# Music Source Restoration with Ensemble Separation and Targeted Reconstruction

arXiv: https://arxiv.org/abs/2603.16926 · HTML: https://mlsanity.com/paper/2603.16926/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 6 of its 7 references that are in the map.
- First posted on arXiv 13 Mar 2026.

## Builds on (most cited first, 6 of 6)

- [Fr'echet Audio Distance: A Metric for Evaluating Music Enhancement Algorithms](https://mlsanity.com/paper/1812.08466/) (2018)
- [Music Source Separation with Band-Split RoPE Transformer](https://mlsanity.com/paper/2309.02612/) (2023)
- [Apollo: Band-sequence Modeling for High-Quality Audio Restoration](https://mlsanity.com/paper/2409.08514/) (2024)
- [SonicMaster: Towards Controllable All-in-One Music Restoration and Mastering](https://mlsanity.com/paper/2508.03448/) (2025)
- [Zimtohrli: An Efficient Psychoacoustic Audio Similarity Metric](https://mlsanity.com/paper/2509.26133/) (2025)
- [MSRBench: A Benchmarking Dataset for Music Source Restoration](https://mlsanity.com/paper/2510.10995/) (2025)

## Abstract (arXiv)

The Inaugural Music Source Restoration (MSR) Challenge targets the recovery of original, unprocessed stems from fully mixed and mastered music. Unlike conventional music source separation, MSR requires reversing complex production processes such as equalization, compression, reverberation, and other real-world degradations. To address MSR, we propose a two-stage system. First, an ensemble of pre-trained separation models produces preliminary source estimates. Then a set of pre-trained BSRNN-based restoration models performs targeted reconstruction to refine these estimates. On the official MSR benchmark, our system surpasses the baselines on all metrics, ranking second among all submissions. The code is available at https://github.com/xinghour/Music-source-restoration-CUPAudioGroup