Investigating an optimum suppression rate of steady-state portions of speech that improves intelligibility the most as a pre-processing approach in reverberant environments

Abstract

Steady-state suppression has been proposed as a pre-processing approach to improve speech intelligibility in reverberant environments [Arai et al., Acoust. Sci. Technol. 23, 229–232 (2002)]. The goal of this work is to find the suppression rate that improves intelligibility the most. This is done by exploring the relationship between the suppression rates of steady-state portions of speech signals in a reverberant environment and syllable identification. This needs to be investigated because the optimum suppression rate might depend on the amount of overlap-masking, which in turn is determined by the reverberation. A syllable identification test was conducted with 21 normal-hearing listeners in two reverberant environments and six suppression rates. Results show a significant improvement in intelligibility for a 40% and a 50% suppression rate for a reverberation time of 1.1 s and a 60% suppression rate for a reverberation time of 1.3 s. Findings confirm that intelligibility is affected by both suppression rate and reverberation time. Results also show that an optimal suppression rate might not be derived from reverberation time alone. The spectral balance of the room response should also be taken into consideration. [Research supported by JSPS (16203041 and 176911).]