Metraipolis › archive › dc_a7ea8af6449e

Parallel Time-Band Mixing with Learned Observation-Adding for Robust ASR Front-Ends — arXiv:2608.30326

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.30326
License (as archived)
CC0-1.0
Archived
2026-10-17T07:52:36.390Z
SHA-256 of the archived text
68aa1cf6e1e7056cda2b7fab4c1abe548fc92c2de040eb622af21fa6c49944c4
Stable URL of this page
https://metraipolis.com/documento/dc_a7ea8af6449e

How to cite

Parallel Time-Band Mixing with Learned Observation-Adding for Robust ASR Front-Ends — arXiv:2608.30326 — arXiv (en), https://arxiv.org/abs/2608.30326 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a7ea8af6449e (Metraipolis archive dc_a7ea8af6449e, sha256 68aa1cf6e1e7056cda2b7fab4c1abe548fc92c2de040eb622af21fa6c49944c4)
BibTeX
@misc{dca7ea8af6449e,
  title = {Parallel Time-Band Mixing with Learned Observation-Adding for Robust ASR Front-Ends — arXiv:2608.30326},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.30326},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a7ea8af6449e; sha256 68aa1cf6e1e7056cda2b7fab4c1abe548fc92c2de040eb622af21fa6c49944c4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.30326. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Speech enhancement is often used as a front-end for robust ASR, yet recurrent temporal and cross-band modules introduce sequential dependencies that reduce parallel efficiency. In this paper, we present a sequence-parallel band-split enhancement front-end built on a Parallel Time-Band Mixer (PTBM) block that eliminates within-block recurrent unrolling. PTBM integrates intra-band temporal mixing and per-frame cross-band attention within a unified parallel architecture, enabling efficient contextual modeling across both time and frequency dimensions. The system retains the mask-plus-residual reconstruction interface and introduces learned Observation-Adding (LOA) to suppress ASR-sensitive artifacts without development-set tuning. Experiments on DNS Challenge and CHiME-4 with frozen Whisper back-ends show that the proposed front-end consistently reduces word error rate relative to recurrent band-split baselines while requiring only 0.96 M parameters and 0.58 GMAC/s for the front-end network.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-01 in cs.AI).
Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.30326 — 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 2026-10-17T07:28:26.857Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).