New speech enhancement using Gamma tone filters and Perceptual Wiener filtering based on sub banding

Abstract

A two stage noise reduction approach is proposed in this paper. This paper focuses on Wiener filter based single channel speech enhancement method. First stage consists of Gamma tone filtering method for sub banding and synthesizing; perceptualWiener filter is used to apply gain function on each sub band to reduce the noise at each critical frequency. Performance of perceptual Wiener filter is improved by finding self adaptive averaging factor to estimate a priori SNR. To reduce residual musical noise that remains in the enhanced speech, perceptual weighting filter is developed based on simultaneous and temporal masking effects of human auditory system is used as a second stage in the proposed system. An unvoiced speech enhancement algorithm is also integrated into the system to improve the intelligibility of speech. Through rigorous objective and subjective evaluations, it is shown that the proposed speech enhancement system is capable of reducing noise with little speech distortion in adverse noise environments and overall performance is superior to several competitive methods.