Geneses: Unified Generative Speech Enhancement and Separation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917223716618240 |
|---|---|
| author | Asai, Kohei Nakata, Wataru Saito, Yuki Saruwatari, Hiroshi |
| author_facet | Asai, Kohei Nakata, Wataru Saito, Yuki Saruwatari, Hiroshi |
| contents | Real-world audio recordings often contain multiple speakers and various degradations, which limit both the quantity and quality of speech data available for building state-of-the-art speech processing models. Although end-to-end approaches that concatenate speech enhancement (SE) and speech separation (SS) to obtain a clean speech signal for each speaker are promising, conventional SE-SS methods suffer from complex degradations beyond additive noise. To this end, we propose \textbf{Geneses}, a generative framework to achieve unified, high-quality SE--SS. Our Geneses leverages latent flow matching to estimate each speaker's clean speech features using multi-modal diffusion Transformer conditioned on self-supervised learning representation from noisy mixture. We conduct experimental evaluation using two-speaker mixtures from LibriTTS-R under two conditions: additive-noise-only and complex degradations. The results demonstrate that Geneses significantly outperforms a conventional mask-based SE--SS method across various objective metrics with high robustness against complex degradations. Audio samples are available in our demo page. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_18456 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Geneses: Unified Generative Speech Enhancement and Separation Asai, Kohei Nakata, Wataru Saito, Yuki Saruwatari, Hiroshi Sound Audio and Speech Processing Real-world audio recordings often contain multiple speakers and various degradations, which limit both the quantity and quality of speech data available for building state-of-the-art speech processing models. Although end-to-end approaches that concatenate speech enhancement (SE) and speech separation (SS) to obtain a clean speech signal for each speaker are promising, conventional SE-SS methods suffer from complex degradations beyond additive noise. To this end, we propose \textbf{Geneses}, a generative framework to achieve unified, high-quality SE--SS. Our Geneses leverages latent flow matching to estimate each speaker's clean speech features using multi-modal diffusion Transformer conditioned on self-supervised learning representation from noisy mixture. We conduct experimental evaluation using two-speaker mixtures from LibriTTS-R under two conditions: additive-noise-only and complex degradations. The results demonstrate that Geneses significantly outperforms a conventional mask-based SE--SS method across various objective metrics with high robustness against complex degradations. Audio samples are available in our demo page. |
| title | Geneses: Unified Generative Speech Enhancement and Separation |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2601.18456 |