Speech recognition in noise and reverberation in multichannel auralizations of two real rooms

Test stimuli is many speech-in-noise-tests contain little to no reverberation whereas people often report on difficulties specifically in understanding speech in reverberant conditions. The aim of this project is to investigate how advanced spatial sound technologies can be used in hearing diagnostics and to develop novel tools with more complex and authentic sound scenes. A recently developed auralization method (Spatial Decomposition Method) may be well suited for bringing advanced spatial sound technology into clinically feasible environments. Here, we present results of our study, where we have used the Finnish Matrix Sentence Test with multichannel auralizations of two real spaces with approx. one and two second reverberation times to measure speech reception thresholds in noise between anechoic/dry and reverberant conditions. Perhaps surprisingly, the results indicate a small but significant improvement in SRTs with one second reverberation time compared to the reference case without reverberation.