錦鯉池旁日式餐廳中的特寫：穿粉紅大衣的女孩俏皮地甩頭髮，髮尾卻掉進醬油碟裡

提示詞

提示詞：
燈光
2000 年代初期電視劇的高調光線：明亮、柔和且修飾效果佳的正面補光照亮兩張臉，沒有強烈陰影，平淡中性的日間光線——不是金色調。

音訊
沒有音樂。她溫暖健談的低聲呢喃貫穿整個鏡頭，他附和的哼聲逐漸有些勉強（"うん…… うん……"），室內環境音、輕柔的錦鯉池水聲，以及她最後明亮的一句「……でしょ？」

正向鎖定條件
2000 年代初期亞洲電視劇質感：強烈柔焦、亮部暈光、輕微光暈、溫暖柔和的色調、細緻的電視電影膠片顆粒。皮膚與牙齒超寫實——每次微笑時都能清楚看見那個小斑點。他的不自在保持禮貌又帶著喜愛，絕不流露厭惡。固定鏡頭；只有焦點轉移是動態變化。

場景背景
傳統日式榻榻米餐廳內，錦鯉池旁的矮桌近景：年輕女子著迷地笑到一半，俏皮地甩動頭髮——髮尾卻直接落進醬油碟裡。2000 年代初期亞洲電視劇。

啟用參考
女孩參考：20 出頭，長長的蜂蜜棕色波浪髮，帶瀏海，戴白色針織貝雷帽，穿亮粉色大衣，先俏皮撩人，接著神情垮下來。她的髮尾是動作關鍵。與參考完全一致。
男子參考：畫外；只能聽見他溫暖的笑聲和一聲輕微的鼻音噴笑。與參考完全一致。
地點參考：有錦鯉池的榻榻米大廳；右側平台上最靠近池子的矮桌。光禿禿的漆面桌上：她那一側附近放著一只淺淺的小醬油碟。背景：柔和溫暖的散景。

首幀與空間調度
首幀：四分之三側面近景——她的頭部和肩膀位於畫面左側，x 40%、y 42%，朝畫面右側對著畫外的約會對象笑；小醬油碟清晰地放在桌面上，位於畫面右下方，x 68%、y 78%。從第一幀開始，她的臉和醬油碟都在畫面中。

格式模式
單一連續鏡頭，4 秒。即時速度。沒有字幕，沒有音樂。

光學
29°短焦望遠鏡頭，攝影機距離 1.3 公尺，景深足以在同一個近景構圖中清楚呈現她的臉和醬油碟。

攝影機
固定在三腳架上，只進行一次緩慢、平順的俯仰移動：鏡頭向下俯仰，跟隨她甩出的頭髮移向醬油碟，在沾到醬油的髮梢上停留，接著向上俯仰回到她的臉，捕捉她意識到狀況的瞬間。節奏溫和、從容，帶有 2000 年代電視劇的步調。不要變焦、不要水平移動、不要手持拍攝。

動作時間安排
0.0s–1.2s：她因為他說了什麼而笑——神采飛揚，滿心歡喜——並俏皮地完整甩動一次頭髮，用優雅的手勢把一側長髮撥過肩膀。
1.2s–2.2s：鏡頭向下俯仰，跟隨落下的頭髮：甩過去的髮尾掠過她的肩膀，直接落進淺淺的醬油碟裡，浸入深色液體後停住；一滴醬油緩緩聚在髮梢。她仍然笑著

LIGHTING
High-key early-2000s television lighting: bright, soft, flattering frontal fill on both faces, no hard shadows, flat neutral daytime — not golden.

AUDIO
No music. Her warm chatty murmur carrying the take, his agreeing hums going slightly strained ("うん…… うん……"), room tone, soft koi water, her bright final "……でしょ？"

POSITIVE LOCKS
Early-2000s Asian broadcast drama texture: heavy soft diffusion, blooming highlights, mild halation, warm muted palette, fine telecine grain. Ultra photoreal skin and teeth — the fleck reads clearly at every smile. His discomfort stays polite and fond, never disgusted. Static frame; focus pulls are the only movement.

SCENE CONTEXT
Close-up at a low table beside a koi pond in a traditional Japanese tatami dining hall: charmed mid-laugh, the young woman does a flirty hair toss — and the ends of her hair land straight into the soy sauce dish. Early-2000s Asian television drama.

ACTIVE REFERENCES
Girl reference: early 20s, LONG wavy honey-brown hair with bangs, white knit beret, hot-pink coat, flirty then deflated. Her hair ends are action-critical. 100% matches the reference.
Man reference: off-frame; only his warm laugh and one tiny nasal snort are heard. 100% matches the reference.
Location reference: tatami hall with koi pond; the low table closest to the pond on the right platform. On the bare lacquered table: a small shallow soy sauce dish near her side. Background: soft warm bokeh.

FIRST FRAME AND SPATIAL BLOCKING
First frame: close-up in three-quarter profile — her head and shoulder frame-left, x 40%, y 42%, laughing toward screen-right at her off-frame date; the small soy dish sharp on the tabletop at lower frame-right, x 68%, y 78%. Both her face and the dish present from frame one.

FORMAT MODE
Single continuous take, 4 seconds. Real-time. No subtitles, no music.

OPTICS
29° short telephoto, camera 1.3 meters back, focus deep enough to hold her face and the dish in the same close composition.

CAMERA
Locked-off tripod base with ONE slow, smooth TILT as the only movement: the frame TILTS DOWN following her swept hair to the soy dish, holds on the dipped tips, then TILTS BACK UP to her face for the realization. Gentle, deliberate, 2000s-drama pace. No zoom, no pan, no handheld.

ACTION TIMING
0.0s–1.2s: she laughs at something he said — bright, flattered — and does a flirty full hair toss, sweeping the long side of her hair over her shoulder with a graceful hand.
1.2s–2.2s: TILT DOWN with the falling hair: the swept ends swing past her shoulder and land directly INTO the shallow soy dish, settling in the dark liquid; one slow drip gathers at the tips. She keeps laughing

• 提示詞為日文：直接貼上即可，模型能理解。