Act-Two
Use Act-Two to transfer a visible-face performance to one character source. The selected reference segment supplies the performance timing; body control and expression intensity determine how strongly its movement and facial expression guide the result.
Properties
- Character input: exactly one image or video; performance input: one selected 3–30 second video segment.
- Controls include body-motion transfer, expression intensity from 1 to 5, output ratio, and seed.
How to use this model
- 1
Upload exactly one character image or character video with an unobstructed face.
- 2
Select 3–30 seconds from the performance recording.
- 3
Set body control, expression intensity, and output ratio explicitly.
- 4
Review source framing and face visibility before starting the generation.
Parameters
Character image
character_imageUse an image with a clearly visible face.
Character video
character_videoUse a video with a clearly visible face.
Performance video
reference_videoBody control
body_controlExpression intensity
expression_intensityOutput ratio
ratio16:9 landscape
9:16 portrait
1:1 square
4:3 landscape
3:4 portrait
21:9 ultrawide
Seed
seed