Tang Dynasty Aesthetic Street Photography Set

Full prompt
Facial Identity: Use the attached facial identity as the sole facial reference. The four photos must make the woman immediately recognizable as exactly the same person. Body: Adult female in her early 20s. Use the attached body reference as the sole reference for body proportions, maintaining realistic anatomy, natural curves, balanced proportions, and authentic skin texture. Theme: Tang Dynasty color photography experiment: Night tour of East Market, boating on Qu Lake, Cuju (ancient football) in the district, and spring banquet at Qu River. Subject: Generate four independent vertical 2:3 realistic photos, no collages. Character identity remains consistent, but hairstyle, clothing, actions, camera angles, and photographic language must change reasonably according to their respective scenes. First Photo: 'A Sudden Smile in East Market Lamp Shadows': The character walks through the crowded East Market of Chang'an, hearing a stall owner's joke and turning her face toward her companion; the corners of her eyes curve first, and her smile is not yet fully unfolded. Hairstyle is a Tang-style Fanwan bun, paired with peacock blue brocade inner layer, cinnabar red half-sleeved top, antique gold bead pattern, and misty purple shawl. Uses 35mm color street photography; vermillion lanterns, turquoise wine flags, spice stalls, and crowds create dense color layers, with natural occlusion of passersby and lanterns allowed at the edges. Second Photo: 'A Brief Pause on the Qu Lake Boat': The character sits at the front of a small painted pleasure boat in the palace garden, looking down as she is attracted by a koi leaping beside the boat. Hairstyle is a side-leaning Tang-style Duoma bun, decorated with a white jade comb and light gold flower hairpin. Clothing uses pearl white, light cyan, pale pink, and aqua silk. Uses 80mm medium format-style natural compression, letting the light cyan water surface, pinkish-white palace wall reflections, boat side, and stone bridge form pastel geometry, with large areas of negative space in the environment preserving a quiet afternoon feel. Third Photo: 'Saving the Ball Before It Crosses the Line in Cuju': The character has just saved a Cuju ball with her knee; her body is slightly off-balance, eyes following the ball, and she's laughing out loud. Hair is tied in a tight high Chui bun, wearing a crimson inner layer, turquoise short outer robe, antique gold belt, and dark split-skirt pants. Uses 24mm close-range wide-angle, low waist camera angle, slight upward tilt, and slanted composition. The Cuju ball and knee form moderate perspective effect; arms and hands are cropped out. Low-power daytime fill flash only enhances the clarity of the eyes and fabric. Fourth Photo: 'Looking Up at the Qu River Spring Banquet': The character sits at the edge of a feast by the winding stream, hearing an off-screen friend call her, and lifts her eyes from the direction of a fruit plate. Hairstyle is an upward-expanding Tang-style Paojia bun, adorned with gold leaf step-shakers, white jade flower hairpins, and short pearl chains. Clothing uses rouge red inner layer, pearl white large-sleeved robe, stone green high-waisted woven belt, and pale apricot shawl. Uses a slightly high camera angle, 80mm mid-telephoto with deep depth of field, simultaneously preserving the relationship between the main character, companions, woven mats, pleasure boats, band, and the winding stream. Background characters are shown sharing fruit, talking, listening to music, or looking at the water. Lighting: All character faces use broad and continuous natural light or actual shop light. Water surface, lanterns, and tree shadows should only affect the background and clothing edges, not creating broken light, white circular spots, or color pollution on the forehead, nose, cheeks, or lips. Texture: Four photos respectively present dense color street photography, pastel spatial geometry, close-range wide-angle fill flash, and high-angle social scene photography. Colors are clear but not over-saturated, retaining real grain, slight tilts, edge cropping, and instantaneous motion. The character is beautiful but does not present a plastic AI face; no black and white or unified filters. A low-profile small handwritten signature 'Zhong' is added to the bottom right corner. Negative Prompts: Collage, multi-frame, identity drift, four photos using the same hairstyle, same pose, same camera angle, modern ponytail, cheap ancient costume, static catalog pose, uniform commercial smile, complex hand movements, extra arms, extra fingers, wrong fingers, lantern light spots on face, water ripple light spots, tree shadows across face, fill flash white spots, plastic skin, black and white photo, modern architecture, modern sportswear, garbled text, other signatures.
Original prompt
面部身份: 使用附带的面部身份作为唯一面部参考。四张照片必须让人一眼认出是完全相同的女性。 身体: 20岁出头的成年女性。使用附带的身体参考作为唯一身体比例参考,保持真实解剖结构、自然曲线、平衡比例和真实皮肤纹理。 主题: 盛唐彩色摄影实验:东市夜游、曲池游船、坊间蹴鞠、曲江春宴。 主体: 分别生成四张独立的纵向2:3写实照片,不要拼图。人物身份保持一致,但发型、服装、动作、机位和摄影语言必须根据各自场景合理变化。 第一张「东市灯影里的突然一笑」: 人物穿过长安东市拥挤灯市,听见摊主玩笑后侧脸转向同行者,眼角先弯,笑意尚未完全展开。发型为唐代反绾髻,搭配孔雀蓝织锦内层、朱砂红半臂衣、旧金联珠纹与烟紫披帛。采用35mm彩色街头摄影,朱红灯笼、青绿酒旗、香料摊和人流形成密集颜色层次,边缘允许路人与灯笼自然遮挡。 第二张「曲池游船的短暂停顿」: 人物坐在宫苑窄小彩绘游船前部,被船边跃起的锦鲤吸引而低头。发型为偏向一侧的唐代堕马髻,佩戴白玉梳与浅金花簪。服装使用珍珠白、淡青、浅粉和水蓝色绢罗。采用80mm中画幅式自然压缩,让浅青水面、粉白宫墙倒影、船舷与石桥形成粉彩几何,大面积环境余白保留安静午后感。 第三张「蹴鞠越线前的一次救球」: 人物刚以膝侧救回即将越线的蹴鞠球,身体略失平衡,眼睛追着球并笑出声。头发梳成紧凑高椎髻,穿绛红内层、青蓝短外袍、旧金腰带与深色分裁裙裤。采用24mm近距广角、腰部低机位、轻微上仰和倾斜构图。蹴鞠球与膝部形成适度近大远小,双臂与双手裁在画外。日间低功率补闪只增强眼神和衣料清晰度。 第四张「曲江春宴的一次抬头」: 人物坐在曲水宴席边缘,听见画外朋友叫她,从果盘方向抬起眼睛。发型为向上展开的唐代抛家髻,佩戴金叶步摇、白玉花钗与短珍珠链。服装使用胭脂红内层、珍珠白大袖衫、石青高腰织带与淡杏披帛。采用略高机位、80mm中焦与较深景深,同时保留主人物、同席女伴、织席、游船、乐队与曲水关系。背景人物分别分果、交谈、听乐或看水面。 光线: 所有人物面部使用宽而连续的自然光或店铺实际光照明。水面、灯笼和树影只能作用于背景与衣料边缘,不得在额头、鼻梁、脸颊和嘴唇制造碎光、圆形白斑或彩色污染。 质感: 四张分别呈现密集彩色街拍、粉彩空间几何、近距广角补闪和高位社会场景摄影。色彩明净但不过度饱和,保留真实颗粒、轻微倾斜、边缘裁切和瞬间动作。人物漂亮但不呈现塑料AI脸,不使用黑白或统一滤镜。 画面右下角加入低调小号手写签名「Zhong」。 负面提示: 拼图,多分镜,身份漂移,四张使用相同发型,相同姿势,相同机位,现代马尾,廉价古装,静态目录摆拍,统一商业微笑,复杂手部动作,多余手臂,多余手指,错误手指,脸部灯笼光斑,水波光斑,树影跨脸,补闪白斑,塑料皮肤,黑白照片,现代建筑,现代运动服,文字乱码,其他签名。
Use this prompt in the VdoBloom image editor.
Opens the image editor with this prompt — add your own photo to generate.