A production technique where a previously generated video clip — not just a still image — is uploaded as a reference input for a new AI video generation, used to visually anchor continuity across scenes (e.g. closing a narrative loop back to an opening shot).
Distinct from Build-Animate-Extract Loop (Frame Extraction for Iterative Generation), which extracts a single still frame from a prior output to use as a new start-frame image: here the whole clip is fed in as an element/reference alongside a still anchor image, giving the model motion and appearance context from an actual generated result rather than just a photo.
Seen in a Seedance 2.0/Higgsfield workflow (Higgsfield @-Element Asset System) where a decade-styled reference photo anchored the character's appearance for a given era while an earlier generated clip was additionally uploaded as a reference to anchor the closing shot's look back to the opening scene.
Apply: When a later scene must visually match an earlier one (e.g. to close a loop), upload both the still anchor/reference image and the earlier generated video clip as references for the new prompt, rather than relying on the anchor image alone.
Из тем: Holding Continuity Across Independently Generated Shots