Contribution

Robustness of Instrument Identification to Spatial Variations and Reverberation in Musical Mixtures

* Presenting author
Day / Time: 18.03.2025, 14:40-15:00
Type: Lecture (in a structured session)
Abstract ID: DAS-DAGA2025/460
Abstract: Spatial cues from sound source positioning and reverberation impact auditory scene analysis. For renditions of music in a concert hall, both aspects shape the sound perception for the listener. Furthermore, the music itself and its orchestration emphasizes specific audible elements. However, it is unclear how instrument identification in musical mixtures as a musical scene analysis ability is affected by these variables.Musically trained listeners identified instruments from a set of eight Western orchestral instruments. Participants used a pitch cue to recognize a target instrument within short, four-instrument excerpts from Wagner’s Tristan prelude. An additional experiment was performed with controlled stationary chord structures of varied complexity. The spatial distribution of instruments in the simulated hall and its reverberation were manipulated using the room-acoustic modeling software TASCAR, with playback through a 16-loudspeaker array in a semi-anechoic room.Against initial hypotheses, spatial and reverb conditions had only marginal impact on listeners’ identification accuracies. However, instrument-dependent effects of mixture complexity and pitch position within the chord arose. These results suggest that instrument identification is a musical scene analysis ability that is robust to spatial variations and reverberation. Furthermore, the instrument-dependent findings complement considerations on orchestration and blend in orchestration practice.