VOICE IMPROVEMENT USING ADAPTIVE FILTER

Abstract

The removal of noise from speech signals has applications ranging from cellular communications to front ends for speech recognition systems. In this paper, an optimal estimate of adaptive filtering using least mean algorithm has been implemented for the observed noisy speech. The algorithm yields better results in noise reduction with significantly less distortions and artificial noise. In this paper, it gives the concept of speech enhancement in a practical approach, using different speech enhancement algorithms. Extraction of high resolution information signals is important in all practical applications. The Least Mean Square (LMS) algorithm is a basic adaptive algorithm has been extensively used in many applications as a consequence of its simplicity and robustness. In this paper we present a novel adaptive filter for de-noising the speech signals based on unbiased and normalized adaptive noise reduction (UNANR) algorithm. The UNANR model does not contain a bias unit, and the coefficients are adaptively updated by using the steepest-descent algorithm. The adaptive filter essentially minimizes the mean-squared error between a primary input, which is the noisy speech, and a reference input, which is either noise that is correlated in some way with the noise in the primary input or a signal that is correlated only with speech in the primary input. To measure the ability of the proposed implementation, signal to noise ratio improvement (SNRI) is calculated.