mlsanityData to 7 Oct 2026

Music Source Separation with Band-split RNN

arXiv 2209.15174 ↗First posted 30 Sep 2022eess.AS cs.LG cs.SDcited by 69

Not on the map. Why?

Abstract (arXiv)

The performance of music source separation (MSS) models has been greatly improved in recent years thanks to the development of novel neural network architectures and training pipelines. However, recent model designs for MSS were mainly motivated by other audio processing tasks or other research fields, while the intrinsic characteristics and patterns of the music signals were not fully discovered. In this paper, we propose band-split RNN (BSRNN), a frequency-domain model that explictly splits the spectrogram of the mixture into subbands and perform interleaved band-level and sequence-level modeling. The choices of the bandwidths of the subbands can be determined by a priori knowledge or expert knowledge on the characteristics of the target source in order to optimize the performance on a certain type of target musical instrument. To better make use of unlabeled data, we also describe a semi-supervised model finetuning pipeline that can further improve the performance of the model. Experiment results show that BSRNN trained only on MUSDB18-HQ dataset significantly outperforms several top-ranking models in Music Demixing (MDX) Challenge 2021, and the semi-supervised finetuning stage further improves the performance on all four instrument tracks.

Key facts. Cited by 69 papers in the map (data through 7 Oct 2026). Builds on 53 of its 81 references that are in the map. First posted on arXiv 30 Sep 2022.

Builds on

53 of 81 references.

References · 53

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  4. 2016Densely Connected Convolutional Networkscited by 6,691
  5. 2016Layer Normalizationcited by 5,579
  6. 2020Knowledge Distillation: A Surveycited by 1,578
  7. 2019Self-training with Noisy Student improves ImageNet classificationcited by 1,304
  8. 2016Stacked Hourglass Networks for Human Pose Estimationcited by 913
  9. 2016Language Modeling with Gated Convolutional Networkscited by 688
  10. 2019Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learningcited by 507
  11. 2018Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separationcited by 400
  12. 2020Meta Pseudo Labelscited by 372
38 more
  1. 2015Deep clustering: Discriminative embeddings for segmentation and separationcited by 211
  2. 2017Teacher-Student Curriculum Learningcited by 197
  3. 2020Attention is All You Need in Speech Separationcited by 184
  4. 2020Real Time Speech Enhancement in the Waveform Domaincited by 183
  5. 2019Dual-path RNN: efficient long sequence modeling for time-domain single-channel speech separationcited by 182
  6. 2020Knowledge Distillation Meets Self-Supervisioncited by 142
  7. 2018Wave-U-Net: A Multi-Scale Neural Network for End-to-End Audio Source Separationcited by 125
  8. 2019Music Source Separation in the Waveform Domaincited by 98
  9. 2021Hybrid Spectrogram and Waveform Source Separationcited by 90
  10. 2020Dual-Path Transformer Network: Direct Context-Aware Modeling for End-to-End Monaural Speech Separationcited by 67
  11. 2020FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancementcited by 66
  12. 2019Demucs: Deep Extractor for Music Sources with extra unlabeled data remixedcited by 53
  13. 2021VoiceFixer: Toward General Speech Restoration with Neural Vocodercited by 50
  14. 2021Decoupling Magnitude and Phase Estimation with Deep ResUNet for Music Source Separationcited by 48
  15. 2018The 2018 Signal Separation Evaluation Campaigncited by 41
  16. 2021Music Demixing Challenge 2021cited by 38
  17. 2021DPT-FSNet: Dual-path Transformer Based Full-band and Sub-band Fusion Network for Speech Enhancementcited by 33
  18. 2017Large-Scale Domain Adaptation via Teacher-Student Learningcited by 28
  19. 2022FullSubNet+: Channel Attention FullSubNet with Complex Spectrograms for Speech Enhancementcited by 28
  20. 2020D3Net: Densely connected multidilated DenseNet for music source separationcited by 24
  21. 2021KUIELab-MDX-Net: A Two-Stream Neural Network for Music Demixingcited by 24
  22. 2018Improved Speech Enhancement with the Wave-U-Netcited by 23
  23. 2020The Cone of Silence: Speech Separation by Localizationcited by 21
  24. 2015Joint Optimization of Masks and Deep Recurrent Neural Networks for Monaural Source Separationcited by 20
  25. 2022RemixIT: Continual self-training of speech enhancement models via bootstrapped remixingcited by 20
  26. 2018MMDenseLSTM: An efficient combination of convolutional and recurrent neural networks for audio source separationcited by 19
  27. 2020Meta-learning Extractors for Music Source Separationcited by 18
  28. 2021A Unified Model for Zero-shot Music Source Separation, Transcription and Synthesiscited by 18
  29. 2021S-DCCRN: Super Wide Band DCCRN with learnable complex feature for speech enhancementcited by 12
  30. 2022On the Use of Deep Mask Estimation Module for Neural Source Separation Systemscited by 11
  31. 2016Deep Clustering and Conventional Networks for Music Separation: Stronger Togethercited by 10
  32. 2021CWS-PResUNet: Music Source Separation with Channel-wise Subband Phase-aware ResUNetcited by 9
  33. 2020Ultra-Lightweight Speech Separation via Group Communicationcited by 8
  34. 2021Teacher-Student MixIT for Unsupervised and Semi-supervised Speech Separationcited by 8
  35. 2020Group Communication with Context Codec for Lightweight Source Separationcited by 7
  36. 2021Continual self-training with bootstrapped remixing for speech enhancementcited by 6
  37. 2021TPARN: Triple-path Attentive Recurrent Network for Time-domain Multichannel Speech Enhancementcited by 6
  38. 2021Semi-Supervised Singing Voice Separation with Noisy Self-Trainingcited by 5

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 12; 2024: 19; 2025: 22; 2026: 16.

69 papers.

Sort:

By area:

  1. 2025Kimi-Audio Technical Reportcited by 210
  2. 2025SepPrune: Structured Pruning for Efficient Deep Speech Separationcited by 31
  3. 2023ReZero: Region-customizable Sound Extractioncited by 10
  4. 2024TIGER: Time-frequency Interleaved Gain Extraction and Reconstruction for Efficient Speech Separationcited by 10
  5. 2023Subnetwork-to-go: Elastic Neural Network with Dynamic Training and Customizable Inferencecited by 9
  6. 2023A Generalized Bandsplit Neural Network for Cinematic Audio Source Separationcited by 8
  7. 2024Apollo: Band-sequence Modeling for High-Quality Audio Restorationcited by 8
  8. 2024OmniSep: Unified Omni-Modality Sound Separation with Query-Mixupcited by 8
  9. 2024A Stem-Agnostic Single-Decoder System for Music Source Separation Beyond Four Stemscited by 7
  10. 2023Benchmarks and leaderboards for sound demixing taskscited by 6
  11. 2023GASS: Generalizing Audio Source Separation with Large-scale Datacited by 6
  12. 2024Listen, Chat, and Remix: Text-Guided Soundscape Remixing for Enhanced Auditory Experiencecited by 6
38 more
  1. 2024Gull: A Generative Multifunctional Audio Codeccited by 5
  2. 2025Music Source Restorationcited by 5
  3. 2023Sound Demixing Challenge 2023 Music Demixing Track Technical Report: TFC-TDF-UNet v3cited by 4
  4. 2024SynthSOD: Developing an Heterogeneous Dataset for Orchestra Music Source Separationcited by 4
  5. 2024SonicSim: A customizable simulation platform for speech processing in moving sound source scenarioscited by 4
  6. 2025MGE-LDM: Joint Latent Diffusion for Simultaneous Music Generation and Source Extractioncited by 4
  7. 2023Multi-scale clustering and source separation of InSight mission seismic datacited by 3
  8. 2024Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNetcited by 3
  9. 2024DeFT-Mamba: Universal Multichannel Sound Separation and Polyphonic Audio Classificationcited by 3
  10. 2025A Survey of Deep Learning for Complex Speech Spectrogramscited by 3
  11. 2026Habibi: Laying the Open-Source Foundation of Unified-Dialectal Arabic Speech Synthesiscited by 3
  12. 2023RTFS-Net: Recurrent Time-Frequency Modelling for Efficient Audio-Visual Speech Separationcited by 2
  13. 2024A Two-Stage Band-Split Mamba-2 Network For Music Separationcited by 2
  14. 2025Score-informed Music Source Separation: Improving Synthetic-to-real Generalization in Classical Musiccited by 2
  15. 2025MACS: Multi-source Audio-to-image Generation with Contextual Significance and Semantic Alignmentcited by 2
  16. 2025User-guided Generative Source Separationcited by 2
  17. 2025Efficient and Fast Generative-Based Singing Voice Separation using a Latent Diffusion Modelcited by 2
  18. 2023Self-refining of Pseudo Labels for Music Source Separation with Noisy Labeled Datacited by 1
  19. 2023MBTFNet: Multi-Band Temporal-Frequency Neural Network For Singing Voice Enhancementcited by 1
  20. 2024Remixing Music for Hearing Aids Using Ensemble of Fine-Tuned Source Separatorscited by 1
  21. 2024Prevailing Research Areas for Music AI in the Era of Foundation Modelscited by 1
  22. 2024An Ensemble Approach to Music Source Separation: A Comparative Analysis of Conventional and Hierarchical Stem Separationcited by 1
  23. 2024Improving Music Source Separation with Diffusion and Consistency Refinementcited by 1
  24. 2025Unleashing the Power of Natural Audio Featuring Multiple Sound Sourcescited by 1
  25. 2025Training-Free Multi-Step Audio Source Separationcited by 1
  26. 2025Target speaker anonymization in multi-speaker recordingscited by 1
  27. 2025MARS-Sep: Multimodal-Aligned Reinforced Sound Separationcited by 1
  28. 2025The Spheres Dataset: Multitrack Orchestral Recordings for Music Source Separation and Information Retrievalcited by 1
  29. 2026Summary of The Inaugural Music Source Restoration Challengecited by 1
  30. 2026Difference-Driven Gating: Adaptive Feature Fusion for U-Net Decodercited by 1
  31. 2023Pac-HuBERT: Self-Supervised Music Source Separation via Primitive Auditory Clustering and Hidden-Unit BERTcited by 0
  32. 2023Adversarial Generative NMF for Single Channel Source Separationcited by 0
  33. 2024Resource-constrained stereo singing voice cancellationcited by 0
  34. 2024Improving Real-Time Music Accompaniment Separation with MMDenseNetcited by 0
  35. 2024Activity-Guided Industrial Anomalous Sound Detection against Interferencescited by 0
  36. 2024S2Cap: A Benchmark and a Baseline for Singing Style Captioningcited by 0
  37. 2025MAJL: A Model-Agnostic Joint Learning Framework for Music Source Separation and Pitch Estimationcited by 0
  38. 2025Solving Copyright Infringement on Short Video Platforms: Novel Datasets and an Audio Restoration Deep Learning Pipelinecited by 0

50 most cited of 69 All 69 (JSON)