Temporally enhanced speech is more intelligible in reverberant environments

Proc. of the Western Pacific Acoustics Conference (WESPAC), Seoul, 2006 (Invited Paper)

Temporally enhanced speech is more intelligible in reverberant environments

T. Arai and N. Hodoshima

Abstract: Reverberation causes degradation in speech comprehension, especially for elderly people, the hearing-impaired and non-native listeners. In order to prevent intelligibility degradation, we developed several pre-processing techniques, where signals are processed before being radiated through the loudspeakers of a public address system. The two main techniques are modulation filtering and steady-state suppression, both of which enhance the temporal dynamics of speech. Previously we had found that the important frequencies of temporal dynamics for speech perception, the modulation frequencies, lie between 1-16 Hz (Arai et al., 1999). Therefore, we configured the modulation filtering to emphasize these modulation frequencies (Kusumoto et al., 2005). In this paper, we mainly discussed steady-state suppression (Arai et al., 2001, 2002), which suppresses steady-state portions of speech to reduce overlap-masking and improve speech intelligibility for young, elderly, and non-native listeners in reverberant environments (Hodoshima et al., 2006a; Miyauchi et al., 2005; Hodoshima et al., 2006b). Especially, we discussed how the “pre-processing” technique, such as the steady-state suppression, is effective with recent results including speech-rate slowing with the steady-state suppression exceeding a simple speech-rate slowing approach (Arai et al., 2005).

Keywords: Speech enhancement, Reverberation, Speech Intelligibility, Steady-State Suppression

[PDF (142 kB)]