Unsupervised Audio Source Separation using Generative Priors

TL;DR

This paper introduces an unsupervised audio source separation method that leverages generative priors trained on individual sources, using spectral domain loss functions and gradient descent for effective source recovery without labeled data.

Contribution

The paper presents a novel unsupervised approach for audio source separation using generative priors and spectral domain optimization, reducing dependence on labeled training data.

Findings

Outperforms classical unsupervised baselines.

Effective in separating spoken digits and instrument sounds.

Uses spectral domain loss for high-quality source estimates.

Abstract

State-of-the-art under-determined audio source separation systems rely on supervised end-end training of carefully tailored neural network architectures operating either in the time or the spectral domain. However, these methods are severely challenged in terms of requiring access to expensive source level labeled data and being specific to a given set of sources and the mixing process, which demands complete re-training when those assumptions change. This strongly emphasizes the need for unsupervised methods that can leverage the recent advances in data-driven modeling, and compensate for the lack of labeled data through meaningful priors. To this end, we propose a novel approach for audio source separation based on generative priors trained on individual sources. Through the use of projected gradient descent optimization, our approach simultaneously searches in the source-specific latent spaces to effectively recover the constituent sources. Though the generative priors can be defined in the time domain directly, e.g. WaveGAN, we find that using spectral domain loss functions for our optimization leads to good-quality source estimates. Our empirical studies on standard spoken digit and instrument datasets clearly demonstrate the effectiveness of our approach over classical as well as state-of-the-art unsupervised baselines.