521내 사진이 80년대 다크 판타지 영화 한 장면

Output size 1920x1080 pixels, 16:9 landscape thumbnail. Adapt the vertical layout below to a wide canvas: place all text as one column in the left 55% of the canvas, keeping the same order and size relationships (the zone percentages describe the order within that column), and place the main illustration in the right 45%. Keep the text large enough to read at thumbnail size.

Create exactly one finished image, not a description. Use the requested aspect ratio; default to 16:9. Fill the entire canvas with one uninterrupted cinematic frame, without letterbox bars, borders, panels or typography.

PHOTO RULE: Use the attached photo as the identity reference. Preserve the subject's facial proportions, eye shape, nose, mouth, skin tone, approximate age, hairline and distinctive visible features. Keep the face large enough to recognize, unobscured and sharply readable within the gently softened film image. Change clothing into original, weathered fantasy travel clothes appropriate to the scene. Do not replace the face with a generic attractive hero. For multiple supplied people preserve all identities with clear faces. If no photo is attached, invent an original fictional adult named Haru with no recognizable resemblance to any real person; finish without asking for a photo. The name is internal only and must never appear in the image.

SCENE: "【장면】". Interpret only the selected scene:
- 성: a rain-dark ruined castle gate at blue hour, towering handmade stonework, warm torchlight grazing a weathered cloak, distant miniature turrets disappearing into mist; quiet solemn anticipation.
- 숲: an ancient tangled forest at predawn, damp roots and moss in the foreground, suspended mist and restrained pale green backlight, a half-hidden original carved stone guardian far behind the subject; hushed uncanny wonder.
- 동굴: a cavern with layered mineral walls, an underground pool and a distant original stone arch, one low amber practical lantern contrasting with cold violet reflected light; tense exploratory curiosity.
If scene is empty or unrecognized, use 성. Keep the same cinematic medium across all choices; change setting, lighting, palette and mood rather than output type.

Make this look like an original frame from a 1980s live-action dark fantasy production: tactile practical sets, hand-built miniatures integrated with convincing perspective, sculpted props, rough woven cloth, tarnished metal, damp rock, restrained theatrical haze and motivated practical lighting. Frame the subject at a three-quarter angle with eyes visible, a medium-close foreground presence and strong layered environmental depth. Keep facial identity readable rather than burying it in shadow. Use organic fine 35mm film grain, gentle diffusion, subtle highlight halation and slightly softened deep shadows, with believable skin texture. No modern digital gloss, cartoon outlines, painted illustration, plastic skin, excessive noise or aggressive teal-orange grading. Avoid gore. Create original costume, creature and architecture designs; do not imitate any named movie, actor, character, artist or existing franchise. No brands, logos, signatures, watermarks, captions, letters or numbers. Output the image directly, without explanations.

[Final output format]
Output size 1920x1080 pixels, 16:9 landscape thumbnail. Adapt the vertical layout below to a wide canvas: place all text as one column in the left 55% of the canvas, keeping the same order and size relationships (the zone percentages describe the order within that column), and place the main illustration in the right 45%. Keep the text large enough to read at thumbnail size.
This selected output format takes priority over any fixed canvas size or aspect ratio elsewhere in this template. Adapt the composition, text zones and margins to the selected canvas without stretching or cropping essential content. Device safe areas apply only when a device is selected. Never print these format instructions on the image.