Noise-Aware Speech Separation with Contrastive Learning

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Zizheng, Chen, Chen, Chen, Hsin-Hung, Liu, Xiang, Hu, Yuchen, Chng, Eng Siong
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913187998203904
author Zhang, Zizheng
Chen, Chen
Chen, Hsin-Hung
Liu, Xiang
Hu, Yuchen
Chng, Eng Siong
author_facet Zhang, Zizheng
Chen, Chen
Chen, Hsin-Hung
Liu, Xiang
Hu, Yuchen
Chng, Eng Siong
contents Recently, speech separation (SS) task has achieved remarkable progress driven by deep learning technique. However, it is still challenging to separate target speech from noisy mixture, as the neural model is vulnerable to assign background noise to each speaker. In this paper, we propose a noise-aware SS (NASS) method, which aims to improve the speech quality for separated signals under noisy conditions. Specifically, NASS views background noise as an additional output and predicts it along with other speakers in a mask-based manner. To effectively denoise, we introduce patch-wise contrastive learning (PCL) between noise and speaker representations from the decoder input and encoder output. PCL loss aims to minimize the mutual information between predicted noise and other speakers at multiple-patch level to suppress the noise information in separated signals. Experimental results show that NASS achieves 1 to 2dB SI-SNRi or SDRi over DPRNN and Sepformer on WHAM! and LibriMix noisy datasets, with less than 0.1M parameter increase.
format Preprint
id arxiv_https___arxiv_org_abs_2305_10761
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Noise-Aware Speech Separation with Contrastive Learning
Zhang, Zizheng
Chen, Chen
Chen, Hsin-Hung
Liu, Xiang
Hu, Yuchen
Chng, Eng Siong
Sound
Audio and Speech Processing
Recently, speech separation (SS) task has achieved remarkable progress driven by deep learning technique. However, it is still challenging to separate target speech from noisy mixture, as the neural model is vulnerable to assign background noise to each speaker. In this paper, we propose a noise-aware SS (NASS) method, which aims to improve the speech quality for separated signals under noisy conditions. Specifically, NASS views background noise as an additional output and predicts it along with other speakers in a mask-based manner. To effectively denoise, we introduce patch-wise contrastive learning (PCL) between noise and speaker representations from the decoder input and encoder output. PCL loss aims to minimize the mutual information between predicted noise and other speakers at multiple-patch level to suppress the noise information in separated signals. Experimental results show that NASS achieves 1 to 2dB SI-SNRi or SDRi over DPRNN and Sepformer on WHAM! and LibriMix noisy datasets, with less than 0.1M parameter increase.
title Noise-Aware Speech Separation with Contrastive Learning
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2305.10761