A wavelet-NLMS algorithm implemented in FBSS for acoustic noise reduction and speech enhancement

In this paper, we address the problem of noise reduction and speech enhancement by adaptive filtering algorithm. Recently, the well-known forward blind source separation (FBSS) structure has been largely studied and intensively used to reduce acoustic noise components and to enhance speech signal. This FBSS structure is often combined with an adaptive algorithm to make efficient the adaptation of the cross-filters that identify the noise to be suppressed at the output. In this paper, we propose to use wavelet transform decomposition of the FBSS structure to get a new two-channel wavelet forward BSS (WFBSS) algorithm. The proposed algorithm provides a better compromise between time and frequency resolution and robustness of the noise reduction process when compared with the classical two-channel forward BSS (FBSS) algorithm. Simulation results prove the efficiency of the proposed WFBSS algorithm in comparison with conventional ones in terms of several objective and subjective criteria.