Open Chinese voice dataset: 218 speakers, 80k+ clips, 98%+ phoneme accuracy
vista8 · x · 2026-10-12
While adding text-to-speech narration to Qiaomu Reader, the author found a Chinese voice dataset with 218 speakers and over 80,000 recorded clips. The audio is emotionally neutral and annotated with gender, age, and native accent, with 98%+ phonetic accuracy after professional labeling and quality checks.
It's handy for open-source TTS models like F5, Spark, and IndexTTS2, which all require a reference Chinese voice clip for synthesis — this dataset saves you from recording one yourself. Official site and samples are linked in the thread.
More from Models
- ThursdAI recap: OpenAI drops 722 math manuscripts, Haiku 5.5 hits 10 cents — altryne · 2026-10-13
- 800 Blackwells vs 100,000: can labs without GPUs survive the open-weights race? — altryne · 2026-10-13
- Swapping only the harness moves a model 35 points: 8090 Labs tests 7 models across 5 coding agents — pstAsiatech · 2026-10-13
- Koltun's spatial-cognition demo: GPT-4o 'can't find the bathtub'; puzzle training lifts a 9B model from 13% to 83% — le_james94 · 2026-10-13
- Meta's Superintelligence Lab ships Muse Wonder, its first frontier model — TechCrunch AI · 2026-10-13
- Meta launches Muse Wonder frontier model family, installable with one curl command — jiayi_pirate · 2026-10-13