Galerie de prompts IA OpenZhenZhen · gpt-image-2.5 · image

Royaume des miroirs centenaire de Shibuya

Shibuya Century Mirror Realm

Royaume des miroirs centenaire de Shibuya — découvrez cette création avec gpt-image-2.5 sur OpenZhenZhen, son prompt original et un exemple de génération à réutiliser.

Royaume des miroirs centenaire de Shibuya — AI image creation
Royaume des miroirs centenaire de Shibuya

Prompt

Prompt chinois

创作一张4:3电影感超现实写实街景照片：夏雨刚停后的涩谷十字路口。前景沥青上铺着一层宽阔而浅的镜面雨水，清晰映出城市，同时保留可信的细微涟漪。画面中心是一位明确成年的女性，穿铬银色羽绒服，跪在水边俯身伸出指尖触水；她的面部忠实绑定固定参考ref-10a-v1。只在水面倒影中，将同一女性转化为1926年的成年振袖和服艺伎，梳辫发型，并向上伸手，让两个时代的指尖恰好隔着反射表面相触。明确表现为同一人的时代变形倒影，而非两个无关主角。水面反射边界上方是2026年的涩谷：弧形透明OLED广告屏、渗入湿沥青的青色与洋红霓虹、密集雨伞、少量半透明雨衣和发光雨伞无人机；全息屏呈现抽象演出图形，不出现可识别名人或角色。水面倒影下方呈现1926年暖褐色街景：低矮木店与推拉门、克制的手绘日文招牌（仅写“酒”和“呉服”）、人力车、早期方正机动车、纸伞和柔和煤气灯。历史场景必须明确读作倒影与另一时代，绝不是字面上的水下城市。指尖接触处形成清楚的同心涟漪，并逐渐变为发光像素；一把历史纸伞似乎从倒影中升起并转化为透明雨伞无人机，少量上升水滴变成微小LED粒子。上方现代都市使用electric cyan, magenta and cool silver，下方历史倒影使用warm amber and sepia；默认分别为电光青、洋红与冷银色，以及温暖琥珀与褐色。视觉焦点始终是两指相触和时代边界，冷色现代上层与暖色历史倒影清楚区分。表现细腻电影级雨水反光、雾中克制体积光、自然镜头景深和震撼奇异感；除这组刻意的同一人时空倒影外，避免不合理镜面复制。不要额外突出的行人、可读广告文案、指定店招以外的招牌、标志或水印。保持4:3构图，以工具支持的最高细节呈现，不承诺工具不支持的8K像素分辨率。

Prompt anglais

Create a cinematic 4:3 photorealistic surreal street photograph of Shibuya Crossing just after a summer shower. A broad, shallow mirror-like sheet of rainwater covers the foreground asphalt and reflects the city with crisp but physically plausible ripples. Compose the image around ONE clearly adult woman in a chrome-silver puffer jacket kneeling at the water's edge and reaching down with one fingertip. Her face is faithfully identity-bound to fixed reference ref-10a-v1. In the reflection only, transform that same woman's appearance into an adult 1926-era geisha in a furisode kimono and braided hairstyle; she reaches upward so the two versions' fingertips meet exactly at the reflective surface. Treat this as one woman and her transformed reflection, not two unrelated subjects. Above the reflection boundary, show contemporary 2026 Shibuya: curved transparent OLED billboards, cyan and magenta neon bleeding across wet asphalt, dense umbrellas, a few translucent raincoats and glowing umbrella drones, with abstract holographic concert graphics rather than recognizable celebrities or characters. In the water reflection below, reveal 1926 Shibuya as a warm sepia streetscape: low wooden shops with sliding doors, restrained hand-painted Japanese signs reading only “酒” and “呉服”, a rickshaw and an early boxy motor taxi, paper parasols and softly glowing gas lamps. Make the historical scene read unmistakably as a reflection and alternate era, never as a literal underwater city. At the touch point, create clear concentric ripples that shift into luminous pixels; let a single historical paper parasol appear to rise from the reflected scene and transition into a transparent umbrella drone, while a few falling droplets turn into tiny LED-like particles. Use electric cyan, magenta and cool silver for the vivid upper city and warm amber and sepia for the reflected past; defaults: electric cyan, magenta and cool silver above, warm amber and sepia below. Keep the woman's fingertip meeting and the era boundary as the central visual event, with a clean separation between the cool modern upper world and warm historical reflection. Detailed cinematic rain reflections, restrained volumetric light through mist, natural lens depth and emotionally charged wonder; avoid impossible mirror duplication beyond the deliberate one-person time-shift reflection. No extra prominent people, no legible billboard copy, no extra signs beyond the two specified shop signs, no logos or watermark. Preserve 4:3 framing; use the highest supported detail without promising an unsupported 8K pixel resolution.

Voir le modèle API

Découvrir plus de prompts image et vidéo · Générer des images · Générer des vidéos · Créer sur le canevas IA