A shot list needs translating before a model can read it

The shot

A traditional shot list records, for each setup, the subject, the action, the shot size, the camera move, the lens or style, and the duration a scene needs. Turning that list into a usable request for a generation tool means restating those fields in terms a specific interface actually accepts. Runway's own Text to Video Prompting Guide documents this translation directly for its Gen-4.5 model, stating that an effective prompt separates "visual components" (subject appearance, environment, lighting, composition or framing, style) from "motion components" (subject action, environmental motion, camera motion, motion style and timing, direction and speed).

What the documents show

The guide states a recommended, non-mandatory structure: "[Camera] shot of [a subject/object] [action] in [environment]. [Supporting component descriptions]," and gives worked examples such as a medium shot of a cowboy on a horse with a described action and a list of style terms. This is Runway's own documented convention, offered as a starting template rather than a required syntax; the same page states that "structure and order are far less important than clearly conveying an idea." Runway's separate Image to Video Prompting Guide documents a different division of labor for a shot that starts from a reference frame: the uploaded image is stated to define "composition, subject matter, lighting, and style," while the prompt is reserved almost entirely for motion. A shot list's reference-image field and its camera-move field map onto two different parts of that interface rather than one combined prompt.

The workflow

Building a generation brief from a shot list, per this documentation, means sorting existing shot-list fields into the categories the tool actually reads: an uploaded reference image or asset for anything about appearance and setting, a written camera-and-action line for anything about motion, and Runway's documented duration range of two to ten seconds per Gen-4.5 generation, recorded on its own Gen-4.5 model guide, as a hard constraint on how long a single requested shot can run before it must be split or extended.

What the tool does not change

None of the cited documentation claims the interface interprets a shot list on its own, resolves ambiguity the way a first assistant director or director of photography would in a shot-list meeting, or decides which shots in a sequence are worth generating at all. The translation from a production's shot list to a tool's accepted input fields is documented as a manual step performed by the operator.

• Which fields in this shot list belong in a reference image rather than in the prompt text?
• Does the requested shot fit inside the vendor's documented duration limit, or does it need to be split?
• Is the camera-and-action line specific enough to survive being separated from the visual description?

This is an editorial workflow guide built from what Runway's own documentation supports today; a different vendor's accepted fields would require a different translation, since none of the reviewed documentation describes a shared, tool-independent format for a generation brief.