Voice and accompaniment separation in music using self-attention convolutional neural network

TL;DR

This paper introduces a novel self-attention convolutional neural network for music source separation, specifically isolating vocals from accompaniment, leveraging musical repetitions for improved performance.

Contribution

The work presents a new self-attention integrated CNN architecture that exploits long-term dependencies in music for enhanced source separation.

Findings

19.5% relative improvement in vocals separation SDR

Outperforms state-of-the-art systems like MMDenseNet and MMDenseLSTM

Utilizes musical repetitions for better reconstruction

Abstract

Music source separation has been a popular topic in signal processing for decades, not only because of its technical difficulty, but also due to its importance to many commercial applications, such as automatic karoake and remixing. In this work, we propose a novel self-attention network to separate voice and accompaniment in music. First, a convolutional neural network (CNN) with densely-connected CNN blocks is built as our base network. We then insert self-attention subnets at different levels of the base CNN to make use of the long-term intra-dependency of music, i.e., repetition. Within self-attention subnets, repetitions of the same musical patterns inform reconstruction of other repetitions, for better source separation performance. Results show the proposed method leads to 19.5% relative improvement in vocals separation in terms of SDR. We compare our methods with state-of-the-art systems i.e. MMDenseNet and MMDenseLSTM.