Ken Burns effect API: Timeline stills are static holds
Sume Timeline holds a still image static; motion is accepted and ignored with a motion_ignored warning. zoompan is on the video filter allowlist.

Sume's Timeline does not zoom or pan a photo. A still in video[].source_url is a static hold, and a motion field on it is accepted and ignored, with a motion_ignored warning on the result. A slow Ken Burns move has to come from a clip that already moves.
Opus Clip's Sept 18 Motion Studio note describes turning up to 30 images and plain language into a finished clip. Sume's Timeline is the assembly step, not an animator. Details: Timeline 1.0 docs, read 2026-10-01.
What does Timeline do with a still?
| Input | Result |
|---|---|
Still, no motion | Static hold for duration seconds |
Still with motion | Held static; motion_ignored warning |
| Moving clip | Plays from source_in for duration |
| Short clip in a long slot | Padded or looped, with a warning |
Is the warning a failure?
No. The docs say soft warnings, including ignored still motion, padded or looped short sources and snapped transitions, are not failures. The job completes and bills as usual; you just do not get the move you asked for. Check warnings[] on GET /v1/jobs/:id/result before you ship.
Where does zoompan come in?
The compiler's filter allowlist for the video filter surface includes zoompan, which is ffmpeg's zoom-and-pan filter. Video filter takes a Sume-hosted clip and a filters-only filtergraph with inline options. I have not confirmed from the docs that a still image is accepted there, so the safe reading is: give it a clip, not a bare photo. Confirm against the live catalog and a test run before building on it.
What can I do today for a photo sequence?
Hold each photo for its slot, add transition types such as fade or dissolve between slots, and lay a voice or music spine under it. That is the pattern in image slideshow video API and product photos to a Timeline video. If you need real camera motion on a photo, generate it as video first and put that clip in the slot.
Sources
Related posts
More in Use cases
- AI kids story video generator: scenes of 2 to 30 seconds
Make a children's story video as one scene per request: wan-3.0 takes 2 to 30 seconds per scene, then join up to 200 slots in a Timeline 1.0 render.
- Kling 4.0 Creative Recreation vs Sume reference ingest manifest
Kling 4.0 Creative Recreation borrows a reference video's pacing and shot structure. Sume reference ingest returns the shots, text and audio facts first.
- AI real estate walkthrough from photos: chain first/last-frame clips
Kling 4.0 takes up to 10 keyframes for a walkthrough. Sume takes a first and last frame per clip, so chain one clip per room pair and join them on a timeline.
- 가상인물 후기 광고: experience lines in a Sume avatar script
공정위는 가상인물 표시를 해도 사용경험처럼 말하면 실제와 맞지 않을 때 부당할 수 있다고 합니다. 아바타 스크립트에 쓰는 1인칭 후기 문장을 점검하세요.
Written by Sume