SynthSOD: Developing an Heterogeneous Dataset for Orchestra Music Source Separation

TL;DR

This paper introduces SynthSOD, a new multitrack dataset for orchestra music source separation, addressing the lack of comprehensive datasets for extracting similar-sounding orchestral sources.

Contribution

SynthSOD is a novel, realistic, and heterogeneous dataset created using simulation techniques, enabling improved training for orchestra source separation models.

Findings

Baseline model trained on SynthSOD performs well on synthetic data.

Model generalizes to real-world orchestral recordings.

SynthSOD enhances the development of source separation techniques for orchestral music.

Abstract

Recent advancements in music source separation have significantly progressed, particularly in isolating vocals, drums, and bass elements from mixed tracks. These developments owe much to the creation and use of large-scale, multitrack datasets dedicated to these specific components. However, the challenge of extracting similarly sounding sources from orchestra recordings has not been extensively explored, largely due to a scarcity of comprehensive and clean (i.e bleed-free) multitrack datasets. In this paper, we introduce a novel multitrack dataset called SynthSOD, developed using a set of simulation techniques to create a realistic (i.e. using high-quality soundfonts), musically motivated, and heterogeneous training set comprising different dynamics, natural tempo changes, styles, and conditions. Moreover, we demonstrate the application of a widely used baseline music separation model trained on our synthesized dataset w.r.t to the well-known EnsembleSet, and evaluate its performance under both synthetic and real-world conditions.