Directing note: motion scores in open models
Two open models expose a single number for how much the picture moves: Step-Video-TI2V's motion_score, 5 by default, and Open-Sora's motion score, 4 by default. Both are written into the request rather than described in words. As of 2026-09-25.
Step-Video's notes give the trade plainly: a low score is steadier but can lack dynamism, a high score allows more intense action. Open-Sora adds a dynamic setting that lets a scorer pick the number.
A score is worth fixing early in a scene and keeping, so shots that follow one another move at a similar pace.
1From the docs
During inference, you can use the following command to generate videos with motion score (the default score is 4)
HPC-AI Tech, Open-Sora repository
The default motion_score = 5 is suitable for general use.
StepFun, Step-Video-TI2V repository
Select from the available camera Motion options to guide the camera's movement within the generated video
Adobe, Firefly help: generate videos using text prompts
- Motion scoreA motion score, 4 by default, fed into the prompt; a dynamic setting lets a scorer choose it
- Motion dynamics controlA motion_score setting, 5 by default; 2 for more stability, 10 or 20 for more intense action
- CameraCamera motion options to pick: zoom in or out, move left or right, tilt up or down, static and handheld
- Camera angles and shot typesShot sizes and angles are defined one by one, from extreme close-up to wide, low-angle and dutch angle
- Camera controlsMore than fifty named motion presets, from crash zooms to crane shots and bullet time
- Demo promptsCamera movement is written into the prompt; the demo prompts circle, tilt and zoom the camera
- Multi-shot examplesCamera moves are written into each shot of the prompt, as in its tracking-shot examples
- Motion ControlsMotion Control presets for Motion 2.0, from crane and dolly moves to orbits and bullet time
2Camera language in a pipeline
Where “Motion scores in open models” has to be written by hand, SceneMixer's pipeline writes the shots; its guides explain:
SceneMixer turns a novel into a watchable short-drama series in one pipeline: upload the novel (Chinese or English) → AI extracts characters, scenes, and an episode-by-episode structure → generates consistent character and scene art → writes storyboard scripts → renders video segment by segment (16:9 by default, 9:16 in one setting) → composites full episodes.
SceneMixer, Turn a Novel into a Video Series with AI (read 2026-10-13)
SceneMixer writes dialogue into the script analysis, assigns each line to a shot with a speaker and a short delivery note, and the video model performs it — picture and audio generated together, so the mouth forms that sentence in the language you chose.
SceneMixer, AI Dialogue Generator (read 2026-10-13)
3Other notes
Examples are quoted from the models' docs; no model was run.