PROMPT RADAR / PROMPT人像时尚
个人资料 / 头像 - Photorealistic Dark-Haired Woman Portrait
A detailed prompt for creating a photorealistic portrait of a woman based on a reference image, focusing on natural lighting and professional camera settings.
@AmeliaAi12 0 热度 0 浏览
ORIGINAL 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”。
01 / 原始方法
完整视觉 Prompt
Use the provided reference image as the main visual reference. Create a highly photorealistic close-up portrait of a young woman with {argument name="hair style" default="long, straight, silky dark-brown hair"}, parted naturally in the center and falling loosely over her shoulders. Preserve the same facial structure, soft features, natural eyebrows, almond-shaped dark eyes, delicate nose, smooth fair skin, and soft pink-red lips. She is looking toward the camera with a gentle, subtle smile and a calm, elegant expression. She is wearing a simple {argument name="clothing" default="light-gray sweatshirt"} with a relaxed fit. Capture her from the shoulders up in a slightly turned three-quarter pose, with her face angled gently toward the camera. Use a {argument name="background" default="warm, cozy indoor background with soft brown and beige tones"}, naturally blurred with shallow depth of field. Photorealistic skin texture, realistic hair strands, natural facial details, soft warm ambient lighting, subtle highlights, realistic shadows, cinematic portrait photography, 85mm lens look, shallow depth of field, high detail, professional photography, natural color grading, ultra-HD, realistic and elegant, no text, no watermark, no distortion.
02 / 复刻路径
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。
- 主体:A detailed prompt for creating a photorealistic portrait of a woman based on a reference image, focusing on natural lighting and professional camera settings.
- 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
- 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
- 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