Handling Background Noise in Neural Speech Generation
Audio and Speech Processing
2021-02-25 v1 Sound
Abstract
Recent advances in neural-network based generative modeling of speech has shown great potential for speech coding. However, the performance of such models drops when the input is not clean speech, e.g., in the presence of background noise, preventing its use in practical applications. In this paper we examine the reason and discuss methods to overcome this issue. Placing a denoising preprocessing stage when extracting features and target clean speech during training is shown to be the best performing strategy.
Cite
@article{arxiv.2102.11906,
title = {Handling Background Noise in Neural Speech Generation},
author = {Tom Denton and Alejandro Luebs and Felicia S. C. Lim and Andrew Storus and Hengchin Yeh and W. Bastiaan Kleijn and Jan Skoglund},
journal= {arXiv preprint arXiv:2102.11906},
year = {2021}
}
Comments
5 pages, 3 figures, presented at the Asilomar Conference on Signals, Systems, and Computers 2020