Speech intelligibility enhancement in noisy reverberant conditions

Abstract

The ideal binary mask (IBM) that was found to yield substantial improvements in speech intelligibility in noise has recently been extended to reverberant conditions by regarding the direct sound and early reflections of target speech as the desired signal. In this paper, we investigate the effect of the division between early and late reflections on the intelligibility of the IBM-processed signal in noisy reverberant conditions. The desired signal was obtained by applying the IBMs to the noisy reverberant mixture signal, and then presented to normal-hearing listeners for word recognition. Results showed that the IBMs with different divisions between early and late reflections provided substantial improvements in speech intelligibility over the unprocessed mixture signals in all conditions tested, and there were small, but statistically significant, differences in speech intelligibility between the different IBMs in some conditions tested.