Back to MiniMax H3 prompts
Audio edit / Reference Generation
Official manual translationTranslated from a MiniMax H3 manual exampleDialogue and performance replacement
A new line may require a bounded performance change, not a full scene regeneration.
Actual H3 output / 10s
Prompt -> result
What this output demonstrates
The existing farewell scene is reinterpreted through a new spoken plea and a matching emotional performance.
- Control pattern
- A new line may require a bounded performance change, not a full scene regeneration.
- Check in your run
- Long replacement dialogue can overrun the source shot and break lip timing.
Shot structure
0s5s10s
INPUTR2V
ACTION01
CAMERA01
AUDIOON
Working principle
A new line may require a bounded performance change, not a full scene regeneration.
Watch for
Long replacement dialogue can overrun the source shot and break lip timing.
Adapt this brief
Prompt draft
Replace the woman dialogue in Video 1 with the line from Audio 1: "Please do not leave. This time, let us not let go of each other." Adjust only the necessary facial and emotional performance to fit the new plea. Preserve the characters, scene, camera, timing, and surrounding ambience.Replace the subject and reference roles first. Change camera, timing, and sound one variable at a time.