A Real-time Objective Speech Analysis Tool for Analysis of Impaired Speech
* Presenting author
Abstract:
In the context of analysis and treatment of voice and speech disorders, computer-assisted examinations are a promising possibility for faster, more objective, and comparable analyses. Our aim is to create a modular, flexible, and easy-to-use tool to support and enable more possibilities in the field of speech therapy and medicine. Therefore, the choice and number of speech tasks and many other parameters are adjustable. Furthermore, various additional functions such as surveys and patient information queries are included. During the examination, real-time processing allows direct feedback to the patient, detection of missing speech, new tasks, and extraction of additional information from the recordings. We present such a tool and further want to investigate the effect of real-time feedback in the context of the Lombard effect. Therefore, ambient noise, possibly with different sound scenarios, should be used to indicate a voice that is too quiet and to animate the patient to raise his voice according to the noise level. The patient's reaction to these changes can then be used to analyze, but also to train the voice, based on the approach of speech therapy concepts such as Lee Silverman Voice Treatment (LSVT).