PROMPT RADAR / IMAGE人像时尚

个人资料 / 头像 - Neo-Noir Cinematic Rain Portrait

An ultra-realistic cinematic night portrait of a man in a black raincoat with an umbrella, featuring dramatic lighting and a moody urban atmosphere.

@iamrealsnow 0 热度 0 浏览
个人资料 / 头像 - Neo-Noir Cinematic Rain PortraitORIGINAL 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。 先锁定人物身份与服装轮廓,再用光线、焦段和动作消除“普通写真感”。

完整视觉 Prompt

Ultra-realistic cinematic night portrait of the person from the uploaded photo, preserving the exact facial features, hairstyle, skin tone, and identity. He is wearing a {argument name="outerwear" default="premium black waterproof raincoat"} with {argument name="gloves" default="black leather gloves"}. In one hand, he holds a {argument name="accessory" default="large open black umbrella"}, tilted slightly backward to cover his head and shoulders while keeping his entire face fully visible. His face is illuminated by a soft, bright directional key light that enhances his facial structure, sharp jawline, and natural skin texture. Expression is calm, confident, and slightly thoughtful, looking directly at the camera. The background features a blurred nighttime metropolitan street with glowing streetlights, cinematic bokeh, wet asphalt reflecting warm city lights, and subtle rainfall. Raindrops are barely visible, creating a moody, atmospheric ambience. Warm orange and golden highlights contrast with cool blue shadows, giving a premium neo-noir cinematic look. Professional editorial photography, shallow depth of field, ultra-detailed, HDR, 85mm lens, f/1.4, realistic skin texture, natural proportions, dramatic lighting, 8K, masterpiece, photorealistic.
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 cinematic night portrait of a man in a black raincoat with an umbrella, featuring dramatic lighting and a moody urban atmosphere.
  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:Snow · 收录于 2026年8月8日

访问原始来源

沿着这条线继续看