Noise estimation for speech enhancement algorithms with post-smoothness processor incorporating global posterior SNR

摘要

Performances of speech enhancement algorithms depend greatly on the accuracy of the estimated noise. In this paper, we explain in details the relationship between noise estimation and denoised speech quality. We particularly show the importance of noise smoothing over frames on denoising quality. This study leads to the development of a new technique to smooth the estimated noise power spectrum over frequency bins of the same frame. Compared to inter-frame smoothing, experimental results show that the proposed intra-frame smoothing has a good impact on the denoised speech. Quality is evaluated over three dimensions: speech distortion, residual background noise and overall quality.