Music Source Separation with Deep Equilibrium Models

TL;DR

This paper introduces DEQ-UMX, a deep equilibrium model-based approach for music source separation that outperforms the original UMX while reducing model size, demonstrating the effectiveness of implicit architectures in audio tasks.

Contribution

The paper proposes a novel DEQ-based architecture for music source separation, adapting deep equilibrium models to improve performance and reduce parameters.

Findings

DEQ-UMX outperforms original UMX in separation quality.

DEQ-UMX reduces model parameters by 30%.

Implicit architectures are effective for audio signal processing.

Abstract

While deep neural network-based music source separation (MSS) is very effective and achieves high performance, its model size is often a problem for practical deployment. Deep implicit architectures such as deep equilibrium models (DEQ) were recently proposed, which can achieve higher performance than their explicit counterparts with limited depth while keeping the number of parameters small. This makes DEQ also attractive for MSS, especially as it was originally applied to sequential modeling tasks in natural language processing and thus should in principle be also suited for MSS. However, an investigation of a good architecture and training scheme for MSS with DEQ is needed as the characteristics of acoustic signals are different from those of natural language data. Hence, in this paper we propose an architecture and training scheme for MSS with DEQ. Starting with the architecture of Open-Unmix (UMX), we replace its sequence model with DEQ. We refer to our proposed method as DEQ-based UMX (DEQ-UMX). Experimental results show that DEQ-UMX performs better than the original UMX while reducing its number of parameters by 30%.