🎬 Fun-CineForge
Workflow: Upload short video → Add clip information (or upload .srt subtitle file) → Upload reference audio (optional) → Preprocessing, model loading, and inference → Output dubbed video
🎛️ Dubbing clip configuration
👤 Age
👤 Gender
📝 Instructions for use
| Fields | Descriptions |
|---|---|
| Dubbing script | The content of this clip (supports Chinese/English) |
| Clue description | Please refer to the sample format to explain the dubbing requirements, focusing on describing the speaker's gender, age, tone, and emotion |
| Timestamps | Start and end timestamps (accurate to milliseconds). The model is sensitive to timestamps; it is recommended to use timestamps adjacent to the audio clip. Duration ≤ 30s/clip |
| Age/Gender | Speaker attribute options |
| Reference audio | Voice cloning reference (Optional) |
⚠️ Note: Ensure that the timestamps of each clip don't overlap and don't exceed the video duration. The model will perform time alignment based on the timestamps, with weak supervision aligning lip movements.