ONLINE SPEECH ENHANCEMENT USING DIFFERENT ADAPTIVE FILTERS

Abstract

Adaptive filtering has become a spacious area of researcher since last few decades in the field of communication. Adaptive noise cancellation is an approach used for noise reduction in speech signal. The speech signal easily gets contaminated with background noise. Channel noise addition makes this speech signal even poorer. Speech signal and noise signal both change continuously with time, then to separate them only adaptive filtering is desirable. This paper deals with cancellation of noise on speech signal using two old (LMS and NLMS) and one new UNANR algorithm. The UNANR (Unbiased and Normalized Adaptive Noise Rejection) model does not contain any bias unit, and the coefficients are adaptively updated by using the steepest-descent algorithm. Two modulation techniques, AM and FM are applied separately in combination with two communication channels i.e. AWGN and Rician. Signal quality parameter PSNR and RMSE measured and compared with respect to SNR. The results show that the performance of the UNANR based algorithm is superior to that of the LMS algorithm in noise reduction.