社交媒体帖子 - Tang Dynasty Aesthetic Street Photography Set

唐代美学 · 街头摄影 · 古风人像 · AI摄影

分类：人像摄影

更新于 2026-10-24T15:48:03.446471686Z

提示词

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.

模型：GPT-5.4
综合：7.8 / 10
提示词质量：9.1 / 10
图片表现：8.4 / 10
图文一致：6.3 / 10
提示词专业且画面设计很强，但当前示例图只完成了第一张场景表达，单图质量好、上线可参考，若目标是四张整组发布则一致性与任务完成度明显不足。