PROMPT RADAR / PROMPT视频叙事
个人资料 / 头像 - Hyper-Realistic Eye Detail Portrait
A close-up portrait of a woman with green-hazel eyes, emphasizing realistic skin texture and golden hour lighting.
@jzaib4269 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。 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
Create an insanely photorealistic, stop-scrolling vertical 3:4 close-up portrait designed to look like a once-in-a-lifetime photograph captured spontaneously on a smartphone during peak golden hour. The frame is intensely intimate, filling almost the entire composition with the face of a {argument name="subject" default="beautiful young adult woman"}, with the viewer immediately drawn into her {argument name="eye color" default="mesmerizing green-hazel eye"} as the unmistakable focal point.
Her eye should be extraordinarily detailed—individual iris fibers, subtle color variations, realistic reflections of the setting sun, naturally separated eyelashes, and tiny imperfections that make the gaze feel genuinely alive. Her defined natural eyebrow, softly sculpted nose, and full rosy-red lips remain visible within the tight crop, while long dark-brown hair naturally sweeps across her face, with dozens of delicate flyaway strands catching the sunlight and crossing her forehead, cheek, and nose.
Use dramatic, low-angle {argument name="lighting" default="sunset light"} pouring directly across her face, creating an unforgettable molten-gold/orange glow with deep cinematic micro-shadows and brilliant highlights along the eye, cheekbone, nose, and individual strands of hair. Preserve real human skin texture—visible pores, fine peach fuzz, subtle tonal variation, tiny natural imperfections, and realistic facial detail—avoiding plastic skin, excessive retouching, or an artificial AI appearance.
The background should dissolve into an extreme creamy bokeh, with only vague warm-gray and earthy outdoor textures remaining, creating powerful separation from her face. Add subtle smartphone-camera characteristics: natural HDR, slight lens softness around the edges, authentic exposure falloff, delicate film grain, realistic optical imperfections, and a tiny amount of motion in loose hair strands.
Visual goal: the image should feel accidentally perfect—raw, intimate, mysterious, cinematic, emotionally captivating, and so photorealistic that viewers might initially mistake it for a real viral photograph. Maximum realism, breathtaking eye detail, natural beauty, golden-hour atmosphere, editorial photography, cinematic color science, 85mm lens aesthetic, f/1.8 depth of field, ultra-high detail, 8K realism.
Negative prompt: plastic skin, waxy face, excessive makeup, beauty filter, over-smoothed skin, CGI appearance, cartoon, illustration, distorted eyes, asymmetrical facial features, extra eyelashes, malformed hair, unrealistic reflections, oversaturated colors, excessive sharpening, artificial HDR, blurry eye, low detail, duplicate facial features.
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 close-up portrait of a woman with green-hazel eyes, emphasizing realistic skin texture and golden hour lighting.
- 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