Saved in:
Bibliographic Details
Main Authors: Ma, T. Aleksandra, Yin, Sile, Yang, Li-Chia, Zhang, Shuo
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.21448
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913974029647872
author Ma, T. Aleksandra
Yin, Sile
Yang, Li-Chia
Zhang, Shuo
author_facet Ma, T. Aleksandra
Yin, Sile
Yang, Li-Chia
Zhang, Shuo
contents Speech enhancement in audio-only settings remains challenging, particularly in the presence of interfering speakers. This paper presents a simple yet effective real-time audio-visual speech enhancement (AVSE) system, RAVEN, which isolates and enhances the on-screen target speaker while suppressing interfering speakers and background noise. We investigate how visual embeddings learned from audio-visual speech recognition (AVSR) and active speaker detection (ASD) contribute to AVSE across different SNR conditions and numbers of interfering speakers. Our results show concatenating embeddings from AVSR and ASD models provides the greatest improvement in low-SNR, multi-speaker environments, while AVSR embeddings alone perform best in noise-only scenarios. In addition, we develop a real-time streaming system that operates on a computer CPU and we provide a video demonstration and code repository. To our knowledge, this is the first open-source implementation of a real-time AVSE system.
format Preprint
id arxiv_https___arxiv_org_abs_2507_21448
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Real-Time Audio-Visual Speech Enhancement Using Pre-trained Visual Representations
Ma, T. Aleksandra
Yin, Sile
Yang, Li-Chia
Zhang, Shuo
Audio and Speech Processing
Emerging Technologies
Machine Learning
Speech enhancement in audio-only settings remains challenging, particularly in the presence of interfering speakers. This paper presents a simple yet effective real-time audio-visual speech enhancement (AVSE) system, RAVEN, which isolates and enhances the on-screen target speaker while suppressing interfering speakers and background noise. We investigate how visual embeddings learned from audio-visual speech recognition (AVSR) and active speaker detection (ASD) contribute to AVSE across different SNR conditions and numbers of interfering speakers. Our results show concatenating embeddings from AVSR and ASD models provides the greatest improvement in low-SNR, multi-speaker environments, while AVSR embeddings alone perform best in noise-only scenarios. In addition, we develop a real-time streaming system that operates on a computer CPU and we provide a video demonstration and code repository. To our knowledge, this is the first open-source implementation of a real-time AVSE system.
title Real-Time Audio-Visual Speech Enhancement Using Pre-trained Visual Representations
topic Audio and Speech Processing
Emerging Technologies
Machine Learning
url https://arxiv.org/abs/2507.21448