Modeling Noise Influence to Speech Intelligibility Non-Intrusively by Reduced Speech Dynamic Range

The noise influence to speech signal waveform can be characterized by reduced speech dynamic range (rDR). This motivated the present work to propose an rDR-based intelligibility measure (denoted as rDRm) that could be used to non-intrusively (i.e., do not require clean reference speech signal) predict speech intelligibility in noise and is computed only using the dynamic range extracted from the noisecorrupted speech. The rDRm indices were evaluated with intelligibility scores obtained from normal-hearing listeners presented with sentences corrupted by four types of maskers in a total of 22 conditions. High correlation (r=0.93) was obtained between rDRm values and listeners’ sentence recognition scores, and this correlation was comparable to those computed with existing intrusive and non-intrusive intelligibility measures. This suggests that the dynamic range of speech signal may work as a simple but efficient predictor of speech intelligibility in noise, whose computation does not need access to the clean reference speech signal.