OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image

杂乱室内中的暖色逆光侧颜

Warm Backlit Profile in a Cluttered Room

与已配置身份参考一致的明确成年女性在杂乱室内微微回望；暗色结构外套、凌乱长发的暖金轮廓光、半透明前景与浅景深营造安静怀旧的电影氛围。

杂乱室内中的暖色逆光侧颜 — AI image creation
杂乱室内中的暖色逆光侧颜

提示词 / Prompt

中文提示词

创作一张电影感、亲密安静的人像，主体为与已配置身份参考ref-10a-v1一致的明确成年女性（25岁及以上），置身于类似工作室、储藏室或旧书店的杂乱室内。人物身体侧向镜头之外，头部轻轻越过肩膀回望观者；表情柔和沉思，双唇微启，头部略微倾斜。长发为深棕至黑色，松散微卷、稍显凌乱；暖金色夕阳从后方照亮发梢，在散落发丝边缘形成细腻光晕。服装为暗色、有结构感且带低调纹理的夹克或西装外套，保持不透明、端庄。背景可见模糊木架、纸张、盒子和普通包装物，借助浅景深柔化。仿佛隔着玻璃或透明防尘塑料片观察画面，利用反射形成层次，但不得遮住面部。眼睛清晰对焦，前景与背景化为柔和散景。自然阳光从左侧进入，方向明确却柔和，点缀克制的体积尘埃、轻微光晕和胶片颗粒。戏剧性明暗中保留足够补光以看清面部细节；整体如摄影故事集中偶然捕捉的片段，梦幻、静谧、亲密而怅惘。以日式摄影集构图作为宽泛编辑方向，不模仿特定摄影师。不要额外人物、可见品牌、文字、水印、换脸、性化造型、遮挡面部的浓重阴影或固定相机/镜头参数。

English prompt

Create a cinematic, intimate portrait of the clearly adult woman (25+) matching the configured identity reference, standing in a cluttered indoor space that feels like a working studio, storeroom or old bookshop. Show her in profile with her body angled away and her head gently turned over one shoulder toward the viewer; use a soft contemplative expression, slightly parted lips and a subtle head tilt. Her long dark-brown-to-black hair is loose, wavy and lightly tousled, with warm golden-hour backlight tracing a delicate halo along stray strands. Dress her in a dark structured jacket or blazer with understated texture; keep clothing fully opaque and modest. Place indistinct shelves, paper, boxes and ordinary packaging in the background, softened by shallow depth of field. Shoot as if viewing through a translucent foreground obstruction such as glass or a protective plastic sheet, creating layered reflections without obscuring her face. Keep the eyes in clear focus while foreground and background fall into gentle bokeh. Use natural sunlight entering from the left, directional yet soft, with restrained volumetric dust, mild lens glow and film-like grain. Balance dramatic contrast with enough fill to preserve facial detail; mood is dreamy, quiet, intimate and wistful, like a candid fragment from a photographic storybook. Use Japanese photobook-inspired framing as a broad editorial direction without imitating a specific photographer. No extra people, visible brand, text, watermark, face swap, sexualized styling, harsh obscuring shadow or fixed camera/lens metadata.

查看 API 模型

探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作