SCENE CONTEXT
A playful vertical selfie video of two second-year Japanese female university students, best friends, both bubbly and full of energy, in the empty tiered lecture hall of the university they both attend, shown in <<<image_3>>>, after the last class of the day. Recreate the rhythm of the hand-sweep gag from <<<video_1>>>: the video is built on deliberate contrast in three parts. (1) It opens with both girls dead serious, like an ID photo. (2) Then the hands sweep across the front girl's face about five times per second, and at every single pass she appears with a completely different exaggerated expression, flickering by rapid-fire — it starts with silly funny faces and builds into a fierce angry face. (3) In the final second, the act collapses and both friends burst into laughter together.
ACTIVE REFERENCES / REFERENCE USAGE
<<<image_1>>> defines the front character's identity, hair and complete outfit: bleached honey-blonde choppy short hair with dark roots at the parting and wispy bangs, two small kelly-green snap hair clips, white cropped ribbed T-shirt, open cobalt-blue zip-up track jacket with two white stripes down each sleeve, small silver hoop earrings. No hat.
<<<image_2>>> defines the rear character's identity, hair and complete outfit: glossy straight black chin-length bob with blunt bangs, white T-shirt, bright kelly-green chunky ribbed knit cardigan with dark buttons and long sleeves, black wide-leg trousers. No hair clip. The blonde girl's green hair clips match her best friend's green cardigan, and both wear plain white T-shirts, the way close best friends do.
Use the large facial portrait on each sheet for facial identity. Take only their faces, hair and clothing from the sheets; do not inherit the grey background, the flat lighting or the three-view layout.
<<<image_3>>> defines the lecture hall: rows of long light-oak desks with grey fabric seats rising in steps behind them, a stairway aisle on the left, pale cream walls, a wooden double door at the top back, tall windows with pale vertical blinds on screen-right with green campus trees in warm sunlight outside, flat white ceiling lights. Adapt its viewing angle to the selfie camera.
<<<video_1>>> controls only the hand choreography, the sweep timing, the small head dips and the relative placement of the two people. Its performers' faces, hair, clothing, cap, room and facial expressions do not transfer. Any watermark or logo in <<<video_1>>> does not appear.
FIRST FRAME
The video begins directly through the phone's front-facing camera. <<<image_1>>> holds the phone at arm's length at about eye level, with her holding hand and the phone outside the frame. Her face and upper chest fill the lower-centre foreground. <<<image_2>>> is close behind her, slightly offset toward screen-right, her chin almost resting on <<<image_1>>>'s shoulder, relaxed and comfortable like best friends, with her face clearly visible beside <<<image_1>>>'s head. Both look into the lens with dead-serious straight faces, lips pressed together, eyes steady, like an ID photo.
WORLD AND SPATIAL BLOCKING
They sit at one of the long desks in a middle row, facing the front of the hall, <<<image_2>>> just behind and beside <<<image_1>>>; the phone camera looks back at them from the front. Behind their heads, the empty rows of desks and seats rise in steps toward the back of the hall, softly out of focus, with the bright windows and green trees toward screen-right. Keep the hall consistent with <<<image_3>>>; the tight selfie framing crops most of it. No other students anywhere in the hall.
<<<image_1>>> holds the phone throughout. <<<image_2>>> has both hands free to perform the sweeps around <<<image_1>>>'s face. Keep their front-to-back arrangement the whole time.
SHOT FORMAT
One continuous 10-second vertical 9:16 selfie take. The selfie image fills the whole vertical frame, with extra headroom above <<<image_1>>>'s head for the hand movements. No cuts, no external view of them filming.
OPTICS AND CAMERA
Natural front-camera perspective at arm's length, both faces readable. The camera stays almost stationary during the gag, with only slight natural hand drift. During the final laughter, allow a small believable wobble from <<<image_1>>>'s shaking shoulders while keeping both faces in frame. No zoom, no orbit, no reframing.
ACTION / PERFORMANCE TIMING
<<<image_2>>> reproduces the reference's rapid alternating hand sweeps from 0.75s to 9.0s, exchanging the upper and lower hand positions, with the same sweep directions and pace as <<<video_1>>>. She is an energetic girl deliberately holding a poker face: her face stays completely straight until 9.0s, with only her bright eyes showing she is holding back a laugh.
From 0.75s to 9.0s, <<<image_2>>>'s hands pass across <<<image_1>>>'s face about five times per second, every 4 to 5 frames, exactly at the pace of <<<video_1>>>. EVERY single hand pass reveals a different face: her expression changes behind each passing hand, about every 0.2 seconds, so the faces flicker by rapid-fire like a flipbook. Each face is big, clear and fully formed the instant the hand clears. Consecutive faces are opposites in mouth shape and eye opening (closed mouth then open mouth, wide eyes then shut eyes), so every reveal is a jolt.
PART 1 — SERIOUS
0.00–0.75s (frames 0–17): both girls dead serious, looking straight into the lens, not a hint of a smile. <<<image_2>>> raises one open hand above <<<image_1>>>'s head and places the other below her chin.
PART 2 — RAPID-FIRE FACES (starts with funny faces, builds into anger)
Two faces in each half second, one per hand pass, in this order:
0.75–1.25s (frames 18–29): cross-eyed with both cheeks puffed like a balloon → a huge wide-open toothy grin.
1.25–1.75s (frames 30–41): tongue stuck out to one side with one eye shut → eyes and mouth stretched wide in a shocked "O".
1.75–2.25s (frames 42–53): lips pushed far forward and twisted sideways → a wailing cry face, mouth pulled down, eyes squeezed.
2.25–2.75s (frames 54–65): lower jaw jutting out, eyes half-lidded → a sparkly idol smile with a wink.
2.75–3.25s (frames 66–77): nostrils flared, eyebrows shot up to the hairline → a sulky pout with puffed lower lip.
3.25–3.75s (frames 78–89): eyes rolled up to the ceiling, mouth hanging open → a smug, one-sided grin.
3.75–4.25s (frames 90–101): cheeks sucked in like a fish → an excited silent scream, mouth wide.
4.25–4.75s (frames 102–113): eyes closed, head drooping as if dozing → eyes popping wide, startled.
4.75–5.25s (frames 114–125): lips rolled in over the teeth → a big beaming grin.
5.25–5.75s (frames 126–137): an ugly sobbing face → a dreamy, melting bliss with eyes closed.
5.75–6.25s (frames 138–149): cross-eyed with the tongue tip touching her upper lip → nose wrinkled in disgust, "ew".
6.25–6.75s (frames 150–161): cheeks puffed tight → a sharp gasp, eyebrows high.
Now the faces turn from funny to cross, getting angrier with every pass:
6.75–7.25s (frames 162–173): a suspicious side-eye squint → a flat, unimpressed blank stare.
7.25–7.75s (frames 174–185): an exasperated eye-roll → an irritated frown, lips pressed hard.
7.75–8.25s (frames 186–197): narrowed glaring eyes → teeth clenched, a low growl face.
8.25–9.00s (frames 198–215): FULL FURY, the final and strongest face — eyes blazing wide in a fierce glare straight into the lens, eyebrows slammed down and together, nose wrinkled, teeth bared and clenched. She holds this fury through the last few hand passes, and it looks even fiercer after each one.
Negative for 0–9s: no talking, no smiling on <<<image_2>>>, no two consecutive reveals with the same face, no face changing while it is visible, no laughing before 9.0s, n
场景背景
一段充满趣味的竖屏自拍视频,主角是两名日本大二女生,她们是闺蜜,性格活泼、精力充沛。场景位于她们共同就读大学的空旷阶梯教室(如 <<<image_3>>> 所示),时间是当天最后一节课后。重现 <<<video_1>>> 中“手部扫过”桥段的节奏:视频由三个部分的刻意对比构成。(1)开头两人表情严肃,如同证件照。(2)随后手以每秒约五次的频率快速扫过前方女生的脸,每次扫过后她都呈现出完全不同的夸张表情,如翻页般快速切换——从滑稽鬼脸逐渐升级为愤怒凶相。(3)在最后一秒,表演崩塌,两人一起开怀大笑。
活跃引用 / 引用用法
<<<image_1>>> 定义前方角色的身份、发型和完整着装:漂染的蜂蜜金色短发,发根处有深色分界线,留有稀疏刘海;佩戴两个小凯利绿按扣发夹;身穿白色短款罗纹 T 恤;外搭敞开的钴蓝色拉链运动夹克,每只袖子上有两条白色条纹;戴小银色圈状耳环。无帽子。
<<<image_2>>> 定义后方角色的身份、发型和完整着装:光泽感直黑发齐下巴波波头,配齐刘海;身穿白色 T 恤;外搭亮凯利绿粗罗纹针织开衫,配深色纽扣和长袖;下穿黑色阔腿裤。无发夹。金发女生的绿色发夹与她闺蜜的绿色开衫相呼应,两人都穿着纯白 T 恤,正如亲密闺蜜间的穿搭默契。
使用每张图上的大面部肖像确定面部特征。仅提取她们的脸部、头发和服装;不要继承灰色背景、平光照明或三视图布局。
<<<image_3>>> 定义阶梯教室:一排排浅色橡木长桌,后方逐级升高的灰色布艺座椅;左侧有楼梯过道;淡奶油色墙壁;后方顶部有一扇木质双开门;屏幕右侧有高窗,配有淡色垂直百叶窗,窗外可见阳光温暖下的绿色校园树木;天花板装有扁平白色灯具。将其视角调整为自拍相机角度。
<<<video_1>>> 仅控制手部动作编排、扫过时机、轻微低头动作以及两人的相对位置。其演员的脸部、头发、服装、帽子、房间及面部表情均不转移。<<<video_1>>> 中的任何水印或标志不得出现。
首帧
视频直接通过手机前置摄像头开始。<<<image_1>>> 手持手机,手臂伸直,高度约与视线齐平,持机手和手机本身在画面外。她的脸部和上胸部占据下方中央前景。<<<image_2>>> 紧靠在她身后,略微偏向屏幕右侧,下巴几乎搁在 <<<image_1>>> 的肩膀上,姿态放松舒适,宛如闺蜜,脸部清晰可见于 <<<image_1>>> 头部旁。两人直视镜头,表情严肃,嘴唇紧闭,眼神稳定,如同证件照。
世界与空间布局
她们坐在中间排的一张长桌前,面向教室前方,<<<image_2>>> 紧贴在 <<<image_1>>> 身后侧方;手机相机从前方向后拍摄她们。在她们头部后方,空荡的桌椅排成阶梯状向教室后方延伸,柔和虚化,明亮的窗户和绿树位于屏幕右侧。保持教室环境与 <<<image_3>>> 一致;紧凑的自拍构图裁剪了大部分背景。教室内没有其他学生。
<<<image_1>>> 全程手持手机。<<<image_2>>> 双手空闲,可在 <<<image_1>>> 脸周围执行扫过动作。全程保持前后排列关系。
镜头格式
一个连续的 10 秒竖屏 9:16 自拍镜头。自拍图像填满整个竖屏画面,<<<image_1>>> 头部上方留有额外空间以供手部动作展示。无剪辑,无外部拍摄视角。
光学与相机
自然的前置摄像头透视,手臂长度距离,两张脸清晰可读。在桥段期间相机几乎静止,仅有轻微的自然手部漂移。在最后的大笑阶段,允许因 <<<image_1>>> 肩膀抖动产生微小且可信的晃动,同时保持两张脸在画面内。无变焦,无环绕,无重新构图。
动作 / 表演时间轴
<<<image_2>>> 在 0.75s 至 9.0s 期间重现参考视频中快速交替的手部扫过动作,交换上下手位置,扫过方向和速度同 <<<video_1>>>。她是一个精力充沛的女孩,刻意保持面无表情:直到 9.0s 她的脸完全不动,只有明亮的眼睛显示出她在忍笑。
从 0.75s 到 9.0s,<<<image_2>>> 的手以每秒约五次的频率扫过 <<<image_1>>> 的脸,每 4 到 5 帧一次,严格遵循 <<<video_1>>> 的节奏。每一次手部扫过都揭示一张不同的脸:她的表情在每只手扫过后改变,大约每 0.2 秒一次,因此面孔如翻页书般快速闪现。当手移开瞬间,每张脸都大而清晰且完整成型。连续的面孔在嘴形和睁眼程度上形成对立(闭嘴后张嘴,睁眼后闭眼),使每次揭示都带来冲击感。
第一部分 — 严肃
0.00–0.75s(第 0–17 帧):两人表情严肃,直视镜头,毫无笑意。<<<image_2>>> 将一只张开的手举至 <<<image_1>>> 头顶上方,另一只手置于其下巴下方。
第二部分 — 快速变脸(始于鬼脸,渐变为愤怒)
每半秒两张脸,每次手部扫过对应一张,顺序如下:
0.75–1.25s(第 18–29 帧):斗鸡眼,双颊鼓起如气球 → 巨大的露齿咧嘴笑。
1.25–1.75s(第 30–41 帧):舌头伸向一侧,单眼闭合 → 眼睛和嘴巴张大呈惊讶的“O”形。
1.75–2.25s(第 42–53 帧):嘴唇向前撅起并歪向一边 → 嚎啕大哭脸,嘴角下垂,双眼紧闭。
2.25–2.75s(第 54–65 帧):下颚突出,眼皮半垂 → 闪耀的偶像微笑并眨眼。
2.75–3.25s(第 66–77 帧):鼻孔扩张,眉毛挑至发际线 → 撅嘴不满,下唇鼓起。
3.25–3.75s(第 78–89 帧):眼球翻向上方天花板,嘴巴张开 → 得意的单边坏笑。
3.75–4.25s(第 90–101 帧):脸颊吸入如鱼嘴 → 兴奋的无声尖叫,嘴巴大张。
4.25–4.75s(第 102–113 帧):闭眼,头低垂似打盹 → 眼睛突然睁大,受惊状。
4.75–5.25s(第 114–125 帧):嘴唇卷起盖住牙齿 → 灿烂的笑容。
5.25–5.75s(第 126–137 帧):丑陋的哭泣脸 → 梦幻般的陶醉 Bliss,闭眼。
5.75–6.25s(第 138–149 帧):斗鸡眼,舌尖碰上唇 → 厌恶地皱鼻,“恶心”。
6.25–6.75s(第 150–161 帧):脸颊紧紧鼓起 → 尖锐倒吸一口气,眉毛高挑。
现在面孔从滑稽转为恼怒,随着每次扫过变得更加愤怒:
6.75–7.25s(第 162–173 帧):怀疑的斜视眯眼 → 平淡无表情的凝视。
7.25–7.75s(第 174–185 帧):无奈的翻白眼 → 皱眉不悦,嘴唇紧抿。
7.75–8.25s(第 186–197 帧):眯起的怒视双眼 → 咬牙切齿,低沉咆哮脸。
8.25–9.00s(第 198–215 帧):全面暴怒,最后也是最强烈的面孔——双眼圆睁,目光凶狠直视镜头,眉毛压低并聚拢,鼻子皱起,龇牙咧嘴。她在最后几次手部扫过中保持这种暴怒,且每次扫过后显得更为凶猛。
0–9s 负面约束:<<<image_2>>> 不说话,不微笑,连续两次揭示不得为同一张脸,可见状态下脸不得变化,9.0s 前不得大笑,n
Comments & Questions (0)
No comments yet
Be the first to comment!