Speech denoising based on perceptual weighting filter
Enqing Dong, Xiaoxiang Pu · 2008
A new speech denoising algorithm for white noise environment based on perceptual weighting filter is proposed, which combines the spectrum subtraction and adopts auditory perception properties in the traditional Wiener filter. The main contribution of the algorithm is to modify the traditional Wiener filter by adopting the perceptual weighting function via the linear prediction technique. The algorithm can redistribute the noise spectrum according to the clear speech, which not only maintains the advantage of the Wiener filter, but also further reduces the influence of the noise. The simulation result shows that the algorithm can reduce the residual noise effectively, and improve the quality of the speech and have very low computational complexity.