TR2021-144

Convolutive Prediction for Monaural Speech Dereverberation and Noisy-Reverberant Speaker Separation


Abstract:

We propose to exploit the linear-filter structure of reverberation within a supervised deep learning based monaural speech dereverberation framework. The key idea is to first estimate the direct-path signal of the target speaker using a DNN and then identify signals that are decayed and delayed copies of the estimated direct-path signal, as these can be reliably considered as reverberation. We then modify the proposed algorithm for speaker separation in reverberant and noisy-reverberant conditions. State-of-the-art speech dereverberation and speaker separation results are obtained on the REVERB, SMS-WSJ, and WHAMR! datasets.

 

  • Related Publication

  •  Wang, Z.-Q., Wichern, G., Le Roux, J., "Convolutive Prediction for Monaural Speech Dereverberation and Noisy-Reverberant Speaker Separation", arXiv, August 2021.
    BibTeX arXiv
    • @article{Wang2021aug,
    • author = {Wang, Zhong-Qiu and Wichern, Gordon and Le Roux, Jonathan},
    • title = {Convolutive Prediction for Monaural Speech Dereverberation and Noisy-Reverberant Speaker Separation},
    • journal = {arXiv},
    • year = 2021,
    • month = aug,
    • url = {https://arxiv.org/abs/2108.07376}
    • }