Driven by a first-frame image at a 1:2 ratio, the frame is stretched narrower than a phone's portrait screen; the prompt itself runs to about thirty words, using a static camera, barely-there movement, and no score to compress the whole shot into a suspended feeling. The three finished clips land on a giant beast, a mecha battlefield, and a rainy-night fall — same constraints, different first frames, different worlds each time.
This prompt runs under thirty words, using three negations to lock down camera and sound: static camera, no camera movement, no BGM — preempting the model's habit of adding motion and music on its own. Vague phrasing like 'time is slowed' and 'suspended' specifies no concrete action, handing narrative and content entirely to the first-frame image. The prompt itself only sets tone and constraints — a minimal control-word approach built for image-to-video generation.
Want tighter control? Editor-expanded reference version (not the original prompt):
Static shot, with the framing held completely unchanged across the whole clip. Keep only the faintest traces of motion — a slight tremor at the edge of fabric, light slowly crawling across a surface, a barely perceptible shift somewhere in the background — everything else stays close to motionless. Light comes from a single, unchanging source with constant color temperature; no shifts between light and shadow, so the sense of time flattens into a suspended state. Material detail should stay legible — skin texture, fabric folds, metal reflections — without any of it actually moving. Pace by subtraction: no build-up, no climax, almost nothing perceptibly changes across the 10 seconds, and that restraint of 'nothing happening' is what creates the frozen feeling. Leave the sound entirely blank — no score, no ambient effects, absolute silence. Negative constraints: no dolly, pan, or tilt, no zoom, no transitions, no large deliberate character movement, no sudden lighting change.
Use a first-frame image with a 1:2 aspect ratio to generate 1:2 videos.
1:2のアスペクト比の最初のフレーム画像を使用して、1:2の動画を生成する。
✍️ Editor’s Notes
This isn't a scene description, it's a workflow note: the first-frame image's own aspect ratio directly locks the output video's aspect ratio, skipping the need to declare frame dimensions in the prompt text and avoiding the risk of the written description clashing with the image's own ratio. In image-to-video generation, the first frame's composition, ratio, and lighting are a stronger constraint than any written description — this is the classic trick of pushing parameters into the image instead of the text, and it shows this example's real focus is the workflow, not the prompt itself.
