Match the visible body to the action
For a dance or full-body movement, use an image that shows the limbs and leaves room around the character. A close portrait is better suited to a smaller gesture than a full-body routine.
0 / 2500
Reference image defines characters, background, and other elements. Size needs to be ≥300px, aspect ratio 2:5–5:2.
Use Kling 3.0 Motion Control on Magnific AI to animate a character image with movement from a reference video. The picture guides the character’s appearance; the video supplies the performance. Create a rendered clip rather than a motion-capture file or a rigged 3D animation.
Start with two inputs that work together: a clear character image and a continuous video of the action.
For a dance or full-body movement, use an image that shows the limbs and leaves room around the character. A close portrait is better suited to a smaller gesture than a full-body routine.
Use a clip where the main performer stays visible. A continuous shot with clear movement gives a more useful guide than rapid cuts, heavy obstruction, or several people moving over one another.
Character orientation determines whether the image or the reference video guides the character’s facing direction. It is different from choosing portrait or landscape output.
Use Image when the facing direction in your character picture is important to the shot. Check that the reference action can work with that orientation.
Use Video when the performer’s facing direction should guide the result. Match the character framing to the movement, especially when the performer turns or travels through the shot.
The reference clip already describes the performance. Keep any extra instructions focused on elements it does not define.
Add a short description of the surroundings or camera behavior when needed. Avoid writing an unrelated sequence of movements that competes with the reference performance.
Choose 720p or 1080p for the output. Watch the entire clip and inspect hands, feet, body proportions, and facial features as the subject turns or passes behind objects.
Motion control is useful when you can show the movement more clearly than you can describe it.
Use a clear routine to explore how a character looks in motion. Frame the image to include the joints and limbs used by the performer.
Transfer a wave, a greeting, or another recognizable action to a character image. Similar body proportions between the reference and character make the intended movement easier to interpret.
Choose the input that should determine the result.
Image to video creates movement from a still image and your written direction. Motion control uses a video performance as the movement guide, making it useful when timing and gestures are central to the idea.
For a person or character delivering an audio track, the avatar tool is the more direct starting point. Motion control is for reference-driven movement, not a replacement for a dedicated audio-to-portrait workflow.
Input matching, orientation, and choosing the right animation workflow.
Provide a character image and a reference video of the movement. Use a clearly visible subject in each. You can add a text prompt to guide the scene, but the video remains the main performance reference.
The supported range depends on character orientation: 3–10 seconds for Image orientation and 3–30 seconds for Video orientation. Choose the mode for the facing direction you want, then use a reference within its duration range.
Image orientation follows the facing direction in the character picture. Video orientation follows the performer in the reference clip. This setting concerns the character’s orientation, not the frame’s aspect ratio.
Yes. Start with a clear character image whose proportions and visible body parts fit the reference action. Review limbs and facial details in the result; heavily stylized shapes may need a simpler performance.
A difficult pose, hidden limbs, rapid movement, or mismatched framing can make the transfer harder to interpret. Try a clearer reference, a more compatible character image, or a simpler action, then compare the result through the full clip.
No. This tool generates a video. It does not provide a skeleton, animation rig, or motion-capture file for a 3D application. Use it to create a visible character performance rather than an editable motion dataset.
Pair a character image with a performance you want it to follow.
Create Motion Video