Posterior Transition Modeling for Unsupervised Diffusion-Based Speech Enhancement
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912462098399232 |
|---|---|
| author | Sadeghi, Mostafa Ayilo, Jean-Eudes Serizel, Romain Alameda-Pineda, Xavier |
| author_facet | Sadeghi, Mostafa Ayilo, Jean-Eudes Serizel, Romain Alameda-Pineda, Xavier |
| contents | We explore unsupervised speech enhancement using diffusion models as expressive generative priors for clean speech. Existing approaches guide the reverse diffusion process using noisy speech through an approximate, noise-perturbed likelihood score, combined with the unconditional score via a trade-off hyperparameter. In this work, we propose two alternative algorithms that directly model the conditional reverse transition distribution of diffusion states. The first method integrates the diffusion prior with the observation model in a principled way, removing the need for hyperparameter tuning. The second defines a diffusion process over the noisy speech itself, yielding a fully tractable and exact likelihood score. Experiments on the WSJ0-QUT and VoiceBank-DEMAND datasets demonstrate improved enhancement metrics and greater robustness to domain shifts compared to both supervised and unsupervised baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_02391 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Posterior Transition Modeling for Unsupervised Diffusion-Based Speech Enhancement Sadeghi, Mostafa Ayilo, Jean-Eudes Serizel, Romain Alameda-Pineda, Xavier Sound Machine Learning Audio and Speech Processing We explore unsupervised speech enhancement using diffusion models as expressive generative priors for clean speech. Existing approaches guide the reverse diffusion process using noisy speech through an approximate, noise-perturbed likelihood score, combined with the unconditional score via a trade-off hyperparameter. In this work, we propose two alternative algorithms that directly model the conditional reverse transition distribution of diffusion states. The first method integrates the diffusion prior with the observation model in a principled way, removing the need for hyperparameter tuning. The second defines a diffusion process over the noisy speech itself, yielding a fully tractable and exact likelihood score. Experiments on the WSJ0-QUT and VoiceBank-DEMAND datasets demonstrate improved enhancement metrics and greater robustness to domain shifts compared to both supervised and unsupervised baselines. |
| title | Posterior Transition Modeling for Unsupervised Diffusion-Based Speech Enhancement |
| topic | Sound Machine Learning Audio and Speech Processing |
| url | https://arxiv.org/abs/2507.02391 |