
原始文档
GPT Image 2 · en
Scene:
A square cinematic frame built with absolute left-right symmetry and a straight-on camera at human height. The palette is powder pink, mint, butter yellow, pale blue, and faded coral, with soft daylight and a miniature dollhouse clarity. The architecture should feel handcrafted, tidy, slightly theatrical, and emotionally precise rather than casual or naturalistic. Stage the image inside an elevator vestibule with centered brass doors, mirrored call panels, matching luggage carts, and a runner carpet splitting the composition into equal halves.
Subject:
Place one or two front-facing figures exactly on the central axis, with costumes, props, and furnishings arranged like curated museum pieces. The people should feel composed and deadpan, never candid, and the room must read as a cutaway stage set where every shelf, lamp, chair, and curtain participates in the symmetry. A porter stands square to the lens with white gloves and a parcel balanced at chest height, while a guest in a lemon coat waits two steps behind on the same axis.
Important details:
The signature rules are central balance, flat frontal staging, pastel temperature control, crisp prop spacing, and a charming handmade scale. Preserve clean rectangles, color-coded rooms, disciplined eye-line, and the sense that the camera is respectfully observing an invented world. Avoid messy realism, loose camera motion, asymmetrical clutter, and dramatic lens distortion. Let the charm come from the repetition of sconces, floor tiles, and parcel ribbons, all rendered with handmade neatness and pastel restraint.
Use case:
Symmetrical Pastel Dollhouse Cinematography single study 9/9.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no dutch angle, no chaotic crowd, no gritty grime, no harsh neon, no visible brand packaging, no broken anatomy, no low-detail background, no accidental text fragments.使用方法
- 复制完整原始提示词,并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用,核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型,不代表该示例由这个模型生成。
- 核对名称、数字与画面文字;不同模型版本的结果可能不同。





