The Sound Demixing Challenge 2023 $\unicode{x2013}$ Music Demixing Track
arXiv:2308.06979 · doi:10.5334/tismir.171
Abstract
This paper summarizes the music demixing (MDX) track of the Sound Demixing Challenge (SDX'23). We provide a summary of the challenge setup and introduce the task of robust music source separation (MSS), i.e., training MSS models in the presence of errors in the training data. We propose a formalization of the errors that can occur in the design of a training dataset for MSS systems and introduce two new datasets that simulate such errors: SDXDB23_LabelNoise and SDXDB23_Bleeding. We describe the methods that achieved the highest scores in the competition. Moreover, we present a direct comparison with the previous edition of the challenge (the Music Demixing Challenge 2021): the best performing system achieved an improvement of over 1.6dB in signal-to-distortion ratio over the winner of the previous competition, when evaluated on MDXDB21. Besides relying on the signal-to-distortion ratio as objective metric, we also performed a listening test with renowned producers and musicians to study the perceptual quality of the systems and report here the results. Finally, we provide our insights into the organization of the competition and our prospects for future editions.
arXiv commentPublished in Transactions of the International Society for Music Information Retrieval (https://transactions.ismir.net/articles/10.5334/tismir.171)
References in corpus (19)
- Distilling the Knowledge in a Neural Network
- Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results
- DivideMix: Learning with Noisy Labels as Semi-supervised Learning
- Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels
- Unsupervised Sound Separation Using Mixture Invariant Training
- Objective Measures of Perceptual Audio Quality Reviewed: An Evaluation of Their Application Domain Dependence
- Hybrid Spectrogram and Waveform Source Separation
- KUIELab-MDX-Net: A Two-Stream Neural Network for Music Demixing
- Learning with Instance-Dependent Label Noise: A Sample Sieve Approach
- Investigating U-Nets with various Intermediate Blocks for Spectrogram-based Singing Voice Separation
- Robust Subspace Recovery Layer for Unsupervised Anomaly Detection
- Moisesdb: A dataset for source separation beyond 4-stems
- Improved Natural Language Generation via Loss Truncation
- Leveraging Hierarchical Structures for Few-Shot Musical Instrument Recognition
- Benchmarks and leaderboards for sound demixing tasks
- Hybrid Transformers for Music Source Separation
- ProMix: Combating Label Noise via Maximizing Clean Sample Utility
- Self-refining of Pseudo Labels for Music Source Separation with Noisy Labeled Data
- Semi-Supervised Singing Voice Separation with Noisy Self-Training