Kling motion control camera movement and orientation

In Kling motion control, character orientation decides whether the camera follows the video or the image. What Kling says, and Sume's character_orientation.

4 min readSume
All posts

In Kling motion control, camera movement depends on the character orientation setting. By default the character's movements, expressions, camera movements and orientation follow the motion reference video; choose the image-matching option and the orientation follows your still instead.

This is from the Video 2.6 section of Kling's Motion Control User Guide, read 2026-09-29. Sume exposes the same choice as a request field, per the Sume API reference and the API reference docs.

What are the two orientation options?

Kling names them by what they match. The default is Character Orientation Matches Video; the alternative is Character Orientation Matches Image.

From Kling's Motion Control guide, read 2026-09-29.
OptionWhat follows the reference videoWhat else
Matches Video (default)Movements, expressions, camera movements and orientation.Other details can be controlled with the prompt.
Matches ImageMovements and expressions.Orientation aligns with the still. Camera movements and other elements can be customized through prompts.

Which option lets me move the camera?

Kling says the Matches Image choice supports camera movement, and its guide shows five examples of it: zoom in, zoom out, camera up, camera down and fixed position. The guide also says to avoid camera movements inside the motion reference itself, since they can truncate the result.

The practical reading: keep the reference clip's camera still, then, if you want the camera to move, use Matches Image and describe the move in the prompt.

What does Sume's character_orientation do?

POST /v1/kling/3.0/motion-control takes character_orientation with two values, video and image. The reference describes it as whose framing wins: video, the default, keeps the driving video's orientation, and image keeps the still's orientation.

Note the prompt wording. Sume's field description says motion and framing follow the driving video and that the prompt only steers appearance details. Kling's guide says prompts can customize camera movement under Matches Image. Because the two descriptions differ, do not assume that a camera-move prompt will work on Sume; test it on a short clip first.

curl -X POST https://api.sume.com/v1/kling/3.0/motion-control \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: motion-orient-001" \
  -d '{
    "image_url": "https://example.com/portrait.png",
    "motion_video_url": "https://example.com/walk-6s.mp4",
    "duration_seconds": 6,
    "character_orientation": "image"
  }'

Which orientation should I pick?

Choose Matches Video (video on Sume) when you want the reference clip's framing and camera to carry over to the result. Choose Matches Image (image) when your still already has the composition you want to keep, and the reference should supply only the character's movement and expressions.

That split is our reading of Kling's two descriptions above, not a rule Kling states in those words.

Can I switch orientation after the clip is made?

Not in the docs. The orientation is a request field, so a different choice is a new request with its own price; send a new Idempotency-Key, because the same key with a changed body answers 409 idempotency_conflict. For the rest of the fields, see Motion control API: animate an image with a driving video.

Sources

Related posts

More in Models

All Models posts

Written by Sume