Metraipolis › archive › dc_21023eda758b

Repetition, Not Length: Isolating the Counting Failure in Neural Text-to-Speech — arXiv:2609.36974

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.36974
License (as archived)
CC0-1.0
Archived
2026-10-07T16:56:47.215Z
SHA-256 of the archived text
750a593c703e0ab5f2d7db522f35840bc1343b8e7d39d1fc81489c01143afe30
Stable URL of this page
https://metraipolis.com/documento/dc_21023eda758b

How to cite

Repetition, Not Length: Isolating the Counting Failure in Neural Text-to-Speech — arXiv:2609.36974 — arXiv (en), https://arxiv.org/abs/2609.36974 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_21023eda758b (Metraipolis archive dc_21023eda758b, sha256 750a593c703e0ab5f2d7db522f35840bc1343b8e7d39d1fc81489c01143afe30)
BibTeX
@misc{dc21023eda758b,
  title = {Repetition, Not Length: Isolating the Counting Failure in Neural Text-to-Speech — arXiv:2609.36974},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.36974},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_21023eda758b; sha256 750a593c703e0ab5f2d7db522f35840bc1343b8e7d39d1fc81489c01143afe30},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36974. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text-to-speech models loop, truncate and lose count on text that repeats a phrase many times. We show that repetition itself is what breaks them, not the length that comes with it. Every repeated sentence in our test set is paired with a control of matched sentence and word count in which no word ever repeats back-to-back. Six models from three architectures render the controls almost perfectly and fail the repeated twins: 94.3% against 18.2% exactly right at k >= 6. The gap survives greedy decoding, repetition-penalty sweeps, four independent speech recognisers and 420 analysis specifications without once reversing sign; a held-out fourth architecture lands within a point of its predicted gap, and one of two non-autoregressive baselines shows the same failure. Varying the period of the text shows the failure grows smoothly with periodicity, half of it surviving when no word is adjacent to itself.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2609.36974 — 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 2026-10-07T16:27:03.261Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).