What the cartoon preset changes
Photographic texture, skin pores, fabric weave and lens noise are replaced by flat colour areas separated by dark outlines. Lighting is simplified into two or three tones per surface, which is what gives the footage its drawn feel.
Camera movement, subject movement and scene layout come from your source video, so a locked-off interview stays an interview and a handheld walk stays a walk. The model does not invent new camera moves.
Footage that converts well
- Single subject against a plain or moderately busy background
- Even daylight or soft indoor lighting with good contrast
- Medium and close shots where faces fill a reasonable part of the frame
- Slow to moderate motion such as talking, walking or gesturing
Footage that struggles
- Fast whip pans and strobing lights, which can cause flicker between frames
- Crowds and tiny faces, where detail turns into smudges
- Screen recordings with small text, which becomes illegible
- Very dark or heavily compressed clips
Prompt tweaks that help
Style presets are text prompts for a generative video model, so results vary with your footage; test a short clip before processing a long one.
You can append short instructions to the preset such as 'warm sunset palette' or 'thinner outlines'. Keep additions to one or two ideas; stacking many conflicting adjectives usually weakens the effect.
Output details
Exports are 720p MP4. The upstream model processes the visual track only, so Inkframe re-attaches your original audio when the server has ffmpeg available, and tells you in the project page when it could not.