PROMPT RADAR / PROMPT视频叙事
个人资料 / 头像 - Realistic Girl Shower Snapshot
A realistic snapshot-style prompt for generating a photo of a young girl with wet hair and soap foam on her face and hands, mimicking a candid iPhone photo.
@Adam38363368936 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
它最有价值的部分,是把一个模糊想法推进到了可以被模型执行的程度。 把一段描述拆成起始状态、动作变化、镜头响应和结束画面,连续性比形容词更重要。
01 / 原始方法
完整视觉 Prompt
Uses ultra-high-definition visual effects with a realistic style. The subject is a sweet, fresh-faced young girl with {argument name="hair style" default="wet, messy brown long hair"}, wearing a {argument name="top" default="black top"}. Her face and hands are covered in {argument name="detail" default="white soap bubbles"}, with one hand raised to her eyes and the other in front of her, moving playfully. The background is a bathroom scene, and the image mimics a snapshot taken from a mediocre iPhone perspective, featuring motion blur, awkward angles, and poor composition. It includes a clear soft-focus filter effect; the girl is winking slightly with a gentle smile and a playful expression, creating a fresh and vibrant atmosphere with a realistic and vivid texture that conveys a refreshing and joyful mood. Aspect ratio 9:16. Aspect ratio 9:16.
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 realistic snapshot-style prompt for generating a photo of a young girl with wet hair and soap foam on her face and hands, mimicking a candid iPhone photo.
- 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