacy protection is assured since only a ''fingerprint'' but not the complete audio data is stored, which does not permit the reconstruction of the content or the identification of the speaker.

Knutzen, J.

• Uncertainty in Acoustic Boundary Characterisation and its Influence on the Sound Field in Room Acoustic FE Simulations
  M. Aretz and J. Knutzen [I]
[A]

The room acoustic design of the low frequency characteristics of small rooms still poses a challenging task to acoustic designers and consultants. In this field numerical methods such as the finite-element-method (FEM) are promising contenders for the realistic simulation of low frequency sound fields. While contemporary CAD and meshing tools facilitate the design of high quality geometric models with regular and sufficiently fine meshes, the Achilles' heel of room acoustic FEM simulations appears to be the realistic representation of the room boundaries. The presented study investigates the influence of different boundary representations on the simulated sound field in small rooms. Therefore measurements and FE simulations of the sound field in a scale model room with a well defined geometry and variable boundary conditions have been conducted. Two different approaches are followed to model the porous absorbers on the room boundaries. Firstly the absorber layers are modelled with their exact 3D dimensions using an 'equivalent homogeneous fluid model'. Secondly the acoustic characteristics of the absorber layers are modelled by an appropriate impedance boundary condition, which is determined by measurement or model calculation based on measured flow resistivities. The audibility of differences between simulation and measurement results is assessed by subjective testing.

Kob, M.

• An approach for stable calculation of vocal fold oscillation
  E. Loch, S. Noelle and M. Kob [I]
[A]

The modeling of continuous string-like systems is often performed by spatial and temporal discretization of a problem-dependent wave equation. For arbitrarily chosen discretizations instabilities may appear which in contrary to the expected increase of accuracy decrease the stability of the calculations. In this paper an approach to avoid such instabilities by adopting the temporal discretization to the spacial discretization is presented. Implementations of a simplified vocal fold model have been made in MATLAB and JAVA. The result is tested with a multi mass model of vocal fold oscillation. It is shown that for an arbitrary number of masses a stable calculation is obtained. The application of the method for more complex models of normal and pathologic voice production is planned.

• An approach for numerical calculation of glottal flow during glottal closure
  A. Gömmel, M. Kob, T. Niendorf and C. Butenweg [I]
[A]

Limitations of classical low-degree-of-freedom approaches (multiple-mass models) of modeling vocal fold (VF) oscillation have indicated the need for more advanced models. A couple of finite-element (FE) models have been presented in the last years, which include a number of numerically difficult solver algorithms. The presented model represents a coupled two-dimensional model of VF movement including fluid-structure interaction effects. A structural model of vocal fold tissues and a fluid model of the airflow are combined by the Arbitrary Lagrangian-Eulerian formulation. As solvers ANSYS was chosen for the structure and CFX for the fluid. A special problem is handling the stop of the airflow in the closed phase of the glottal cycle. Since the fluid elements in the glottal gap channel would have to be zero-volume elements, a special modeling strategy is necessary. In this model a loss coefficient is introduced when the glottal gap falls below a certain control volume. As exemplary area of use, results of comparative simulations are shown, which differ in driving pressure values and VF shape. It is shown that typical convergent/divergent shapes are only obtained when a minimum length of a channel between the VF remains.

• Biomechanical modeling of chest-falsetto registers and their transition
  I. Tokuda, M. Zemke, M. Kob and H. Herzel [I]
[A]

The exact definition of registers in the human voice is still under debate. Especially the quantitative analysis of transitions between the registers has not been investigated in much detail yet. Excised larynx experiments show different kinds of voice instabilities that appear close to the transition from chest to falsetto register. Similar phenomena are also observed in vocalization of untrained singers. Towards comprehensive understanding of such register transitions, this paper carries out a biomechanical modeling of the human voice. Our model has a body-cover structure, which is in total composed of fourth masses. A smooth geometry is realized by introducing a polygon shape to the vocal fold model. Acoustical coupling with the vocal tract as well as the trachea is also considered. The simulation study shows that the model can reproduce many complex phenomena such as register jumps, hysteresis, subharmonics, and chaos, observed in excised larynx experiment as well as in vocalization of untrained singers. We study in detail the influence of the subglottal and supraglottal resonances on the transition point of the registers.

• Investigation of voice production in death metal singers
  C. Eckers, D. Hütz, M. Kob, P. Murphy, D. Houben and B. Lehnert [I]
[A]

Singers in ''metal'' style often use a unique form of voice production called ''growling''. The underlying mechanisms of voice production as well as possible potentially harmful influences on the singer's voice have not yet been intensively studied. Supra glottal laryngeal constriction appears to play a major role in metal singing. This constriction results in difficulties to apply standard laryngeal endoscopy for visualization of vocal fold function during this particular phonation style. In this study we present an examination method using a questionnaire including the voice handicap index as well as a combined simultaneous endoscopy, electroglottography and acoustic analysis to investigate in particular the production of death metal singing in seven subjects. None of the subjects complained about persistent voice problems and none had signs of secondary organic voice disorders like vocal fold nodules or inflammation. However, most showed minor irregularities on stroboscopic examination. Further findings reveal that at least two forms of supra glottal laryngeal constriction are used to obtain the typical death metal sound which differ in use of vocal folds and ventricular folds as well as in supra-glottal adduction effort. The study shall help to answer the question whether death metal singing can be performed such that the risk of voice damage can be reduced or avoided.

• Simulation of differences between male and female vocal fold configuration during phonation
  M. Kob, P. Dejonckere, E. Calderón and S. Kaynar [I]
[A]

Statistically, some vocal fold lesions such as vocal fold nodules - considered as a localized tissue reaction to voice loading - are much more frequent in female ( 95%) than in male ( 5%) adult humans. This can only partially be explained by the about two times higher oscillation frequency during voice production in females. In previous studies, gender-related differences in the shape of the vocal fold edge have been observed that can be related to different geometries of the laryngeal framework. With a time-domain model of vocal fold function the oscillation patterns and resulting forces between and inside the vocal folds have been simulated, and the conditions for eliciting a hourglass-shaped vibration pattern (necessary for determining the localized tissue reaction of the nodules) could be clearly defined. Our simulations and measurements support the assumption that collision and tearing forces on the vocal fold edges during phonation are higher and more specifically located in females compared to males.

Koch, C.

• Space-resolved sonochemistry in cleaning vessels in comparison to mechanical effects
  C. Kling, C. Koch and K.-V. Jenderka [I]
[A]

There are many applications of ultrasound in medical science and engineering which are based on the physical effect of cavitation in liquid media such as cleaning and sonochemistry. Unfortunately, cavitation is a stochastic process and is strongly influenced by environmental conditions. This makes it difficult to control or even optimise a certain application process. Up to now, empirical assessment methods are used to account for the influence of parameters like temperature, frequency, vessel geometry, surfactants, input power and others because it lacks of an objective description of the different cavitation processes. Therefore, a project at PTB aims to relate different cavitation effects to each other and to measurement parameters obtained from the underlying sound field.

In this context, sonochemical reaction in a cleaning vessel was studied. Simple oxidation of iodide ions was used to investigate the influence of different parameters. The reaction was restricted to small closed volumes inside the filled vessel to obtain spatial resolution of the cavitation action. The chemical efficiency of ultrasonic irradiation was analysed spectrometrically. The spatially resolved data were compared to quantitative results from mechanical cavitation effects.

Koch, P.

• Properties of Bubbles at Pressure Antinodes in a Standing Sound Field
  T. Davaadorj, P. Koch, A. Pluta, U. Parlitz and W. Lauterborn [C]
[A]

The parameter space for trapping bubbles in an acoustic resonator is limited by various factors, notably dissolution of the bubble by diffusion of the gas inside the bubble into the surrounding liquid, the primary Bjerknes force driving strongly oscillating bubbles away from the pressure antinode and parametric instabilities leading to surface oscillations of large amplitude. The region of stability in the parameter plane made up of the bubble radius at rest and the driving sound pressure amplitude at a frequency of 1 MHz is determined for different static pressures. The temperatures at bubble collapse in this region are calculated. The maximum attainable temperatures appear along the upper stability boundary in bubble resonance conditions.

Kochan, K.

• Robust Parameter Design of an Adaptive Multiple-Channel Active Noise Controller
  K. Kochan and D. Sachau [C]
[A]

Noise induced by machines, vehicle engines or propellers can lead to high tonal sound pressure levels. Due to increasing comfort requirements and occupational health and safety regulations, noise protection is a mandatory component of each modern technical system. With advances in technology, this challenge can in some cases be solved by the active noise control method. As an application example, the active noise control method is applied to the acoustical problem of a loadmaster working area, which is in front of the cargo bay of a propeller driven transport aircraft. The present research is focused on the final design process of such an active noise control system. Here, the hardware and the positions of 8 secondary loudspeakers and 16 error microphones are already defined. But the controller software and parameters can still be adapted in order to improve the performance and robustness of the system. In the present paper the steepest descent frequency domain controller as well as different solutions of the robust controller parameterization will be discussed. Moreover, the active control results inside an acoustical mock-up of the loadmaster working area will be presented.

• ANC Concept study using a 1-D-Model
  N. Hoevelmann, D. Sachau and K. Kochan [C]
[A]

In an advanced - especially propeller driven - aircraft the sound pressure level may reach up to high values caused by propeller blades producing disturbances when passing the fuselage. This excitation induces high tonal noise levels in the low frequency range in the aircraft interior. Due to mass reduction as a major aspect in aircraft design research has been focused on active noise control (ANC). In order to test the performance of ANC-systems test facilities are needed for different technology readiness levels. To analyze the expected acoustic behaviour of future aircrafts under realistic conditions a test rig for acoustic ground test is needed. The present paper will outline the design process of a mechatronic mock-up including both airborne and structure borne noise sources as well as the associated transfer paths. It is planned to use data taken from numerical simulations to adjust a specific noise field inside the mock-up by a proper control scheme. To simulate relevant transfer paths it is planned to integrate modern light weight structures - such as CFK - into the test rig.

Kodejska, M.

• Ferroelectret-film accelerometers with high sensitivities
  M. Kodejska, J. Hillenbrand and G.M. Sessler [C]
[A]

Ferroelectrets are piezoelectric films of charged cellular or porous polymeric electrets and have larger d33-coefficients than lead zirconate titanate (PZT). Compared to PZT and other piezoelectric materials, ferroelectrets are flexible, lightweight, and cheap. These properties allow the construction of light, flat, and sensitive accelerometers of various shapes. Such accelerometers were built with polypropylene (PP) ferroelectrets in single- and in multi-layer arrangements and seismic masses from 1 to 20 g. Film area, number of ferroelectret films, and seismic mass determine sensitivity and resonance frequency of the accelerometer and these parameters were varied for the production of different accelerometers. The frequency responses of the sensitivity of the accelerometers were measured using an electrodynamic vibration exciter. A typical sensitivity of about 3 pC s2/m was obtained for an accelerometer consisting of a single ferroelectret film with d33 500 pC/N in the audio frequency range, a seismic mass of 9 g, and an area of 3 cm2. A theoretical description of the various accelerometers has been achieved using simple models. Experimental and theoretical data were compared and showed good agreement. The results demonstrate that ferroelectret-film accelerometers have advantageous properties, in particular when adapted to specific applications.

Köhler, J.

• A Study of Throat Microphone Performance in Automatic Speech Recognition on Motorcycles
  T. Winkler, S. Pronkine, R. Bardeli and J. Köhler [C]
[A]

Koelewijn, T.

• Auditory capture in an auditory and visual spatial cueing task
  T. Koelewijn, M. van der Hoeven and A. Bronkhorst [I]
[A]

Previous cueing studies show that a sound coming from a particular location in space can capture auditory and visual attention. We conducted two cueing studies where a baseline condition in the form of an auditory spatially diffuse cue was included, using out-of-phase presentation through two loudspeakers. The first study examined how an auditory non-informative spatial cue influences detection and localization of auditory targets. Results show that, compared to the baseline, cues shortened reaction times at small cue-target angles (up to 7) and increased them at larger angles. The reaction time difference was up to 80 ms. Furthermore, cueing seems to have no effect on sound localization: false alarms were normally distributed around the hits. These findings demonstrate that strong auditory capture effects occur that depend on cue-target distance. The second study shows how an auditory non-informative spatial cue influences detection of visual targets. Here the results show performance benefits when the target was preceded by a cue presented at target location and performance costs when a cue was presented elsewhere. Additionally, this study shows that even when attention is highly visually focused by means of a 100% valid endogenous visual cue, auditory stimuli still capture visual attention.

Költzsch, P.

• What do the acoustics of the 'occidental' Europe owe the Arabs?
  P. Költzsch [I]
[A]

Whilst the Arabs translated the results of the ancient Greek science almost without exception into Arabic until the middle of the 9th Century, and therefore made them known for the Arab world, the Greek heritage was largely unknown in Central Europe at the beginning of the 2nd Millennium. About from this time on, a transfer process began towards Central Europe, primarily for the scientific and philosophical discoveries of the Greeks. This happened because of the Arab occupation of Sicily (9th to 11th century) and of Spain (8th to 15th century), through the translations from Arabic into Latin and through multi-transfer paths (universities, libraries, etc.). However the Arab scientists were not only mediators, in an impressive way they also commented on the works of the Greeks and introduced their own new scientific ideas. Consequently the ancient Hellenistic and the Arab knowledge were spread in the 12. /13. Century in Central Europe. The lecture tries to answer the following questions for the area of acoustics: Which works of ancient Greek scientists have been passed on to the Arabs? Which Arab scientists were involved and how have they inspired the ''occidental'' culture in Europe through their own scientific achievements? [Lecture in German]

Koeman, T.

• Background noise: an increasing environmental problem
  J. Jabben, D. Bergmans, E. Schreurs and T. Koeman [C]
[A]

Noise quality in most studies is related to the average noise level (Lden) only, although this indicator merely expresses the average intensity of the soundscape. A complementary feature of increasing importance is the background noise level, which is the noise level that is exceeded during 95 % of the observation time. Inhabitants of areas with a high background noise level may be more vulnerable to annoyance or health effects, as recuperation times are narrowed. Due to strong growth of road, railway and air traffic, background noise levels have steadily increased in the Netherlands. In order to assess the impact of environmental noise levels from transport sources, RIVM in cooperation with the Dutch National Aerospace Laboratory have started a research project in 2007 with the aim of setting up background noise maps for road-, railway- and air traffic. This paper discusses the mapping and characteristic behavior of background noise. The implications of the background noise levels on health and annoyance will be discussed. The results indicate that a further increase of background noise levels in the near future is to be expected that may affect the Dutch population. Hence an intensified reduction policy of noise emission is most urgent.

König, D.

• Slat Noise Source Identification for a High-Lift Configuration
  D. König, S.R. Koh, W. Schröder and M. Meinke [I]
[A]

To investigate the noise sources of a high-lift configuration consisting of a slat and a main wing an efficient hybrid LES/CAA method is used. A freestream Mach number of M=0.16 and an angle of attack of 13 deg is assumed. The Reynolds number, based on the clean chord length and the freestream velocity is Re=1.4*106. The LES is done by a finite volume solver of second order accuracy. The used computational mesh consists of 55 million grid points. The acoustic analyses are performed by solving the acoustic perturbation equations (APE) whereas the perturbed Lamb vector is used as the only source, i.e., vortex sound dominates. The pressure distribution and the mean velocity field are in good agreement with experimental data. In the slat cove shear layer vortical structures similar to those in plane shear layers and impinging jets are observed. The acoustic simulation shows two pronounced areas where sound is generated, i.e., the main wing trailing edge and the slat region, the latter of which is dominant. The slat gap is identified to be responsible for the generated sound between 1 and 3 kHz via space and cross correlations, respectively, between the acoustic pressure and the perturbed Lamb vector.

Körpert, K.

• The new electronic report of noise-exposure-measurements and audiometric-examinations in the AUVA
  W. Wahler, K. Körpert and A. Gaafar [C]
[A]

The AUVA is the Austrian Social Insurance for occupational risks for approximately 3 million employed persons and 1.3 million school children and students. Therefore, a central motivation of the AUVA is the prevention of occupational accidents and diseases, e.g. the noise-induced hearing loss. One major element of the prevention activities are the visits of prevention workers in companies to improve the employees' conditions of hearing by an audiometric examination. To pick out the employees which are heavily exposed, it is necessary to determine the noise exposure level at work places by measurement and calculation. The data entry of the detailed noise-exposure and audiometric studies can be now electronically supported by a new computer application called EPOS. The objective of this paper is to report on the challenges and experiences gained during a five year multidisciplinary software development project and the essential components of the software will be presented.

Koh, S.R.

• Slat Noise Source Identification for a High-Lift Configuration
  D. König, S.R. Koh, W. Schröder and M. Meinke [I]
[A]

To investigate the noise sources of a high-lift configuration consisting of a slat and a main wing an efficient hybrid LES/CAA method is used. A freestream Mach number of M=0.16 and an angle of attack of 13 deg is assumed. The Reynolds number, based on the clean chord length and the freestream velocity is Re=1.4*106. The LES is done by a finite volume solver of second order accuracy. The used computational mesh consists of 55 million grid points. The acoustic analyses are performed by solving the acoustic perturbation equations (APE) whereas the perturbed Lamb vector is used as the only source, i.e., vortex sound dominates. The pressure distribution and the mean velocity field are in good agreement with experimental data. In the slat cove shear layer vortical structures similar to those in plane shear layers and impinging jets are observed. The acoustic simulation shows two pronounced areas where sound is generated, i.e., the main wing trailing edge and the slat region, the latter of which is dominant. The slat gap is identified to be responsible for the generated sound between 1 and 3 kHz via space and cross correlations, respectively, between the acoustic pressure and the perturbed Lamb vector.

Kohler, T.

• On the Classification of Acoustic Sequences for Intervention in Essential Hypertension
  P. Friedrich, T. Kohler and B. Wolf [C]
[A]

Acoustic signals can modulate the human metabolic and central-nervous functions and evoke physiological effects. Especially the anti-hypertensive effect of certain iterative sound-patterns as possible intervention for essential hypertension is examined in many recent studies as well as our own research work. It was shown that the acoustic sequences and sound clusters used for intervention can decrease blood pressure significantly. For a therapy, however, it is necessary to identify the musical active ingredients of the sounds in the sense of active pharmaceutical ingredients. This contribution discusses the systematic analysis of the musical features that are responsible for the anti-hypertensive effect. More than 400 features were extracted and investigated in terms of their relevance concerning a blood pressure lowering effect. The 17 most significant characteristics were used to develop a classifier based on the discriminant analysis that decides whether a sound pattern is sedative or stimulating. This will be