AudioScopeV2: Audio-Visual Attention Architectures for Calibrated Open-Domain On-Screen Sound Separation

Efthymios Tzinis, Scott Wisdom*, Tal Remez, John R. Hershey

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


We introduce AudioScopeV2, a state-of-the-art universal audio-visual on-screen sound separation system which is capable of learning to separate sounds and associate them with on-screen objects by looking at in-the-wild videos. We identify several limitations of previous work on audio-visual on-screen sound separation, including the coarse resolution of spatio-temporal attention, poor convergence of the audio separation model, limited variety in training and evaluation data, and failure to account for the trade off between preservation of on-screen sounds and suppression of off-screen sounds. We provide solutions to all of these issues. Our proposed cross-modal and self-attention network architectures capture audio-visual dependencies at a finer resolution over time, and we also propose efficient separable variants that are capable of scaling to longer videos without sacrificing much performance. We also find that pre-training the separation model only on audio greatly improves results. For training and evaluation, we collected new human annotations of on-screen sounds from a large database of in-the-wild videos (YFCC100M). This new dataset is more diverse and challenging. Finally, we propose a calibration procedure that allows exact tuning of on-screen reconstruction versus off-screen suppression, which greatly simplifies comparing performance between models with different operating points. Overall, our experimental results show marked improvements in on-screen separation performance under much more general conditions than previous methods with minimal additional computational complexity.

Original languageEnglish
Title of host publicationComputer Vision – ECCV 2022 - 17th European Conference, Proceedings
EditorsShai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner
PublisherSpringer Science and Business Media Deutschland GmbH
Number of pages18
ISBN (Print)9783031198359
StatePublished - 2022
Externally publishedYes
Event17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel
Duration: 23 Oct 202227 Oct 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13697 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference17th European Conference on Computer Vision, ECCV 2022
CityTel Aviv


  • Audio-visual sound separation
  • Self-attention


Dive into the research topics of 'AudioScopeV2: Audio-Visual Attention Architectures for Calibrated Open-Domain On-Screen Sound Separation'. Together they form a unique fingerprint.

Cite this