Wavelet transform based coherence function for multi-channel speech enhancement

This paper addresses the problem of enhancing a speech signal acquired by a microphone array used for hands-free voice communication applications. A new algorithm based on the coherence function developed in the wavelet domain and applied to the beamforming output signal is presented. The wavelet coherence function based nonlinear post-filtering provides a further noise suppression. Its performances is comparable to that of the wavelet Wiener filter proposed in our earlier work [1].