Metraipolis › archive › dc_b1eaa2e69596
The Voice on the Line — arXiv:2610.35272
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.35272
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-27T08:13:19.373Z
- SHA-256 of the archived text
070ec0cc4ecb18d5db2052259ba55d31dac981103998a7bcd68f9d8896b7c880- Stable URL of this page
https://metraipolis.com/documento/dc_b1eaa2e69596
How to cite
The Voice on the Line — arXiv:2610.35272 — arXiv (en), https://arxiv.org/abs/2610.35272 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b1eaa2e69596 (Metraipolis archive dc_b1eaa2e69596, sha256 070ec0cc4ecb18d5db2052259ba55d31dac981103998a7bcd68f9d8896b7c880)
BibTeX
@misc{dcb1eaa2e69596,
title = {The Voice on the Line — arXiv:2610.35272},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.35272},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b1eaa2e69596; sha256 070ec0cc4ecb18d5db2052259ba55d31dac981103998a7bcd68f9d8896b7c880},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.35272. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A transcriber listening through one microphone also hears everyone near it. We train causal voice isolators: masking transformers of 140M, 243M and 374M parameters, 20 ms of latency, enrolling the caller from the call itself. Short evaluation windows hide two failures that real calls expose: forgetting the caller when another voice talks alone for a while, and damaging a caller who is alone. Long calls with turns, memory beyond the attention window, real headset crosstalk and drifting caller levels address both, but each trades one against the other. WavLM-Large's own weights, made causal and trained with the model, move the trade: M with them (555M parameters) lowers the error on our development sets from 11.44% to 10.55%; a student trained from scratch to imitate WavLM does not, and w2v-BERT 2.0 and Whisper's encoder, pre-trained on far more audio, help less or not at all. A lone caller loses most at the start of a recording: no training call opened with someone already talking. Training on windows cut from longer calls lowers the error on clean development calls from 3.92% to 2.96%; this is our release. Live, the release makes 16.13% word errors on 60 s AMI windows, against 23.66% for VoxLock, the strongest open isolator, and 10.19% on its own AMI windows, against 11.71%; it is level on Krisp's benchmark (5.74% against 5.63%), and a caller who is alone still loses more words than with VoxLock (5.58% against 2.92% on clean test calls). Distilled into 22.1M parameters without WavLM, it runs live on one CPU thread inside a LiveKit agent, at 33% of a core per call in 20 ms frames, with 10.66% error on our development sets (the release 10.49%). We open-source the release and this CPU model, with a LiveKit plugin, at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Sound (cs.SD) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.35272 — 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 2026-10-27T07:56:00.317Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).