Rank-1 Constrained Multichannel Wiener Filter for Speech Recognition in Noisy Environments

TL;DR

This paper introduces a rank-1 constrained multichannel Wiener filter for speech recognition in noisy environments, demonstrating significant WER reductions in real recordings by optimizing noise reduction with a new residual noise constraint.

Contribution

The paper proposes a novel rank-1 constrained MWF with a residual noise power constraint, improving speech recognition performance over existing multichannel filters in noisy conditions.

Findings

40% relative WER reduction over baseline WDAS

15% relative WER reduction over GEV-BAN

Speech recognition accuracy correlates with MFCC variance

Abstract

Multichannel linear filters, such as the Multichannel Wiener Filter (MWF) and the Generalized Eigenvalue (GEV) beamformer are popular signal processing techniques which can improve speech recognition performance. In this paper, we present an experimental study on these linear filters in a specific speech recognition task, namely the CHiME-4 challenge, which features real recordings in multiple noisy environments. Specifically, the rank-1 MWF is employed for noise reduction and a new constant residual noise power constraint is derived which enhances the recognition performance. To fulfill the underlying rank-1 assumption, the speech covariance matrix is reconstructed based on eigenvectors or generalized eigenvectors. Then the rank-1 constrained MWF is evaluated with alternative multichannel linear filters under the same framework, which involves a Bidirectional Long Short-Term Memory (BLSTM) network for mask estimation. The proposed filter outperforms alternative ones, leading to a 40% relative Word Error Rate (WER) reduction compared with the baseline Weighted Delay and Sum (WDAS) beamformer on the real test set, and a 15% relative WER reduction compared with the GEV-BAN method. The results also suggest that the speech recognition accuracy correlates more with the Mel-frequency cepstral coefficients (MFCC) feature variance than with the noise reduction or the speech distortion level.