Glossary
Accent coverage
How well a speech dataset represents the accents and dialects of a language. Broad coverage makes ASR and TTS systems work for more of the people who use them.
Coverage is described against a target: the regional accents of one language, global varieties of English, or the dialect families of a language such as Arabic. ASR error rates are consistently higher for under-represented accents — a documented fairness gap that buyers work to close. Coverage claims rest on speaker metadata collected with consent.
Why it matters
Model teams buy against specific accent gaps, so accent labels are often the first field they filter on.