Time-Frequency Masking in the Complex Domain for Speech Dereverberation and Denoising

Donald S. Williamson, DeLiang Wang

IEEE/ACM Transactions on Audio Speech and Language Processing · 2017 · 218 citations · 41 references

Concepts

TL;DR

Speech is degraded by background noise and reverberation, reducing perceptual quality and intelligibility. The study aims to perform monaural speech separation in reverberant and noisy settings. A supervised deep neural network estimates a complex ideal ratio mask to enhance magnitude and phase, producing direct speech from reverberant and noisy input, and is evaluated on simulated and real room impulse responses with background noises. The method significantly improves objective speech quality and intelligibility, outperforming related approaches across many reverberant and noisy environments.

Abstract

In real-world situations, speech is masked by both background noise and reverberation, which negatively affect perceptual quality and intelligibility. In this paper, we address monaural speech separation in reverberant and noisy environments. We perform dereverberation and denoising using supervised learning with a deep neural network. Specifically, we enhance the magnitude and phase by performing separation with an estimate of the complex ideal ratio mask. We define the complex ideal ratio mask so that direct speech results after the mask is applied to reverberant and noisy speech. Our approach is evaluated using simulated and real room impulse responses, and with background noises. The proposed approach improves objective speech quality and intelligibility significantly. Evaluations and comparisons show that it outperforms related methods in many reverberant and noisy environments.

References

41