mlsanityData to 19 Oct 2026

DiffVQE: Hybrid Diffusion Voice Quality Enhancement Under Acoustic Echo and Noise

arXiv 2605.08189 ↗First posted 5 May 2026eess.AScited by 0

On the map: Speech and audio language models → Speech enhancement and noise suppression

Abstract (arXiv)

Acoustic echo and background noise pose challenges on speech enhancement in hands-free systems and speakerphones. Discriminatively trained end-to-end methods represent a powerful solution for joint acoustic echo control (AEC) and denoising. However, with the advent of generative methods, diffusion-based approaches have seen remarkable performance in speech enhancement tasks. In this work, to the best of our knowledge, we provide the first (still non-causal) diffusion-based AEC model (DiffVQE) that is reproducible in terms of topology, training data, and training framework. So far, without employing diffusion, Microsoft's discriminative DeepVQE model has been shown to excel any of the ICASSP 2023 AEC Challenge entries achieving remarkable performance. Using data from the Interspeech 2025 URGENT Challenge for a diverse, high-quality training dataset, our DiffVQE excels DeepVQE both in echo and noise control performance, as well as in computational complexity and model size.

Key facts. No citing papers in the map yet (data through 19 Oct 2026). Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models. Builds on 28 of its 44 references that are in the map. First posted on arXiv 5 May 2026.

Builds on

28 of 44 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025P.808 Multilingual Speech Enhancement Testing: Approach and Results of URGENT 2025 Challenge15 Jul
  2. 2025Interspeech 2025 URGENT Speech Enhancement Challenge29 May
  3. 2024Universal Score-based Speech Enhancement with High Content Preservation18 Jun
  4. 2024URGENT Challenge: Universality, Robustness, and Generalizability For Speech Enhancement7 Jun
  5. 2024Efficient High-Performance Bark-Scale Neural Network for Residual Echo and Noise Suppression8 Apr
  6. 2023ICASSP 2023 Acoustic Echo Cancellation Challenge22 Sep
  7. 2022StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberation22 Dec
  8. 2022Speech Enhancement and Dereverberation with Diffusion-based Generative Models11 Aug
  9. 2022Speech Enhancement with Score-Based Generative Models in the Complex STFT Domain31 Mar

Other references · 19

  1. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,295
  2. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 3,082
  3. 2019Common Voice: A Massively-Multilingual Speech Corpuscited by 1,596
  4. 2016Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Networkcited by 1,215
  5. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 552
  6. 2020DNSMOS: A Non-Intrusive Perceptual Objective Speech Quality metric to evaluate Noise Suppressorscited by 356
  7. 2021NISQA: A Deep CNN-Self-Attention Model for Multidimensional Speech Quality Prediction with Crowdsourced Datasetscited by 270
  8. 2017Pyroomacoustics: A Python package for audio room simulations and array processing algorithmscited by 175
  9. 2020Adversarial score matching and improved sampling for image generationcited by 99
  10. 2023TorchAudio-Squim: Reference-less Speech Quality and Intelligibility measures in TorchAudiocited by 82
  11. 2020A consolidated view of loss functions for supervised deep learning-based speech enhancementcited by 43
  12. 2024ICASSP 2024 Speech Signal Improvement Challengecited by 30
7 more
  1. 2021AECMOS: A speech quality assessment metric for echo impairmentcited by 15
  2. 2023DeepVQE: Real Time Deep Voice Quality Enhancement for Joint Acoustic Echo Cancellation, Noise Suppression and Dereverberationcited by 12
  3. 2022Low-Complexity Acoustic Echo Cancellation with Neural Kalman Filteringcited by 5
  4. 2024A Hybrid Approach for Low-Complexity Joint Acoustic Echo and Noise Reductioncited by 4
  5. 2023End-To-End Deep Learning-based Adaptation Control for Linear Acoustic Echo Cancellationcited by 2
  6. 2025Neural Kalman Filters for Acoustic Echo Cancellationcited by 2
  7. 2025EchoFree: Towards Ultra Lightweight and Efficient Neural Acoustic Echo Cancellationcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026DiffVQE2: An Efficient Low-delay Diffusion Model for Acoustic Echo and Noise Control
  2. 2024Extract and Diffuse: Latent Integration for Improved Diffusion-based Speech and Vocal Enhancement
  3. 2024FADI-AEC: Fast Score Based Diffusion Model Guided by Far-end Signal for Acoustic Echo Cancellation
  4. 2024Pre-training Feature Guided Diffusion Model for Speech Enhancement
  5. 2023Diffusion-Based Speech Enhancement in Matched and Mismatched Conditions Using a Heun-Based Sampler
  6. 2023NADiffuSE: Noise-aware Diffusion-based Model for Speech Enhancement
  7. 2025DiTSE: High-Fidelity Generative Speech Enhancement via Latent Diffusion Transformers
  8. 2023A Survey on Audio Diffusion Models: Text To Speech Synthesis and Enhancement in Generative AI