Metraipolis › archive › dc_0ff562bd56ec

Measuring the Verifier's Floor: How Much Apparent Rare-Word Text-to-Speech Error Is ASR Error? — arXiv:2610.31470

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.31470
License (as archived): CC0-1.0
Archived: 2026-10-25T08:36:47.814Z
SHA-256 of the archived text: 4a42a4bef6809e494216de9bedf8b35b8da5b31024d3ef46834bd7d52b58a7f5
Stable URL of this page: https://metraipolis.com/documento/dc_0ff562bd56ec

How to cite

Measuring the Verifier's Floor: How Much Apparent Rare-Word Text-to-Speech Error Is ASR Error? — arXiv:2610.31470 — arXiv (en), https://arxiv.org/abs/2610.31470 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0ff562bd56ec (Metraipolis archive dc_0ff562bd56ec, sha256 4a42a4bef6809e494216de9bedf8b35b8da5b31024d3ef46834bd7d52b58a7f5)

BibTeX

@misc{dc0ff562bd56ec,
  title = {Measuring the Verifier's Floor: How Much Apparent Rare-Word Text-to-Speech Error Is ASR Error? — arXiv:2610.31470},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.31470},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0ff562bd56ec; sha256 4a42a4bef6809e494216de9bedf8b35b8da5b31024d3ef46834bd7d52b58a7f5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31470. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text-to-speech (TTS) is usually evaluated by transcribing its output with an automatic speech recognition (ASR) system, so the word error rate also counts recognizer errors. We measure these errors on human recordings of the test sentences. On LibriSpeech and FLEURS, the evaluation recognizer misses 33-67% of very rare words even in human speech, and all three recognizers miss 22-42%. Checking every word with two more recognizer families lowers the apparent TTS error on words that at least one recognizer hears two- to three-fold (very rare LibriSpeech words: from 10-13% to 3-5%), yet very rare words remain the hardest. Ignoring words the recognizer misses in human speech is biased: the other recognizers find TTS errors far more often on these words (4-6% vs. 0.2-0.3%) and overturn about half of the errors this mask keeps. As half of the remaining adjudicated errors may be spelling variants, a listening test is needed.

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/2610.31470 — 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 2026-10-25T08:31:27.846Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).