A Tang-style beauty with a red peony in her hair among flowers before a palace in sunset light, with lifelike unposed details

Prompt

FIRST FRAME FIRST FRAME（=upload the reference image, using it only as the starting pose; afterward, the character moves naturally）：

A Tang-style beauty, black hair gathered into a high bun, a red peony tucked beside a gold hairpin with dangling ornaments, a floral huadian mark between her brows,

wearing a thin rice-white gauze outer layer over a deep vermilion-orange, gold-painted moxiong ruqun, both hands naturally folded in front of her；

peony bushes beside her, towers and pavilions of the Forbidden City behind her, warm golden dusk light, fine dust particles drifting in the air。

The character’s facial features, makeup, and clothing remain consistent with the first frame; her body stays naturally relaxed throughout, never stiff.

Core requirement: the feel of a real person’s documentary-style performance, removing every trace of an “AI feel”——

real pores and fine facial hair on the skin, natural sheen and subtle movement from breathing，

micro-expressions with imperfect realism (her gaze occasionally wanders, the corners of her mouth move slightly before she smiles，

an irregular blinking rhythm, with occasional almost imperceptible swallowing / lip-pressing / small shifts of weight).

Storyboard (few lines of dialogue · emphasis on realistic performance details)：

SHOT 1｜0–4 seconds｜medium close-up (handheld camera · extremely subtle breathing sway)

She lowers her head and looks at her folded hands, her fingertips unconsciously and lightly fidgeting with the clothing tie，

shifting her weight from one foot to the other, one shoulder relaxing；

then the peonies beside her draw her attention; she naturally raises her eyes and looks to the side, her gaze arrives first, and only then does her head turn。

No dialogue, only natural breathing and ambient sound (wind, distant birdsong)。

SHOT 2｜4–7 seconds｜facial close-up (expressive eyes)

The camera gently pushes in toward her face: her eyelashes blink irregularly, her pupils make tiny wandering adjustments as they find focus，

her nostrils move almost imperceptibly with her breathing, the corners of her lips press together first, then open into a genuine, faint smile，

the smile begins in her eyes——not a formulaic standard smile。

〔almost a breathy whisper · one line〕"这花……开得真好。"

SHOT 3｜7–11 seconds｜waist-up movement

She raises her hand and lightly touches the peony petal beside her, her fingertip genuinely sinking into the petal’s softness，

her wrist turns naturally, the gauze sleeve slips down slightly, and she instinctively pulls it back up with her other hand，

her body turns slightly with the movement; the wind tousles a loose strand of hair and presses it against her cheek, but she leaves it alone。

The entire frame is filled with real, unintentional small movements。No dialogue。

SHOT 4｜11–15 seconds｜return to medium close-up (resolution · natural hold)

She withdraws her hand, her gaze settling forward again, and lets out a soft breath，

there is a touch of genuine melancholy and relief in her eyes; her lips move slightly as if she wants to say something, but ultimately she says nothing，

finally, a not-quite-perfect yet deeply genuine faint smile appears, and she naturally holds that moment。

〔breathy · fading · as if speaking to herself〕"……也好。"

Voice settings: female voice, natural and conversational, with a breathy quality, speech rhythm uneven and shifting with the emotion，

not overly precise or formal, with realistic pauses for breath；only two extremely brief soft lines throughout。

Camera CAMERA：

Handheld feel (extremely subtle natural sway and breathing sensation), multiple shots：

medium close-up／facial close-up／waist-up／return to medium close-up, natural transitions without showy technique；

realistic shallow depth of field, focus occasionally making an extremely subtle breathing-like refocus, palace architecture and peonies blurred in the background。

Mood · lighting：

realistic natural dusk light, side backlight tracing a golden rim along the hair, warm tones；

airborne dust, petals occasionally drifting down, the gauze outer layer and loose hair moving in a real light breeze，

the overall image should feel like footage shot on film, with fine grain and soft highlights without overexposure。

CONSTRAINTS（constraints）：

- 4K ultra high definition, 15 seconds, the character’s makeup and clothing consistent with the first frame (no face or outfit changes between shots)

- Core: real-person documentary-style performance——realistic skin texture, irregular blinking, natural shifts of weight，

micro-expressions with an “imperfect realism”; stiff, vacant behavior or a plastic fake face are strictly forbidden

- Remove the AI feel: avoid excessive skin smoothing, excessive symmetry, overly smooth movement, and a perfect smile

- The two lines of dialogue and lip movements must be strictly synchronized; the breathy voice must sound natural and conversational, with no lip-sync fakery and no overly precise diction

- Wind movement, falling petals, gauze sleeves, and loose hair must follow real physics, naturally and without exaggeration

- The eyes must be expressive, with genuine wandering focus; no glowing eyes or abnormal pupils

- End by holding in the afterglow of a natural small movement, with no black screen or transition

- No subtitles, no watermark, no symbols or markings of any kind, no text

首帧 FIRST FRAME（=上传参考图，仅作起手姿态，之后人物自然活动）：

唐风美人，乌发高髻、金步摇旁簪一朵红牡丹，眉心花钿，

米白薄纱外披罩绛橙描金抹胸襦裙，双手在身前自然交叠；

身畔牡丹丛、身后紫禁城楼台、黄昏暖金光、空气微尘浮动。

人物五官妆容服饰与首帧一致，肢体全程自然松弛，不僵直。

核心要求：真人纪实表演质感，去掉一切「AI 味」——

皮肤有真实毛孔与细绒、油光与呼吸起伏，

微表情有不完美的真实感（眼神偶尔游移、嘴角先动一下再笑、

眨眼节奏不规律、偶有极轻的吞咽/抿唇/调整重心的小动作）。

分镜脚本（少台词·重真实细节表演）：

SHOT 1｜0-4秒｜中近景（手持镜头·极轻微呼吸晃动）

她低头看着交叠的双手，指尖无意识地轻捻衣带，

重心从一只脚换到另一只，肩膀松一下；

随后被身旁牡丹引去注意，自然抬眼侧望，眼神先到、头才转过去。

无台词，仅自然呼吸与环境声（风、远处鸟鸣）。

SHOT 2｜4-7秒｜面部特写（眼神戏）

镜头轻推到面部：眼睫不规律地眨动，瞳孔有细微对焦的游移，

鼻翼随呼吸极轻翕动，唇角先抿了一下、才漾开一个真实的浅笑，

笑意先从眼睛起来——不是程式化的标准微笑。

〔几乎气声·一句〕"这花……开得真好。"

SHOT 3｜7-11秒｜半身动态

她抬手轻触身旁牡丹花瓣，指腹真实地陷进花瓣的软，

手腕自然转动，纱袖滑下一截，她下意识用另一只手把袖子撩回去，

身子随之微微侧转，碎发被风吹乱、贴在脸颊，她不去理。

满画面都是真实的、不刻意的小动作。无台词。

SHOT 4｜11-15秒｜回到中近景（收束·自然定格）

她收回手，目光重新落回前方，轻轻吐了口气，

眼神里带着一点真实的怅然又释然，嘴唇微动想说什么、终究没说，

最后是一个没那么完美、却很真实的浅笑，自然停在那一刻。

〔气声·渐弱·像自语〕"……也好。"

声音设定：女声，自然生活化，气声偏多，语速随情绪不匀，

不字正腔圆、略带真实的呼吸停顿；全程仅两处极短轻语。

镜头 CAMERA：

手持质感（极轻微自然晃动与呼吸感），多分镜：

中近景／面部特写／半身／回中近景，切换自然不炫技；

真实浅景深、焦点偶有极细微的呼吸式重对焦，背景宫阙牡丹虚化。

氛围·光线：

黄昏真实自然光，侧逆光镶发丝金边，暖调；

空气微尘、花瓣偶尔飘落、纱披与碎发随真实微风动，

整体像胶片实拍，颗粒细腻、高光柔和不过曝。

CONSTRAINTS（约束）：

- 4K 超清，15秒，人物妆容服饰与首帧一致（换镜不换脸不换装）

- 核心：真人纪实表演——真实皮肤质感、不规律眨眼、自然重心转移、

微表情有「不完美的真实感」，严禁僵硬呆滞或塑料假脸

- 去 AI 味：避免过度磨皮、过度对称、过度顺滑的运动与完美微笑

- 两处台词与口型严格同步，气声生活化，不假唱、不字正腔圆

- 风动、落花、纱袖、碎发符合真实物理，自然不夸张

- 眼神有戏、有真实焦点游移；禁止眼神发光、瞳孔异常

- 结尾在自然的小动作余韵里定格，无黑屏无转场

- 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.