Identity: the face, skin, body and hair are the subject's own and never change. Where a hairstyle is described below, style the subject's own hair into it; never change the hair's colour, length or texture. Apply only the outfit, makeup, hairstyling, pose and setting described below.
Create a surreal composite portrait.
Subject: the colossal subject rising behind a city skyline, scaled hundreds of times life
size so the subject's head alone is larger than the surrounding tower blocks. Bare natural makeup applied to the subject's own face. The subject wears a loose cream silk shirt with a scattered red
floral print, open at the neck.
Action: leaning forward on the elbows with both hands raised to frame the subject's face,
fingertips resting lightly against the subject's face, forearms rising vertically,
elbows hidden below the rooftops. Head tilted very slightly, chin level, calm unsmiling
gaze directly into the lens.
Setting: a dense Japanese city intersection below the subject at normal scale — a wide crossing
with pedestrians and crosswalks, elevated rail line with a passing train, low-rise
commercial buildings with blank signage panels, glass office towers to the right, trees
in the lower left, clear blue sky with light cloud above.
Visual style: photoreal composite photography, bright natural daylight, consistent
perspective and atmospheric haze between figure and city.
Composition: vertical, the subject's head and framing hands filling the upper half, the city
filling the lower half, the skyline crossing roughly at the middle of the frame.
Lighting: bright soft daylight from the front-left, matched exactly between the figure
and the city below, soft shadows, clear sky.
Text inside: none — all signage and billboards are blank.
Constraints: never scale the subject to normal human size, never place the subject standing in or on
the street, never lower the hands from the subject's face, never darken the sky or add night
lighting, never render readable text, logos or brand names on any signage, never change the subject's face, skin or body, and never change the colour, length or texture of the subject's hair — style it only as described.
Output: 3:4 vertical.
