Novel Wiener sub-band processing schemes for binaural adaptive speech-enhancement

Abstract

A novel application of Wiener filtering is proposed for processing band-limited noisy signals in a binaural adaptive speech-enhancement scheme. Two different Weiner filtering based implementations have been developed, both with the aim of further reducing residual incoherent noise components resulting from application of conventional adaptive sub-band noise cancellers. Initial comparative results achieved in simulation experiments using real anechoic speech corrupted with simulated interference demonstrate that the proposed binaural sub-band schemes employing Wiener filtering are capable of significantly outperforming conventional wideband and sub-band noise-cancellation methods