CineDub: Large-Scale Chinese Television Dubbing for Cinematic Scenes

CineDub-CN provides audio and text pairs for Chinese television dubbing across monologue, narration, and dialogue scenes. FunAudioLLM released this resource in 2026 as the first large-scale dataset for zero-shot movie dubbing in diverse cinematic environments.

Use Cases

• Training MLLM-based dubbing models using monologue and dialogue scene labels
• Evaluating zero-shot performance on diverse cinematic scenes
• Developing cross-lingual speech synthesis systems for Chinese and English

Strengths

• First large-scale Chinese television dubbing dataset
• Includes specific labels for monologue, narration, and dialogue scenes
• Supports zero-shot dubbing model training

Limitations

• Example repository status indicates this may not contain the full CineDub-CN dataset
• Missing explicit column definitions and row counts in metadata

Provenance

Source

FunAudioLLM

Collection Method

pipeline-generated

Freshness

Last updated March 2026.

Geography

China