CASCADE COMBINATION OF WAVELET AND ADAPTIVE FILTER FOR NOISE CANCELLATION

Abstract

Speech enhancement techniques are very important in the field of signal processing for their numerous applications. They are employed in many contexts such as hands-free telephony, hearing aid systems, re-mastering of audio recordings, preprocessing for speech recognition interfaces, etc. In this paper, an efficient cascade combination approach for cancellation of noises in speech signals is discussed. It is possible to obtain higher performance by cascading two or more algorithms. A better noise removal approach is presented by cascading wavelet and adaptive filters. Results show that the cascade approach gives high Peak Signal to Noise Ratio (PSNR) and low Root Mean Square Error (RMSE) than individual performances of wavelet and an adaptive filter.