Describe subject, a second-by-second action timeline, camera, environment, and audio in one block, with negative constraints, so the shot is controllable instead of a lucky roll.
提示词
Generate one continuous 8-second cinematic shot. Subject: [SUBJECT]. Action timeline: 0-2s [ACTION], 2-6s [ACTION], 6-8s [ACTION]. Camera: [LENS + MOVEMENT]. Environment: [PLACE + WEATHER]. Audio: [AMBIENCE]. Keep identity, wardrobe, and lighting consistent. No cuts, subtitles, watermarks, or extra limbs.
替换变量
[SUBJECT]The main subject.
[LENS + MOVEMENT]e.g. 35mm slow dolly-in.
该用哪个模型
- Veo 3.1 — Video generation is a Veo capability, not a text-model one.
示例输入
- [SUBJECT] = "a barista finishing a latte". Camera = "slow push-in".
示例输出
我们只发布来自真实模型运行的示例输出。此模板尚未经过正式测试，因此不显示输出。请用上方的输入自行运行。
为何这一结构有效
- A timeline gives the model temporal control instead of one static idea.
- Negative constraints suppress the usual video artefacts.
常见失败与修复
Motion is chaotic.
修复：Simplify to one clear action per time band.
已知局限
- Video output is non-deterministic; expect several tries per usable shot.