A two-channel fast NLMS algorithm for speech enhancement and acoustic noise reduction

摘要

This paper addresses the problem of acoustic noise reduction and speech enhancement in new telecommunication systems by adaptive filtering algorithms. We propose a new two-channel forward BSS algorithm based on signal prediction to give an automatic algorithm with a very nice behavior at the output. This algorithm is called the two-channel fast normalized least mean square (TCFNLMS) algorithm. This new algorithm shows superiority in comparison with classical two-channel NlMS in terms of the following objective criteria: the segmental signal to noise ratio (SegSNR), the cepstral distance (CD), the system mismatch (SM) and the segmental mean square error (SegMSE).