A New Speech Enhancment algorithm in Hearing Aid based on Wavelet Transform

Abstract

Voice Enhancement systems are used to remove background inference in a speech signal and are become an main component of modern hearing aid. In everyday life the speech communication is vivid uses for the hearing impaired and the numerous other applications. Speech is the fundamental means of human communication. After over thirty years of research enhancement algorithm. Which offer superior noise reduction over current methods? All speech enhancements suffer from distortion for the residual noise due to imperfect noise removal. It is always require to perform denoising in voice processing system operating in highly noise because of wavelet transform is one of the popular techniques used in signal enhancement, In the present paper wavelet thresholding and wavelet packet thresholding method have been used to decrease the noise from the voice signal. A simple threshold method is presented to compute the optimum threshold value. Mean square error(MSE) at different values of SNR is computed to method like traditional speech subtraction, wiener filtering method, spectral subtraction with MMSE etc.The result obtained is compared with the other voice enhancement algorithm given in various reference papers. In comparison to other traditional methods we get improved result in terms of SNR and MSE.Simulation done in MATLAB platform.