mlsanityData to 7 Oct 2026

MetricGAN-U: Unsupervised speech enhancement/ dereverberation based only on noisy/ reverberated speech

arXiv 2110.05866 ↗First posted 12 Oct 2021cs.SD cs.CL eess.AScited by 18

Not on the map. Why?

Abstract (arXiv)

Most of the deep learning-based speech enhancement models are learned in a supervised manner, which implies that pairs of noisy and clean speech are required during training. Consequently, several noisy speeches recorded in daily life cannot be used to train the model. Although certain unsupervised learning frameworks have also been proposed to solve the pair constraint, they still require clean speech or noise for training. Therefore, in this paper, we propose MetricGAN-U, which stands for MetricGAN-unsupervised, to further release the constraint from conventional unsupervised learning. In MetricGAN-U, only noisy speech is required to train the model by optimizing non-intrusive speech quality metrics. The experimental results verified that MetricGAN-U outperforms baselines in both objective and subjective metrics.

Key facts. Cited by 18 papers in the map (data through 7 Oct 2026). Builds on 15 of its 31 references that are in the map. First posted on arXiv 12 Oct 2021.

Builds on

15 of 31 references.

References · 15

  1. 2021SpeechBrain: A General-Purpose Speech Toolkitcited by 404
  2. 2020DNSMOS: A Non-Intrusive Perceptual Objective Speech Quality metric to evaluate Noise Suppressorscited by 235
  3. 2017Supervised Speech Separation Based on Deep Learning: An Overviewcited by 213
  4. 2017SEGAN: Speech Enhancement Generative Adversarial Networkcited by 165
  5. 2021DNSMOS P.835: A Non-Intrusive Perceptual Objective Speech Quality Metric to Evaluate Noise Suppressorscited by 165
  6. 2021MetricGAN+: An Improved Version of MetricGAN for Speech Enhancementcited by 114
  7. 2019MetricGAN: Generative Adversarial Networks based Black-box Metric Scores Optimization for Speech Enhancementcited by 104
  8. 2021Unsupervised Speech Enhancement using Dynamical Variational Auto-Encoderscited by 28
  9. 2020An Open source Implementation of ITU-T Recommendation P.808 with Validationcited by 27
  10. 2021Noisy-target Training: A Training Strategy for DNN-based Speech Enhancement without Clean Speechcited by 15
  11. 2021CycleGAN-based Non-parallel Speech Enhancement with an Adaptive Attention-in-attention Mechanismcited by 7
  12. 2020Noise Tokens: Learning Neural Noise Templates for Environment-Aware Speech Enhancementcited by 5
3 more
  1. 2020iMetricGAN: Intelligibility Enhancement for Speech-in-Noise using Generative Adversarial Network-based Metric Learningcited by 4
  2. 2021Deep Noise Suppression With Non-Intrusive PESQNet Supervision Enabling the Use of Real Training Datacited by 3
  3. 2021TENET: A Time-reversal Enhancement Network for Noise-robust ASRcited by 3

Built on by

Cited by year

Citing papers per year: 2021: 1; 2022: 3; 2023: 6; 2024: 3; 2025: 5.

18 papers.

Sort:

By area:

  1. 2022Speech Enhancement and Dereverberation with Diffusion-based Generative Modelscited by 195
  2. 2021Deep Learning-based Non-Intrusive Multi-Objective Speech Assessment Model with Cross-Domain Featurescited by 32
  3. 2023Self-Supervised Visual Acoustic Matchingcited by 14
  4. 2023Perceive and predict: self-supervised speech representation based loss functions for speech enhancementcited by 8
  5. 2024Self-Supervised Speech Quality Estimation and Enhancement Using Only Clean Speechcited by 7
  6. 2024SKILL: Similarity-aware Knowledge distILLation for Speech Self-Supervised Learningcited by 4
  7. 2022Deep neural network techniques for monaural speech enhancement: state of the art analysiscited by 3
  8. 2023Non Intrusive Intelligibility Predictor for Hearing Impaired Individuals using Self Supervised Speech Representationscited by 3
  9. 2024Gaussian Flow Bridges for Audio Domain Transfer with Unpaired Datacited by 3
  10. 2025A Hybrid Model for Weakly-Supervised Speech Dereverberationcited by 2
  11. 2022A Training and Inference Strategy Using Noisy and Enhanced Speech as Target for Speech Enhancement without Clean Speechcited by 1
  12. 2023Speech Modeling with a Hierarchical Transformer Dynamical VAEcited by 1
6 more
  1. 2023Unsupervised speech enhancement with deep dynamical generative speech and noise modelscited by 1
  2. 2025Autoregressive Speech Enhancement via Acoustic Tokenscited by 1
  3. 2025WhiSQA: Non-Intrusive Speech Quality Prediction Using Whisper Encoder Featurescited by 1
  4. 2023Integrating Uncertainty into Neural Network-based Speech Enhancementcited by 0
  5. 2025U-DREAM: Unsupervised Dereverberation guided by a Reverberation Modelcited by 0
  6. 2025Unsupervised Speech Enhancement using Data-defined Priorscited by 0