Time-Domain Neural Network Approach for Speech Bandwidth Extension

Xiang Hao, Chenglin Xu, Nana Hou, Lei Xie, Eng Siong Chng, Haizhou Li

Abstract

In this paper, we study the time-domain neural network approach for speech bandwidth extension. We propose a network architecture, named multi-scale fusion neural network (MfNet), that gradually restores the low-frequency signal and predicts the high-frequency signal through the exchange of information across different scale representations. We propose a training scheme to optimize the network with a combination of perceptual loss and time-domain adversarial loss. Experiments show the proposed multi-scale fusion network consistently outperforms the competing methods in terms of perceptual evaluation of speech quality (PESQ), signal to distortion rate (SDR), signal to noise ratio (SNR), log-spectral distance (LSD) and word error rate (WER). More promisingly, the multi-scale fusion network requires only 10% of the parameters of the time-domain reference baseline.
BibTeX
@inproceedings{hao2020time,
  author = {Xiang Hao and Chenglin Xu and Nana Hou and Lei Xie and Eng Siong Chng and Haizhou Li},
  title = {{Time-Domain Neural Network Approach for Speech Bandwidth Extension}},
  booktitle = {IEEE International Conference on Acoustics, Speech and Signal Processing},
  year = {2020},
  doi = {10.1109/icassp40776.2020.9054551},
}

Loading related content...

Figure