Spectral-domain speech enhancement for speech recognition

Abstract

Speech recognition performance deteriorates in face of unknown noise. Speech enhancement offers a solution by reducing the noise in speech at runtime. However, it also introduces artificial distortion to the speech signal. In this paper, we aim at reducing the artifacts that have adverse effects on speech recognition. With this motivation, we propose a modification scheme including a smoothing adaptation to frame signal-to-noise ratio (SNR) and a reestimation of a priori SNR for spectral-domain speech enhancement. The experiment shows that the proposed scheme of enhancement significantly improves the performance of the state-of-the-art speech recognition over the baseline speech enhancement techniques.