Acoustic Echo cancellation using variable step size based adaptive filtering with performance measure

Abstract

Acoustic echo is a most frequent occurrence in modern telecommunication systems. This paper describes the different adaptive filtering algorithms to minimize an unwanted echo and improves quality of speech signals. Adaptive filtering has been an active area of research. Here we are using NLMS, MMAX-NLMS and MMAXNLMSvss algorithms to analyze speech signal. In both NLMS and MMAX-NLMS algorithms the step size parameter is fixed. That forces a performance compromise between fast convergence and small steady state misadjustment. So, Partial update adaptive algorithms have been proposed by deriving a variable step-size for the MMAX-NLMS algorithm using its mean square deviation. The proposed MMAX-NLMSvss algorithm is tested with speech input signal and the result shows that it has fast convergence time, lower intricacy compared to the NLMS, MMAXNLMS algorithm.