Instrument Separation of Symbolic Music by Explicitly Guided Diffusion Model

TL;DR

This paper introduces a diffusion model approach for instrument separation in symbolic music, explicitly guiding the process to maintain consistency between mixtures and individual instrument tracks, achieving high-fidelity results.

Contribution

The paper presents a novel diffusion-based method with explicit guidance for instrument separation in symbolic music, improving fidelity and consistency.

Findings

High-fidelity sample generation for multitrack symbolic music

Effective preservation of mixture-instrument consistency

Demonstrated creativity in separated instrument outputs

Abstract

Similar to colorization in computer vision, instrument separation is to assign instrument labels (e.g. piano, guitar...) to notes from unlabeled mixtures which contain only performance information. To address the problem, we adopt diffusion models and explicitly guide them to preserve consistency between mixtures and music. The quantitative results show that our proposed model can generate high-fidelity samples for multitrack symbolic music with creativity.