# Improved singing voice separation with chromagram-based pitch-aware remixing

arXiv: https://arxiv.org/abs/2203.15092 · HTML: https://mlsanity.com/paper/2203.15092/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 19 references that are in the map.
- First posted on arXiv 28 Mar 2022.

## Builds on (most cited first, 12 of 12)

- [Self-training with Noisy Student improves ImageNet classification](https://mlsanity.com/paper/1911.04252/) (2019)
- [Demucs: Deep Extractor for Music Sources with extra unlabeled data remixed](https://mlsanity.com/paper/1909.01174/) (2019)
- [Decoupling Magnitude and Phase Estimation with Deep ResUNet for Music Source Separation](https://mlsanity.com/paper/2109.05418/) (2021)
- [The 2018 Signal Separation Evaluation Campaign](https://mlsanity.com/paper/1804.06267/) (2018)
- [PoCoNet: Better Speech Enhancement with Frequency-Positional Embeddings, Semi-Supervised Conversational Data, and Biased Loss](https://mlsanity.com/paper/2008.04470/) (2020)
- [Joint Optimization of Masks and Deep Recurrent Neural Networks for Monaural Source Separation](https://mlsanity.com/paper/1502.04149/) (2015)
- [MMDenseLSTM: An efficient combination of convolutional and recurrent neural networks for audio source separation](https://mlsanity.com/paper/1805.02410/) (2018)
- [Mixing-Specific Data Augmentation Techniques for Improved Blind Violin/Piano Source Separation](https://mlsanity.com/paper/2008.02480/) (2020)
- [Singing voice separation: a study on training data](https://mlsanity.com/paper/1906.02618/) (2019)
- [Semi-Supervised Singing Voice Separation with Noisy Self-Training](https://mlsanity.com/paper/2102.07961/) (2021)
- [Improving singing voice separation using Deep U-Net and Wave-U-Net with data augmentation](https://mlsanity.com/paper/1903.01415/) (2019)
- [CatNet: music source separation system with mix-audio augmentation](https://mlsanity.com/paper/2102.09966/) (2021)

## Built on by (most cited first, 2 of 2)

- [MBTFNet: Multi-Band Temporal-Frequency Neural Network For Singing Voice Enhancement](https://mlsanity.com/paper/2310.04369/) (2023)
- [MARS-Sep: Multimodal-Aligned Reinforced Sound Separation](https://mlsanity.com/paper/2510.10509/) (2025)

## Abstract (arXiv)

Singing voice separation aims to separate music into vocals and accompaniment components. One of the major constraints for the task is the limited amount of training data with separated vocals. Data augmentation techniques such as random source mixing have been shown to make better use of existing data and mildly improve model performance. We propose a novel data augmentation technique, chromagram-based pitch-aware remixing, where music segments with high pitch alignment are mixed. By performing controlled experiments in both supervised and semi-supervised settings, we demonstrate that training models with pitch-aware remixing significantly improves the test signal-to-distortion ratio (SDR)