A girl in a red hanfu with a fan and plum branch stands in a gallery under falling ginkgo leaves and speaks a soft monologue

Prompt

First frame FIRST FRAME（=upload the reference image, lock it at 100%）：

A woman in ancient Chinese attire, wearing a gold buyao hairpin and pearl tassels draped beside her face, a deep crimson outer robe embroidered with gold over a white inner robe,

holding a landscape folding fan and a branch of white plum/cherry blossoms; standing in a corridor with red pillars, golden side light at dusk,

ginkgo leaves drifting down, light reflecting off the wet green stone ground. Her facial features, hairstyle, clothing, accessories,

fan pattern, and composition must remain exactly consistent with the first frame; no deformation or face changes at any point.

Character · expression · action · dialogue（emotional scene, strict lip-sync with dialogue）：

0–2 seconds: She lowers her head and eyes slightly, looking at the folding fan, with a restrained, faint smile at the corners of her mouth.

〔softly · breathy〕“The wind has risen in this corridor again.”

2–4 seconds: Her smile slowly fades. She lifts her head slightly and turns her gaze to the right side of the frame, her expression growing serious.

〔a slight pause in her tone〕“You said… you’d come back to see these leaves cover the steps.”

4–6 seconds: A hint of uncertainty and worry appears in her eyes and expression; her lips move slightly, as if she wants to speak but stops herself—do not furrow her brows.

〔trailing off mid-sentence〕“Now…”（her voice stops, the sentence unfinished）

6–8 seconds: Her gaze slowly drops, quiet and dejected; sadness settles in her eyes, but she does not cry.

〔almost inaudible · lip movement only〕“Never mind.”

8–10 seconds: She lowers her head and gently closes her eyes, the corners of her lips lifting into the faintest self-mocking smile.

〔murmuring to herself · breathy laugh〕“I suppose I’ve waited too long.”

10–12 seconds: She raises her head toward the warm light, keeping her expression restrained, tenderness and hurt in her eyes.

〔soft but steady〕“I don’t blame you. Really.”

12–15 seconds: She looks ahead/to the right side of the frame, her gaze soft and slightly moist, her lips moving faintly as if she wants to speak but cannot,

then falls silent, gazing steadily into the distance; end on a gentle, cool expression with a hint of melancholy.

〔final line, breathy · fading〕“…I only hope you’re well.”

Voice settings: female voice, gentle and cool, with frequent breathiness; slow pace, phrase endings drop slightly;

emotions restrained, never outwardly dramatic, as if murmuring to herself; voice must stay in sync with facial lip movements throughout.

Camera CAMERA：

Locked medium close-up, extremely slow push-in（subtle push-in）, always focused on her face and eyes;

shallow depth of field, red corridor pillars and drifting leaves blurred in the background. One continuous shot, no cuts or transitions.

Atmosphere · lighting：

Golden side/backlight at dusk, soft warm tones; ginkgo leaves/petals drift down naturally in a light breeze,

light reflecting off the wet stone ground, delicate atmospheric depth; the overall mood is gentle, cool, restrained, and melancholic.

CONSTRAINTS（约束）：

- 4K ultra-high definition, 15 seconds; the image, character, clothing, and setting must be 100% consistent with the first frame

- Strict dialogue and lip sync, clear and natural pronunciation with a breathy quality; no fake singing or timing mismatch

- Facial expressions progress through a 7-stage emotional arc, with smooth transitions and no frame jumps or stuttering

- Expressive eyes; no glowing eyes or abnormal pupils

- Wind, falling leaves, tassels, and fan tassel must move according to realistic physics, gently and without exaggeration

- No black screen or transition at the end; end on a natural freeze-frame

- No subtitles, watermarks, symbols, markings, or text

首帧 FIRST FRAME（=上传参考图，100%锁定）：

古风女子，金步摇发钗、珍珠流苏垂面，绛红描金外袍罩白色里衣，

手持山水折扇与一枝白梅/樱枝；站于红柱长廊，黄昏金色侧光，

银杏叶飘落，湿青石地面有反光。人物五官、发型、服饰、配饰、

扇面纹样、构图与首帧完全一致，全程不得变形、不得换脸。

人物·表情·动作·台词（情绪戏，口型与台词严格对齐）：

0-2秒：微微低头垂眸，目光落在折扇上，嘴角噙含蓄淡笑。

〔轻声·气声〕"这长廊的风，又起了。"

2-4秒：笑意缓缓收敛，轻抬头，视线转向画面右侧，神情转认真。

〔语气微顿〕"你说过……要回来看这树叶落满阶。"

4-6秒：眉眼浮起一丝浅淡疑惑与担忧，唇瓣微动，欲言又止——不皱眉。

〔半句收住〕"如今……"（话音停住，未说完）

6-8秒：视线缓缓下沉，安静失落，眼底沉下来，不落泪。

〔几乎无声·只唇形〕"算了。"

8-10秒：低头轻阖双眼，唇角牵起一抹极淡的自嘲式浅笑。

〔自语·气声笑〕"是我等得太久了吧。"

10-12秒：抬头迎向暖光，神情克制，眼里温柔却委屈。

〔轻而稳〕"我不怪你，真的。"

12-15秒：目视前方/画面右侧，眼神柔和微润，唇瓣轻动似欲语还休，

最终归于安静凝视，定格在温柔清冷、略带忧郁。

〔最后一句气声·渐弱〕"……只盼你，安好。"

声音设定：女声，温柔清冷，气声偏多，语速偏慢，尾音轻微下沉；

情绪克制不外放，似自语呢喃，全程贴合面部口型。

镜头 CAMERA：

固定中近景，极缓慢推近（轻微 push-in），始终对准面部与眼神；

浅景深，背景红柱与飘叶虚化。一镜到底，无切换、无转场。

氛围·光线：

黄昏金色侧逆光，暖调柔和；银杏/花瓣随微风自然飘落，

湿石地面反光，空气感细腻；基调温柔清冷、含蓄忧郁。

CONSTRAINTS（约束）：

- 4K 超清，15秒，画面与首帧人物/服饰/场景 100% 一致

- 台词与口型严格同步，发音清晰自然，气声质感，不假唱、不错位

- 表情按 7 段情绪曲线递进，过渡连贯，不跳帧、不卡顿

- 眼神有戏；禁止眼睛发光、禁止瞳孔异常

- 风动、落叶、流苏、扇穗符合真实物理，轻柔不夸张

- 结尾无黑屏、无转场，自然定格

- 无字幕、无水印、无任何标记符号、无文字

• Needs your photo: attach an image to the request — the prompt works with it.