A Comparison of Generative and Discriminative Methods for Speech Enhancement: Robustness, Complexity, and Hallucination
On the map: Speech and audio language models → Speech enhancement and noise suppression
Abstract (arXiv)
In this study, we conduct a comprehensive comparative analysis of generative and discriminative deep learning-based speech enhancement methods, specifically in noise reduction tasks. Our investigation focuses on evaluating their effectiveness under high and low signal-to-noise ratio conditions, considering both matched and mismatched training scenarios. We further investigate the impact of training data volume, model convergence speed, and interpret the performance differences in terms of objective results for the considered training paradigms. Additionally, we compare the complexity-performance trade-off and the practical viability of these approaches. To further strengthen the evaluation, we study the hallucination characteristics of generative approaches in terms of word error rate and phoneme similarity. The insights derived from this study provide empirical evidence to assist researchers and practitioners in understanding whether the perceptual gains of different approaches justify their computational cost in practical applications.
Key facts. Cited by 2 papers in the map (data through 19 Oct 2026). Placed in Speech enhancement and noise suppression, one of 11 directions in Speech and audio language models. Builds on 36 of its 44 references that are in the map. First posted on arXiv 1 Jun 2026.
Builds on
36 of 44 references.
Closest ancestors · 19
Same direction, newest first
- 2025Leveraging Discriminative Latent Representations for Conditioning GAN-Based Speech Enhancement28 Aug
- 2025FlowSE: Flow Matching-based Speech Enhancement9 Aug
- 2025DeepFilterGAN: A Full-band Real-time Speech Enhancement System with GAN-based Stochastic Regeneration29 May
- 2025Conditional Latent Diffusion-Based Speech Enhancement Via Dual Context Learning17 Jan
- 2024GAN-Based Speech Enhancement for Low SNR Using Latent Feature Conditioning17 Oct
- 2024GALD-SE: Guided Anisotropic Lightweight Diffusion for Efficient Speech Enhancement23 Sep
- 2024Investigating Training Objectives for Generative Speech Enhancement16 Sep
- 2024Universal Score-based Speech Enhancement with High Content Preservation18 Jun
- 2024EARS: An Anechoic Fullband Speech Dataset Benchmarked for Speech Enhancement and Dereverberation10 Jun
- 2023Ultra Low Complexity Deep Learning Based Noise Suppression13 Dec
- 2023Investigating the Design Space of Diffusion Models for Speech Enhancement7 Dec
- 2023SE-Bridge: Speech Enhancement with Consistent Brownian Bridge23 May
7 more
- 2023Reducing the Prior Mismatch of Stochastic Differential Equations for Diffusion-based Speech Enhancement28 Feb
- 2022Analysing Diffusion-based Generative Approaches versus Discriminative Approaches for Speech Restoration4 Nov
- 2022Speech Enhancement and Dereverberation with Diffusion-based Generative Models11 Aug
- 2022DeepFilterNet2: Towards Real-Time Speech Enhancement on Embedded Devices for Full-Band Audio11 May
- 2022CMGAN: Conformer-based Metric GAN for Speech Enhancement28 Mar
- 2021MetricGAN+: An Improved Version of MetricGAN for Speech Enhancement8 Apr
- 2021Real-time Denoising and Dereverberation with Tiny Recurrent U-Net5 Feb
Other references · 17
- 2020Denoising Diffusion Probabilistic Modelscited by 21,267
- 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,295
- 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,103
- 2023Consistency Modelscited by 1,860
- 2018SDR - half-baked or well done?cited by 626
- 2020DNSMOS: A Non-Intrusive Perceptual Objective Speech Quality metric to evaluate Noise Suppressorscited by 356
- 2020DCCRN: Deep Complex Convolution Recurrent Network for Phase-Aware Speech Enhancementcited by 284
- 2019WHAM!: Extending Speech Separation to Noisy Environmentscited by 246
- 2020The INTERSPEECH 2020 Deep Noise Suppression Challenge: Datasets, Subjective Testing Framework, and Challenge Resultscited by 230
- 2019MetricGAN: Generative Adversarial Networks based Black-box Metric Scores Optimization for Speech Enhancementcited by 161
- 2022FastDiff: A Fast Conditional Diffusion Model for High-Quality Speech Synthesiscited by 144
- 2023FunCodec: A Fundamental, Reproducible and Integrable Open-source Toolkit for Neural Speech Codeccited by 95
5 more
- 2024SCOREQ: Speech Quality Assessment with Contrastive Regressioncited by 49
- 2019Bridging the Gap Between Monaural Speech Enhancement and Recognition with Distortion-Independent Acoustic Modelingcited by 36
- 2023Blind Audio Bandwidth Extension: A Diffusion-Based Zero-Shot Approachcited by 22
- 2024Comparative Analysis Of Discriminative Deep Learning-Based Noise Reduction Methods In Low SNR Scenarioscited by 4
- 2024A Hybrid Approach for Low-Complexity Joint Acoustic Echo and Noise Reductioncited by 4
Built on by
Citing papers per year: 2026: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2023Unifying Robustness and Fidelity: A Comprehensive Study of Pretrained Generative Methods for Speech Enhancement in Adverse Conditions
- 2023On the Behavior of Intrusive and Non-intrusive Speech Enhancement Metrics in Predictive and Generative Settings
- 2025PASE: Leveraging the Phonological Prior of WavLM for Low-Hallucination Generative Speech Enhancement
- 2024URGENT Challenge: Universality, Robustness, and Generalizability For Speech Enhancement
- 2023Unsupervised speech enhancement with deep dynamical generative speech and noise models
- 2024FINALLY: fast and universal speech enhancement with studio-like quality
- 2023Diffusion-based speech enhancement with a weighted generative-supervised learning loss
- 2024Diffusion-based Generative Modeling with Discriminative Guidance for Streamable Speech Enhancement