Multi-Task Sub-Band Network For Deep Residual Echo Suppression
Audio and Speech Processing
2023-03-14 v1
Abstract
This paper introduces the SWANT team entry to the ICASSP 2023 AEC Challenge. We submit a system that cascades a linear filter with a neural post-filter. Particularly, we adopt sub-band processing to handle full-band signals and shape the network with multi-task learning, where dual signal voice activity detection (DSVAD) and echo estimation are adopted as auxiliary tasks. Moreover, we particularly improve the time frequency convolution module (TFCM) to increase the receptive field using small convolution kernels. Finally, our system has ranked 4th in ICASSP 2023 AEC Challenge Non-personalized track.
Cite
@article{arxiv.2303.06404,
title = {Multi-Task Sub-Band Network For Deep Residual Echo Suppression},
author = {Jiayao Sun and Dawei Luo and Zhaoxia Li and Jindong Li and Yukai Ju and Yang Li},
journal= {arXiv preprint arXiv:2303.06404},
year = {2023}
}