arXiv cs.CL· Kirill Borodin, Vasilii Kudryavtsev, Nikita Vasiliev, Grach Mkrtchian·· 13 小时前AI 评分31
研究：TTS 评测中 ASR 识别器自身错误占比几何
Measuring the Verifier's Floor: How Much Apparent Rare-Word Text-to-Speech Error Is ASR Error?
AI 导读
研究测量了用 ASR 转写评估 TTS 时识别器自身的错误：在 LibriSpeech 和 FLEURS 上，评估用识别器即使面对人类录音也漏掉 33-67% 的极罕见词，三个识别器均漏掉 22-42%。
正文 · 原文
Abstract: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.
| Comments: | Submitted to IEEE ICASSP 2027 |
| Subjects: | Sound (cs.SD); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.31470 [cs.SD] |
| (or arXiv:2610.31470v1 [cs.SD] for this version) | |
| https://doi.org/10.48550/arXiv.2610.31470 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Kirill Borodin [view email]
[v1]
Thu, 22 Oct 2026 15:59:54 UTC (35 KB)
来源：arXiv cs.CL · arxiv.org