Prompt by metadigital
11 jam yang laluCrowded Underground Subway Candid
Gemini
2.5 Flash[FACIAL & HAIR ANCHOR — GENERATE NEW IMAGE USING UPLOAD ONLY AS REFERENCE]
Generate a completely new cinematic image using the uploaded photo strictly as facial and hairstyle reference — do NOT edit, retain, or reuse the pixel structure, background, pose, or framing of the uploaded photo. Lock the subject's face (eyes, nose, mouth, jaw, skin tone) and hairstyle/hair color/hair texture exactly as shown in the upload — do not impose a different haircut, length, or color than the reference. Seamlessly integrate: match lighting, skin tone, and head angle to the new pose; blend hairline/jaw/neck naturally with zero visible seam or cutout edge.
[MEDIUM & ART STYLE]
Cinematic photography, hyperrealistic editorial film-still quality — modern noir thriller / A24-style character drama. Full-frame camera look, natural depth of field, subtle film grain, true-to-life skin texture and micro-imperfections (avoid an overly clean/AI-polished look).
[SUBJECT & POSE]
The subject stands in three-quarter profile, looking directly back over one shoulder toward the camera with an intense, guarded expression — alive and charged, not stiff or static. Posture still and upright, deliberately calm amid the surrounding motion. Wearing a sharp black suit jacket over a white collared shirt, unbuttoned at the collar, no tie.
[ENVIRONMENT / SETTING]
A crowded underground subway platform at night / rush hour. A silver train with open doors parked beside the platform, side marked with a partial visible car number ending in "...02". Overhead fluorescent strip lighting runs along the ceiling, casting cool, slightly harsh light. Blurred commuters in dark jackets move through foreground and background with motion blur, creating density and urban anonymity — the subject is the only sharp, still point in frame.
[ANATOMY LOCK — per Metadigifun]
Anatomically correct proportions: 5 fingers per hand with natural knuckles, thumb with 2 joints, 1:7 head-to-body ratio, natural shoulder width, arms reaching mid-thigh, 5 toes per foot if visible.
[EXPRESSION UNLOCK]
Face and hairstyle locked to upload; expression is free to match the theme (intense/guarded, over-the-shoulder glance).
[FIX PROTOCOL]
If errors appear in hands, body, or feet: fix only the hands/body/feet — never alter the face or hairstyle.
[COMPOSITION / TECHNICAL]
Aspect ratio 4:5 (portrait editorial film-still crop). Cool cinematic color grade — desaturated teal/blue shadows, warm-white fluorescent highlights on skin. Moderate depth of field: subject tack-sharp, crowd and train motion-blurred.
[NEGATIVE PROMPT]
No distorted or extra fingers, no melted/warped hands, no mismatched or pasted-on face, no visible seam at hairline/jaw/neck, no imposed hairstyle/hair color different from the uploaded reference, no flat/overlay-looking grain or lighting, no plastic/overly smooth AI-rendered skin, no anatomical errors, no warped or unreadable train signage/text, no blurred or distorted face.
Serupa tapi tak sama
Tidak ditemukan