Sora-style (do not paste as-is)
World-sim paragraph
Photoreal cinematic shot of a woman in a wool coat in a Tokyo alley at dusk. Steam, neon, she looks up as a train passes. The camera discovers her face. 12 seconds, immersive physics, native rain audio.
Seedance shape
FORMAT: 16:9, 12s, photoreal dusk. SUBJECT: Woman in a wool coat at a ramen cart; she reads a ticket, then looks up. TIMELINE: 0–5s ticket + steam. 5–12s look-up; only then a slow dolly-in. CAMERA: One push-in triggered by the head raise. AUDIO: Steam, train, light rain. No music. Full block: remap page → dolly-in.
Kling shape
Shot 1: Medium-wide static, ticket in hand. Shot 2: Same axis, dolly 1.5m to MCU as she looks up. Do not orbit. Staging stays identical. Photoreal neon, 16:9. Full block: remap page → dolly-in.
Veo shape
8-second 16:9 photoreal clip. Medium-wide, then a slow dolly-in as she looks up at a passing train. Steam, neon in puddles. Diegetic steam / train / rain. No music, no captions. Full block: remap page → dolly-in.