Skip to content
Fan Zeng (Frank)
Back to resources

AISHELL-3

AISHELL-3 provides neutral, multi-speaker Mandarin recordings with character and pinyin transcriptions.

Visit official source (opens in a new tab)

Research use

Use the parallel representations to compare pronunciation modeling and multi-speaker TTS.

Limitations

The recordings cover neutral delivery and a defined participant pool. Expressive and conversational output require separate evaluation.

License and access

Dataset license
Apache-2.0 (opens in a new tab)
Access status
Direct access

Use the publisher’s resource page and check the terms for the files and version you choose.

Additional sources