Summary

The performance of modern speech technologies (e.g., automatic speech recognizers, speech enhancement techniques, or cocktail-party processors) is drastically deteriorated when sound signals are employed that have been produced in realistic environments. The main disturbing factors introduced by a realistic environment are interfering noise sources as well as reverberation produced by sound being reflected from walls. The human auditory system has developed sophisticated bottom-up as well as top-down processes that enable a listener to understand speech even under such adverse conditions. Hence, the aim of many researchers has been to understand and model the mechanisms that underlie the auditory/human processing of speech in realistic environments and to apply the gained knowledge to improve the performance of modern speech technologies.

The overwhelming aim of the SPAPAC workshop is to bring together researchers that work in the research areas of speech perception/development, speech processing/recognition , and auditory processing/modelling in order to stimulate new interdisciplinary approaches to the complex problem of understanding auditory/speech perception/processing under adverse conditions.

Materials

• Workshop Outcomes Report