PROMPT RADAR / IMAGE插画三维
个人资料 / 头像 - Candid Low-Light Portrait
Generates a realistic late-night flash portrait of a young woman in a dim room, useful for testing candid photo realism and mood.
@nanyuan0412 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
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。 轮廓语言、材质规则和世界观要互相支持,避免把互不相干的风格词堆在一起。
01 / 原始方法
完整视觉 Prompt
Create an ultra-realistic candid low-light portrait photo of {argument name="character name" default="a young East Asian woman"} sitting indoors at night in a dim, slightly messy room, framed from upper chest to head in a vertical 3:4 composition. She faces the camera directly with a calm, slightly melancholy neutral expression, large glossy dark eyes, soft pink blush, natural lips, and pale smooth skin with subtle pores and tiny beauty marks. Her {argument name="hair color" default="black"} hair is shoulder-length, slightly tousled, with airy curtain bangs and a side ponytail on the viewer's left. Add exactly 2 visible hair accessories: a translucent blue butterfly clip near her left temple and a small silver fishbone hairpin clipped to the left strap area near her shoulder. She wears a pale mint-green spaghetti-strap camisole with thin straps and a simple scoop neckline, bare shoulders and collarbones visible. Pose her with one hand raised under her chin, index finger resting along the jaw and cheek, creating an introspective selfie-like pose. Use harsh direct on-camera flash mixed with dark ambient room light, producing cool bluish shadows, visible film grain, slight underexposure, soft focus imperfections, and an authentic late-night phone-camera snapshot feeling rather than polished studio lighting. Background: a dark apartment bedroom or dorm room with a doorway or wardrobe shadow on the left and a refrigerator or cabinet on the right, indistinct objects on top, all blurred and low contrast. Keep the mood intimate, quiet, and realistic, with no text except a tiny cursive watermark-like signature in the lower right if desired.
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:Generates a realistic late-night flash portrait of a young woman in a dim room, useful for testing candid photo realism and mood.
- 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