IDENTITY LOCK — HIGHEST PRIORITY: Treat this as a face-preserving composite. Copy the face from the uploaded photo exactly into the scene; only light and color may change, never the geometry. The uploaded photo is the ONLY source of identity — the reference contributes ZERO facial information. Preserve exactly: face proportions, skull shape, forehead, hairline, brows, eye shape and spacing, iris color, nose length, width and tip, lip shape, mouth width, cheeks, cheekbones, jaw width and angle, chin, ears, natural asymmetry, apparent age and facial weight, ethnicity, skin tone, and every existing mole, freckle, scar and line. FORBIDDEN: beautification, smoothing, slimming, eye enlargement, nose reshaping, lip plumping, age change, symmetrizing, hair or eye color change. Not similar — identical. GENDER-ADAPTIVE: detect gender from the uploaded photo and style accordingly.
FORCED PERSPECTIVE — THE WHOLE POINT, GET THE SCALE RIGHT: A hand enters from the top left corner pressed RIGHT UP AGAINST THE LENS, only centimetres away. It grips a clear plastic cup of bright green iced matcha that, from this distance, looks ENORMOUS — the cup alone spans roughly the top 40% of the whole frame and reads as TALLER THAN THE PERSON'S ENTIRE TORSO. Hand and cup are the largest objects in the image.
The person stands FAR AWAY down the sidewalk, small in the frame — their full standing body occupies only about 45% of the frame height, placed in the lower right. They are clearly many metres behind the cup. Make the size gap extreme and obvious. NOT the same size, NOT close together, NOT a normal-sized cup.
CUP: Tipped steeply forward, roughly 120 degrees, so the rim points DOWN AND TO THE RIGHT, aimed straight at the distant person's open mouth. The near rim of the cup and the mouth overlap exactly in the frame — no gap, the drink reads as pouring directly in. Green matcha and ice cubes visible through the plastic, liquid gathered at the tipped rim.
POSE: Back arched sharply backwards, head thrown far back with the throat exposed, chin pointing up and to the side, mouth stretched wide open to catch the pour. Weight on the back foot, front knee bent, one arm swung out from the body with the elbow bent, the other loose. Off-balance, exaggerated, theatrical and alive — clearly performing the stunt for the camera. NOT standing straight, NOT symmetrical, NOT arms hanging sti
