mlsanityData to 14 Oct 2026

Diff-VS: Efficient Audio-Aware Diffusion U-Net for Vocals Separation

arXiv 2604.01120 ↗First posted 1 Apr 2026eess.AScited by 0

On the map: Diffusion models → Audio, speech and music

Abstract (arXiv)

While diffusion models are best known for their performance in generative tasks, they have also been successfully applied to many other tasks, including audio source separation. However, current generative approaches to music source separation often underperform on standard objective metrics. In this paper, we address this issue by introducing a novel generative vocal separation model based on the Elucidated Diffusion Model (EDM) framework. Our model processes complex short-time Fourier transform spectrograms and employs an improved U-Net architecture based on music-informed design choices. Our approach matches discriminative baselines on objective metrics and achieves perceptual quality comparable to state-of-the-art systems, as assessed by proxy subjective metrics. We hope these results encourage broader exploration of generative methods for music source separation

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Audio, speech and music, one of 14 directions in Diffusion models. Builds on 28 of its 31 references that are in the map. First posted on arXiv 1 Apr 2026.

Builds on

28 of 31 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025Generating Separated Singing Vocals Using a Diffusion Model Conditioned on Music Mixtures26 Nov
  2. 2025Efficient and Fast Generative-Based Singing Voice Separation using a Latent Diffusion Model25 Nov
  3. 2024Simultaneous Music Separation and Generation Using Multi-Track Latent Diffusion Models18 Sep
  4. 2024Diff-A-Riff: Musical Accompaniment Co-creation via Latent Diffusion Models12 Jun
  5. 2023Multi-Source Diffusion Models for Simultaneous Music Generation and Separation4 Feb
  6. 2022Speech Enhancement and Dereverberation with Diffusion-based Generative Models11 Aug
  7. 2022Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain31 Mar
  8. 2022Conditional Diffusion Probabilistic Model for Speech Enhancement10 Feb

Other references · 20

  1. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  2. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 3,045
  3. 2022High Fidelity Neural Audio Compressioncited by 687
  4. 2024Stable Audio Opencited by 197
  5. 2018Wave-U-Net: A Multi-Scale Neural Network for End-to-End Audio Source Separationcited by 125
  6. 2021Hybrid Spectrogram and Waveform Source Separationcited by 91
  7. 2022Universal Speech Enhancement with Score-based Diffusioncited by 75
  8. 2022Music Source Separation with Band-split RNNcited by 70
  9. 2018The 2018 Signal Separation Evaluation Campaigncited by 41
  10. 2023Moisesdb: A dataset for source separation beyond 4-stemscited by 39
  11. 2023Music Source Separation with Band-Split RoPE Transformercited by 37
  12. 2024Investigating Training Objectives for Generative Speech Enhancementcited by 20
8 more
  1. 2020A consolidated view of loss functions for supervised deep learning-based speech enhancementcited by 19
  2. 2024ETTA: Elucidating the Design Space of Text-to-Audio Modelscited by 17
  3. 2017Adversarial Semi-Supervised Audio Source Separation applied to Singing Voice Extractioncited by 11
  4. 2024SCNet: Sparse Compression Network for Music Source Separationcited by 10
  5. 2024Mel-RoFormer for Vocal Separation and Vocal Melody Transcriptioncited by 8
  6. 2023Sound Demixing Challenge 2023 Music Demixing Track Technical Report: TFC-TDF-UNet v3cited by 4
  7. 2025Towards Reliable Objective Evaluation Metrics for Generative Singing Voice Separation Modelscited by 2
  8. 2025Moises-Light: Resource-efficient Band-split U-Net For Music Source Separationcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2023EDMSound: Spectrogram Based Diffusion Models for Efficient and High-Quality Audio Synthesis
  2. 2024Improving Music Source Separation with Diffusion and Consistency Refinement
  3. 2025High-Quality Sound Separation Across Diverse Categories via Visually-Guided Generative Modeling
  4. 2025User-guided Generative Source Separation
  5. 2023ArchiSound: Audio Generation with Diffusion
  6. 2025Audio Generation Through Score-Based Generative Modeling: Design Principles and Implementation
  7. 2023High-Quality Visually-Guided Sound Separation from Diverse Categories
  8. 2025Score Distillation Sampling for Audio: Source Separation, Synthesis, and Beyond