Speech enhancement using the multistage Wiener filter

Abstract

In this paper we develop a subspace speech enhancement approach for estimating a signal which has been degraded by additive uncorrelated noise. This problem has numerous applications such as in hearing aids and automatic speech recognition in noisy environments. The proposed approach utilizes the multistage Wiener filter (MWF). This filter is constructed from a Krylov subspace associated with the Wiener filter for this problem. The principles and performance of this approach are described. The approach provides signals with higher quality compared to the full Wiener filter and the signal subspace method as evident from informal subjective listening tests, the commonly used PESQ objective test and improved speech recognition performance. In informal listening tests the listeners preferred the MWF enhancement over the other enhancement methods.

Speech enhancement by LSTM-based noise suppression followed by CNN-based speech restoration

“…[33]) or most often different models are used in succession to improve the enhancement performance. Applied to classical speech enhancement, this principle is generally used to achieve a higher noise attenuation, e.g., with the multi-stage Wiener filter approach [34], which in turn leads to degradations of the speech quality. Different from that, some studies have focused on first performing speech separation and subsequently enhancing the separated signals using nonnegative matrix factorization [35] or Gaussian mixture models [36].…”