PROMPT RADAR / PROMPT平面设计

个人资料 / 头像 - Two-Tone Studio Portrait

A high-end studio portrait prompt using magenta and blue lighting for a dramatic, editorial fashion aesthetic.

@Sairah_0 0 热度 0 浏览
个人资料 / 头像 - Two-Tone Studio PortraitORIGINAL VISUAL / 来源见文末
编辑先看

它为什么值得进入灵感库

先别急着照抄整段 Prompt:真正有用的是主体、空间和镜头之间的约束关系。 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 a high-end photorealistic studio portrait of a {argument name="subject" default="male subject"}, framed from the chest up, facing directly toward the camera with a calm, neutral and confident expression. The subject has well-defined masculine facial features, realistic natural skin texture, detailed eyes, subtle pores, clean grooming, and neatly styled hair. He is wearing a simple oversized black crew-neck T-shirt. Use a minimal seamless studio background with a smooth {argument name="background gradient" default="purple-to-deep-blue"} gradient. Dramatic cinematic two-tone lighting: {argument name="primary lighting" default="vibrant magenta/pink rim light"} from the left side and cool electric-blue light from the right, creating subtle colored highlights along the face, ears, hair and shoulders while maintaining natural skin tones. Soft frontal fill light keeps facial details clear, with controlled shadows and strong dimensionality. Symmetrical composition, centered head and shoulders, eye-level camera, straight-on perspective, medium close-up, 85mm portrait lens look, shallow depth of field, crisp facial focus, soft background falloff, premium fashion/editorial photography, cinematic color grading, realistic skin, high dynamic range, professional studio photography, ultra-detailed, clean composition, no props, no text, no logos.
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:A high-end studio portrait prompt using magenta and blue lighting for a dramatic, editorial fashion aesthetic.
  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:Sairah · 收录于 2026年8月10日

访问原始来源

沿着这条线继续看