CN-Celeb
A large-scale Chinese celebrity speaker recognition dataset with 1,000 speakers across 11 real-world genres, collected in-the-wild
Duration
274 hours
Languages
1
Sample Rate
16 kHz
Published
2019-11
Description
1Contains over 130,000 utterances from 1,000 Chinese celebrities, totaling approximately 274 hours
2Covers 11 real-world genres: entertainment, interview, singing, drama, movie, vlog, livestream, speech, opera, recitation, and advertisement
3In-the-wild collection with real environmental noise, diverse speaking styles, cross-session and cross-device variations
4Free for research use
Language Details
| Language | Duration |
|---|---|
| Mandarin Chinese | 274 hours |
Publisher