AI 视频生成器

Action Figure

模型

提示词

Real smartphone footage, 4:3, 10 seconds, one continuous take from a completely static, locked-off phone camera. Natural sync sound only: light city ambience, distant traffic, a faint breeze. The camera never pans, zooms, shakes or cuts.

CAMERA AND FRAMING: the phone sits low, about at the height of the ledge he is sitting on, tilted slightly upward, a couple of meters from him. It is a close, low-angle medium shot: while he sits, his body fills most of the frame from the top of his head near the top edge down to his sneaker at the bottom edge. Wide phone lens, deep focus, everything sharp.

SETTING: a city plaza in warm late-afternoon golden sunlight coming from the left. In the foreground at the bottom of the frame is the flat top of a low wall clad in grey ceramic tiles, the ledge he sits on. Behind the ledge: a strip of green shrubs, a small parking area with a few cars, a young tree on the right, a tall street lamp, and a plain beige-tan stone office building with even rows of windows and air-conditioner units filling the center-left of the background. Behind and to the left of it, a taller blue-green glass office tower; on the right edge, a tall mirrored glass tower reflecting the sky. Above everything, a bright blue sky with a few thin white clouds. Light, sunny, clean, real.

THE MAN: a young East Asian man in his early twenties, slim, with dark hair slicked back and a pair of black sunglasses pushed up on top of his head. He wears an oversized white t-shirt with one big round vintage-style graphic printed on the chest in cream, gold and brown tones, baggy wide-leg trousers in a muted olive-grey, bright yellow socks and white sneakers with gold accents, and a black backpack on his back with both straps over his shoulders. In his right hand he holds a red paper coffee cup; in his left hand a black smartphone. He starts sitting on the tiled ledge, turned slightly toward the camera, with his right knee pulled up and his right foot planted on the ledge, his right elbow resting on that knee holding the cup up near his chin, his left leg hanging down over the front of the ledge, and his eyes on the phone in his lap, relaxed.

THE HAND: a giant real woman's hand with light skin and long, glossy lime-green almond-shaped nails, many times larger than him, so that her thumb alone is about as long as his torso. It enters from the right edge of the frame at the height of his hips, coming in horizontally from the side, fingers open, palm toward him. Only the hand and part of the wrist and forearm are ever visible, always from the right or lower right. Its grip is always the same: thumb pressed across the front of his hips and index finger behind his lower back, pinching him at the waist like a collector holding a figurine.

WHAT THE EFFECT IS: the moment the fingers close on him, the living man becomes a rigid toy action figure of himself. His body locks into one fixed standing pose: legs straight and together, feet flat, back straight, chin level, both arms hanging straight down at his sides, right hand still gripping the red cup, left hand still holding the phone, face blank and neutral, eyes open, looking straight ahead. He keeps exactly this pose for the rest of the clip. He never bends, sways, blinks noticeably, or reacts; every movement of his body is a movement of the whole rigid figure produced by the hand, like a solid plastic toy being lifted, stood up, and turned over in someone's fingers. He is fully solid and real: proper weight, contact shadow when he stands on the ledge, no glow, no transformation effect.

TIMELINE: 0 to 1.2 s, he sits calmly on the ledge looking at his phone, thumb scrolling, the cup held up at his chin; nothing else happens. At 1.2 s the giant fingertips slide in from the right edge at hip height. 1.2 to 2.2 s, the thumb and index finger close on his waist. At 2.2 s he snaps into the rigid figure pose. 2.2 to 3.2 s, the hand lifts him straight up off the ledge, the whole stiff body rising as one piece, tilting slightly as it comes up, his feet leaving the tile. 3.2 to 4.2 s, the hand lowers him and stands him upright on top of the tiled ledge, feet flat, facing the camera, arms at his sides; the fingers stay pinched at his hips, thumb in front. Standing, he now appears smaller and further back, full body visible with headroom. 4.2 to 5.2 s, the hand lifts him off the ledge again and brings him forward toward the camera and up, holding him in the air at eye level so that his head, shoulders and torso fill the frame and his legs run out of the bottom edge, the hand and forearm entering from the lower right, thumb across his hips; his blank face looks straight into the lens. 5.2 to 7.2 s, the hand slowly rotates the rigid figure around his vertical axis, turning him to show his left side and then his back, so the black backpack faces the camera, the red cup still visible at his side. 7.2 to 9.0 s, the hand rotates him back the same way, side view then front, until he faces the camera again. 9.0 to 10 s, the hand holds him still in the air facing the camera, pinched at the waist, his neutral face looking straight into the lens, arms straight at his sides, cup in his right hand.

NEVER: the hand never comes from the top of the frame and never grabs his head, shoulders or chest; he never wears a hat or cap; he never goes limp, never dangles, never bends at the knees or elbows, never raises the phone or the cup after 2.2 s; he never smiles, talks or looks around once rigid; the camera never moves; he never floats or moves on his own without the hand; he never changes size except by moving closer to or further from the camera; no cuts, no text, no graphics.

Photoreal, sharp, real sunlight, real skin and fabric, a playful in-camera scale trick on a real city street.

5796 / 50000

参考图片（可选）

画面比例

时长

分辨率

生成音频2 倍积分Basic+

所需积分50

视频预览

加载中...