MiniMax H3 Original MP3 Lip-Sync Workflow Fixes the issue where input audio is regenerated into pseudo-Chinese vocals. This workflow passes through your original MP3/vocal directly to the final video output while maintaining H3's precise lip-sync.
Generation Speed: ~228s for 10-second video (INT8/NVFP4 environment)
No missing custom nodes needed, standard setup.
Description
Update: Added I2V (Image to Video) Simple Version!
- Perfect for single-image animation with original MP3 lip-sync.
- Cleaner and lighter node setup compared to R2V.
- Same zero audio corruption / perfect Japanese vocal sync.