Kirill Borodin, Vasilii Kudryavtsev, Nikita Vasiliev, Grach Mkrtchian · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.31470 · cited by 0 works (OpenAlex) · open-access PDF
텍스트-투어 (TTS)는 일반적으로 자동 음성 인식 (ASR) 시스템을 사용하여 출력을 번역하여 평가되므로 단어 오류 비율은 인식기 오류를 계산합니다. 우리는 테스트 문장의 인간 기록에 이러한 오류를 측정합니다. LibriSpeech 및 FLEURS에서 평가 인식기는 인간의 말에서도 매우 드문 단어의 33-67%를 놓치고 세 가지 인식기 모두 22-42%를 놓치고 있습니다. 두 개의 인식기 가족이있는 각 단어를 확인하면 적어도 하나의 인식기가 두 번 또는 세 번 (매우 드문 LibriSpeech 단어 : 10-13%에서 3-5%)을 들을 수있는 단어에 대한 TTS 오류가 낮아집니다. 그러나 매우 드문 단어는 인간의 말에서 인식기 오류를 무시
Machine translation (Meta M2M100); the English original follows.
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.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Speech recognition, Computer science, Word error rate, Word (group theory), Spelling, Active listening
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.31470, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange