The role of temporal modulation processing in speech/non‐speech discrimination tasks.

In this paper, temporal modulation characteristics of speech and noise from the point of view of speech/non‐speech discrimination are analyzed. Although previous psychoacoustic studies have shown that temporal modulation components below 16 Hz are important for speech intelligibility, there is no reported analysis of modulation components from the point of view of speech/noise discrimination. Our data‐driven analysis of modulation components of speech and noise reveals that speech and noise are more accurately classified by low‐pass modulation frequencies than band‐pass ones [H. You and A. Alwan, in Interspeech Proceedings (2009) pp. 36–39]. Effects of additive noise on the modulation characteristics of speech signals are also analyzed. Based on the analysis, a frequency adaptive modulation processing algorithm for a noise robust automatic speech recognition task is proposed. Speech recognition experiments are performed to compare the proposed algorithm with other noise robust front‐ends, including RASTA ...