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

高架桥上的都市夜色与Soran三层标记

Urban Night on the Overpass with Soran Three-Tier Mark

3:4竖幅都市夜景环境人像：明确成年东亚女性站在高架人行桥上，人物小于城市环境，身穿哑光黑色无袖长裙，低头内收、手靠近脸侧。冷青灰与深墨绿城市基底点缀远处暖橙白灯，单点透视与金属栏杆清楚可读。画面内三处自然材质标记准确写“Soran”，不使用外贴水印。

高架桥上的都市夜色与Soran三层标记 — AI image creation
高架桥上的都市夜色与Soran三层标记

提示词 / Prompt

中文提示词

创作一幅3:4竖幅写实都市夜景环境人像：a clearly adult East Asian woman, about 25 / 一位明确成年、约25岁的东亚女性站在高架人行桥上，人物比例刻意小于城市环境，位于画面下部中央偏左。服装采用a minimalist matte-black sleeveless long dress with modest full coverage / 极简哑光黑色无袖长裙，覆盖得体。姿态略微内收：肩线收拢、头部低垂，一只手靠近脸侧，另一只手臂松松横抱身体。视线落向地面或手部，不直视镜头；情绪安静、疲倦、沉思并略带不安，不做开朗大笑。面部身份遵循已配置的固定参考图。 人物身后是一条笔直多车道城市主路，向单一消失点延伸。高层办公楼、玻璃金属立面、树木、路灯和车辆在浅层空气虚化中仍保持结构可读，不能糊成纯散景。横向金属栏杆在人物身后形成清晰分界。暖白与克制琥珀色车灯、窗灯沿道路纵深向消失点收缩。 色彩采用cold cyan-gray, deep ink green and charcoal, with restrained amber and warm-white lights / 冷青灰、深墨绿与炭黑，点缀克制琥珀和暖白灯光：大面积冷青灰、深墨绿与炭黑城市基底，仅以少量暖琥珀和暖白灯点提亮。人物肤色保持中性偏暖，与冷色城市轻微分离。采用可信的城市环境光，不加棚拍补光；远处背景最亮、人物中等亮度、城市暗部深沉。黑裙内部保留少量明暗层次。 维持自然夜间摄影质感：中等清晰度、建筑轮廓可辨、轻微高感颗粒和克制的高光晕染；不用HDR、过度锐化或美妆精修。视觉叙事表现辽阔开放城市与人物内收姿态的反差。 将“Soran”准确融入画面内恰好三处真实材质，大小写必须完全一致：第一处是在裙装肩部或腰侧的小型灰银织标或金属牌；第二处是在栏杆划痕或旧贴纸残痕中的隐约字样；第三处是在裙摆缝线或栏杆连接件磨损中的微观字样。第一处清楚可辨，另两处近看可见且拼写正确。不得增加漂浮文字、字幕、角标、重复标志或外贴水印。 避免人物占比过大、直视镜头、愉悦微笑、展开式大动作、复杂裙装装饰、背景糊成纯光斑、赛博朋克、霓虹雨夜、高饱和蓝紫、夜店色、橙色滤镜、戏剧棚拍光、硬轮廓光、解剖畸形、多余肢体、随机文字或拼写错误。

English prompt

Create a photorealistic 3:4 vertical urban-night environmental portrait of a clearly adult East Asian woman, about 25 / 一位明确成年、约25岁的东亚女性 standing on an elevated pedestrian bridge. The figure is deliberately small relative to the city and sits in the lower center-left of the frame. She wears a minimalist matte-black sleeveless long dress with modest full coverage / 极简哑光黑色无袖长裙，覆盖得体. Her posture is slightly folded inward: shoulders drawn in, head lowered, one hand near the side of her face and the other arm loosely across her body. She looks toward the ground or her hand, not at the lens; her mood is quiet, tired, contemplative and slightly uneasy, never smiling broadly. Her facial identity follows the configured reference. Behind her, a straight multilane city road recedes toward a single vanishing point. High-rise office buildings, glass-and-metal facades, trees, streetlights and vehicles remain structurally readable through a shallow atmospheric blur, never reduced to pure bokeh. A horizontal metal bridge railing creates a clear boundary behind her. Warm-white and muted amber headlights and windows contract toward the distant vanishing point. Use cold cyan-gray, deep ink green and charcoal, with restrained amber and warm-white lights / 冷青灰、深墨绿与炭黑，点缀克制琥珀和暖白灯光: a broad cold cyan-gray, deep ink-green and charcoal urban base, with only restrained warm amber and warm-white light points. Keep skin neutral-warm and softly separated from the cool city. Use believable ambient city lighting, no studio fill; the distant background is brightest, the person mid-value and the city shadows deep. Preserve a few tonal layers within the black dress. Maintain natural night-photo texture: medium clarity, recognizable architecture, subtle high-ISO grain and restrained highlight bloom, no HDR, over-sharpening or beauty retouching. The visual story contrasts the vast open city with her inward posture. Integrate exactly three in-scene instances of the word “Soran”, all spelled exactly with this capitalization and attached to real materials: first, a small gray-silver woven label or metal plate on the dress near the shoulder or waist; second, a discreet trace in scratches or old sticker residue on the railing; third, a microscopic stitching or metal-wear detail near the hem or railing joint. Make the first instance clearly legible and the other two subtle but correctly spelled on close inspection. Do not add floating text, subtitles, corner marks, repeated logos or watermarks. Avoid an oversized subject, direct eye contact, cheerful smile, expansive gestures, complex dress ornament, pure-bokeh background, cyberpunk, neon rain, saturated blue-purple, nightclub colors, orange filter, dramatic studio light, hard rim light, distorted anatomy, extra limbs, random text or misspellings.

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