Papers › Unsupervised Source Separation By Steering Pretrained Music Models

Unsupervised Source Separation By Steering Pretrained Music Models

25 Oct 2021arXiv:2110.13071archive 2025-07-28

Ethan Manilow, Patrick O'Reilly, Prem Seetharaman, Bryan Pardo

We showcase an unsupervised method that repurposes deep models trained for music generation and music tagging for audio source separation, without any retraining. An audio generation model is conditioned on an input mixture, producing a latent encoding of the audio used to generate audio. This generated audio is fed to a pretrained music tagger that creates source labels. The cross-entropy loss between the tag distribution for the generated audio and a predefined distribution for an isolated source is used to guide gradient ascent in the (unchanging) latent space of the generative model. This system does not update the weights of the generative model or the tagger, and only relies on moving through the generative model's latent space to produce separated sources. We use OpenAI's Jukebox as the pretrained generative model, and we couple it with four kinds of pretrained music taggers (two architectures and two tagging datasets). Experimental results on two source separation datasets, show this approach can produce separation estimates for a wider variety of sources than any tested supervised or unsupervised system. This work points to the vast and heretofore untapped potential of large pretrained music models for audio-to-audio tasks like source separation.

PaperPDFCode

Code

ethman/tagbox mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Audio GenerationAudio Source SeparationMusic GenerationMusic TaggingTAG

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

ConvolutionDense ConnectionsDilated ConvolutionJukeboxLayer NormalizationPosition-Wise Feed-Forward LayerResidual ConnectionVQ-VAE

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections