Metraipolis › archive › dc_210fe27b707d
Masked Autoregressive Speech Enhancement with Continuous Neural Audio Codec Representations — arXiv:2609.03940
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.03940
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T01:12:10.496Z
- SHA-256 of the archived text
348f89b5e339048ee34f7735c162f037e82a683f633a0f64688408ad8de274f1- Stable URL of this page
https://metraipolis.com/documento/dc_210fe27b707d
How to cite
Masked Autoregressive Speech Enhancement with Continuous Neural Audio Codec Representations — arXiv:2609.03940 — arXiv (en), https://arxiv.org/abs/2609.03940 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_210fe27b707d (Metraipolis archive dc_210fe27b707d, sha256 348f89b5e339048ee34f7735c162f037e82a683f633a0f64688408ad8de274f1)
BibTeX
@misc{dc210fe27b707d,
title = {Masked Autoregressive Speech Enhancement with Continuous Neural Audio Codec Representations — arXiv:2609.03940},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.03940},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_210fe27b707d; sha256 348f89b5e339048ee34f7735c162f037e82a683f633a0f64688408ad8de274f1},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.03940. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Most previous work on speech enhancement (SE) based on masked generative modeling relied on discrete token representations of audio signals, obtained using neural audio codecs (NACs). However, a recent study has shown that continuous latent representations of NACs can be advantageous for SE in terms of speech quality and intelligibility. In this work, we propose masked autoregressive SE (MARSE), a method for SE based on iterative decoding of masked clean speech frames using continuous NAC representations of speech. In particular, we investigate a set of different decoding policies, ceteris paribus, that is, using the same DNN (a Conformer model), the same NAC (the DAC codec) and the same training setup. The results show that MARSE enables a flexible trade-off between SE performance and computational cost. Audio examples and code are available online. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-04 in cs.AI). Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.03940 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T00:28:03.015Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-04?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).