Reference to video
Turns one reference image into a controlled video shot while preserving composition and identity.
VELVET / WORKFLOW LIBRARY
Curated production recipes and a creative workflow catalogue for film and video teams, with technical graphs kept behind the scenes. Find the right method, add it to your project and reuse it throughout production.
Turns one reference image into a controlled video shot while preserving composition and identity.
Preserves a character's face, wardrobe and visual language across multiple shots.
Creates a synchronized character performance from portrait, dialogue and audio inputs.
Directs dolly, orbit, pan and push-in camera language with natural instructions.
Transfers pose and movement from a source performer to a target character.
Changes only a selected region of a shot while preserving motion and surroundings.
Extends the beginning or end of existing footage while maintaining visual continuity.
A batch pipeline that turns storyboard frames into shots with a consistent visual language.
Controls an object's path and speed within the frame using reference signals.
Animates complex surfaces, mechanical parts and subtle motion with clarity.
Improves motion legibility and camera tracking in sports and action shots.
Cleans, restores and brings generated or archival footage to delivery resolution.
The catalogue only shows recipes verified on Velvet infrastructure with pinned versions.
Users change the goal, inputs and allowed settings while the technical graph stays protected in the recipe.
Every generation is stored with its sources, recipe version, variations and approval decision.