AISHELL-6-Whisper
Mandarin audio-visual whisper speech dataset with parallel whispered and normal speech
Duration
59.3 hours
Languages
1
Sample Rate
48 kHz
Published
2025-09
Description
1Contains about 59.3 hours of data: 29.8 hours of whispered speech and 29.5 hours of parallel normal speech
2167 speakers participated, with 121 providing synchronized RGB-D video and 46 providing audio only
3Participants read 10-20 minutes of unique poetry texts, covering diverse Mandarin accents with balanced gender distribution
4Recorded in a controlled studio using a single-channel Neumann U87 microphone
5Synchronized frontal face video is provided at 1280x720 resolution and 25 fps
6Licensed under CC BY-NC-SA 4.0 for non-commercial use only
Language Details
| Language | Duration |
|---|---|
| Mandarin Chinese | 59.3 hours |
Publisher
License & Commercial Use
License
CC BY-NC-SA 4.0
Commercial Use
Non-commercial only