A perceptual masking approach for noise robust speech recognition

Abstract

This article describes a modified technique for enhancing noisy speech to improve automatic speech recognition (ASR) performance. The proposed approach improves the widely used spectral subtraction which inherently suffers from the associated musical noise effects. Through a psychoacoustic masking and critical band variance normalization technique, the artifacts produced by spectral subtraction are minimized for improving the ASR accuracy. The popular advanced ETSI-2 front end is tested for comparison purposes. The performed speech recognition evaluations on the noisy standard AURORA-2 tasks show enhanced performance for all noise conditions.

Intra-frame cepstral sub-band weighting and histogram equalization for noise-robust speech recognition

“…Type 4: II requires three HEQ operations, the same as S-HEQ, demonstrating that S-HEQ and WS-HEQ (1) II are similar in computational complexity.…”

“…Equations 12 to 15, by assigning α as less than 1.0, WS-HEQ (1) , which requires three HEQ operations, displays the best behavior, regardless of the selected structure. However, the two types that require only two HEQ operations (i.e., WS-HEQ (2) and …”

“…A significant number of noise-robustness techniques have been proposed to address the noise problem, and one prevailing subset of these techniques is focused on reducing the statistical mismatch of speech features in the training and testing conditions of the recognizer. Typical examples are perceptual masking [1], empirical mode decomposition [2], optimally modified log-spectral amplitude estimation [3], wavelet packet decomposition with AR modeling [4], cepstral mean and variance normalization (MVN) [5], cepstral histogram normalization (CHN) [6,7], MVN with ARMA filtering (MVA) [8], higher order cepstral moment normalization (HOCMN) [9], and temporal structure normalization (TSN) [10]. In some of these methods, the compensation is performed on each individual cepstral channel sequence of an utterance by assuming that these channels are mostly uncorrelated [7].…”

Signal Subspace Speech Enhancement with Oblique Projection and Normalization

“…To reduce the effect of any present tones which caused by increased variance at random frequencies, M anti and Matassoni [19] performed variance normalizat across the critical bands for spectral subtraction speech hancement algorithm. The variance is computed as in (…”

“…Lu [18], employed an optimal smoothing factor, adapted by the variation of signal to spectral deviation ratio in successive frames. Variance normalization was used in spectral subtraction speech enhancement algorithm by Maganti and Matassoni [19] across the critical bands to smoothen the output signal, removing the spikes in the output which reduced the effect of increased variance at random frequencies.…”