PROMPT RADAR / IMAGE平面设计

个人资料 / 头像 - White Suit Man in Hooded Crowd

An ultra-realistic black-and-white cinematic fashion portrait of a confident man in a white suit surrounded by ominous hooded figures.

@OrhanGhazi65942 0 热度 0 浏览
个人资料 / 头像 - White Suit Man in Hooded CrowdORIGINAL VISUAL / 来源见文末
编辑先看

它为什么值得进入灵感库

这条内容值得留下的,不是表面的风格词,而是它提供了一条可以继续实验的画面路径。 先规定信息层级和文字安全区,再谈视觉风格;无法落版的漂亮画面不是完整设计。

完整视觉 Prompt

Create an ultra-realistic cinematic black-and-white editorial portrait of a {argument name="character description" default="handsome, confident young man"} standing calmly in the exact center of a mysterious crowd. He wears a perfectly tailored {argument name="suit color" default="white"} luxury suit jacket with sharp lapels over a slightly open white dress shirt, no tie, elegant and sophisticated. His hair is dark, thick, tousled, and textured; he has a clean-shaven face with subtle stubble, strong jawline, direct calm eye contact, and a composed faintly serious expression. Surround him with exactly 12 visible hooded figures in black robes: 2 large blurred foreground silhouettes framing the lower left and lower right edges, 4 midground hooded figures with partially visible shadowed faces around his shoulders, and 6 background hooded figures fading into darkness behind him. The hooded crowd should feel anonymous, ominous, and cult-like, with faces mostly hidden in deep shadow and only faint eyes/noses visible. Use a tight vertical portrait composition, chest-up framing, centered symmetry, shallow depth of field, dramatic high-contrast lighting, soft diffused light on the man’s face and white suit, deep black background, moody editorial fashion photography, 85mm lens look, realistic skin texture, crisp suit fabric, film-grain texture, no text, no logos, no color, no extra subjects.
查看原作者内容

不要一次改完所有变量

用 GPT Image 2 复刻时,先固定主体和构图,再逐次替换光线、色调或材质中的一个变量,这样更容易判断哪项描述真正起作用。

  1. 主体:An ultra-realistic black-and-white cinematic fashion portrait of a confident man in a white suit surrounded by ominous hooded figures.
  2. 视觉方向:社区实测摄影提示词;按原提示词的光线描述执行
  3. 镜头与构图:按原提示词的镜头与景别执行;按原提示词的构图与画幅执行
  4. 交付约束:auto,遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。
负向约束

遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。

模型建议

GPT Image 2 / Grok Imagine / Midjourney / Flux。先用原始 Prompt 建立基准结果,再根据模型对自然语言、镜头运动或文字排版的能力做局部改写。

来源与编辑原则

保留作者,也保留判断。

本页收录公开案例并提供中文索引、结构化拆解和复刻建议,不冒充原作者作品,也不替代原帖上下文。

作者:ORHAN · 收录于 2026年8月12日

访问原始来源

沿着这条线继续看