Efficient algorithm for Speech Enhancement using Adaptive filter

The present system of speech enhancement is developing by adaptive filtering approach in digital filters. The adaptive filter utilizes the least mean square algorithm for noise removal, but in practical application of LMS algorithm, a key parameter is the step size. As it is known, if the step size is large, the convergence rate of LMS algorithm will be rapid, but the steady-state mean square error (MSE) will increase. That means speech enhancement has some limitations in SNR improvement and rate of convergence. In this project an optimal estimation of adaptive filtering using Unbiased and normalized adaptation noise reduction (UNANR) algorithm has been implemented for the noisy speech. The aim of this paper is to implement various adaptive noise cancellers for speech enhancement based on gradient steepest descent approach. In this paper, we can say that the signal to noise improvement in the input signal after UNANR filtering is much higher and it is also simple to implementation compared to that of LMS filter algorithm. Therefore we conclude that the Unbiased and Normalized adaptation noise reduction (UNANR) algorithm is an efficient adaptive filtering algorithm than least mean square (LMS) algorithm. Keyword: LMS, UNANR, MSE, Adaptive filter etc.