385착시 3D 모서리 전광판 광고

Generate ONE full-bleed photorealistic street photograph directly, no questions, code, collage or poster layout.
Output exactly one finished 2:3 portrait image, 1024x1536 pixels. Keep all important text inside the central 73% of image height: leave the top 12% and bottom 15% free of essential lettering.
PRODUCT: "【상품】"
SHOP: "【가게 이름】"
TIME: "【시간】"
CITY MOOD: "【도시 분위기】"

[FIXED CORNER LED ILLUSION]
An invented contemporary commercial building at an imaginary intersection. Ground-level camera at the designated corner viewing point, looking upward. A massive L-shaped LED billboard spans TWO perpendicular exterior walls, both planes clearly visible. Real straight dark top/bottom bezel, metal cladding, windows and consistent street vanishing points. The screen dominates the middle two thirds; sky, architecture and street fill all image edges. Small ordinary adults on the sidewalk look up, at plausible real architectural scale.
INSIDE the screen, a deep virtual recessed box spans both panels continuously. One enormous hero product appears to project toward the viewer across the corner and over the VIRTUAL lip. Align virtual floor, side walls, contact shadows, foreshortening and occlusion to the same viewpoint. All digital elements stay within the physical outer screen boundaries. Subtle luminous LED pixel texture distinguishes the digital imagery from the rigid real building. This is a photographed anamorphic LED ad, never a physical giant object floating in the street, miniature workshop or tabletop scene.

[PRODUCT AND PHOTO RULE]
If a product photo is attached, use it as identity reference: preserve recognizable silhouette, proportions, colors, layer count, toppings, materials and distinctive decorations; remove original background and integrate its enlarged likeness into the LED virtual box. Do not reproduce third-party logos, packaging branding or source captions. Preserve permitted owner-supplied shop wording through SHOP only. If no photo is attached, create a convincing fictional product matching PRODUCT without asking for a photo. One hero product only; supporting splash or a few product pieces may establish depth, never introduce a different menu.
Food must look edible and photographic: iced latte has irregular transparent ice, condensation, pale milk and brown espresso marbling; tteokbokki has short thick rice-cake cylinders in uneven directions, viscous red sauce, irregular folded fish cake and fine scallions; cake has natural porous sponge, soft cream and strawberry seeds. No plastic sheen, identical repeated pieces or ornamental food coils. Other products keep plausible material texture.

[TIME AND CITY]
Use TIME for illumination: daylight gives natural shadows; sunset gives warm sky and reflected evening light; neon night gives a dark sky, luminous screen and restrained city reflections. Empty TIME defaults to sunset.
Use CITY MOOD for street setting: commercial avenue means an invented busy modern intersection; rainy street means wet pavement, reflections and pedestrians with umbrellas; snowy plaza means open paving, snow at edges and winter coats. Empty CITY MOOD defaults to a contemporary commercial avenue. Match weather, clothes and street lighting to both inputs. Never copy a real landmark, building, screen operator or advertiser.

[EXACT TEXT]
If SHOP is nonempty, print it exactly ONCE in small plain type inside a corner of the LED scene, at middle image height (between 35% and 70%), not a logo. Check every letter. Empty SHOP means no shop text, badge or blank sign. PRODUCT, TIME and CITY MOOD are visual instructions, never printed captions. Visible text uses Korean; preserve supplied owner wording. No invented slogan, price, promotion, health claim, QR code, watermark, real advertiser or unrelated street lettering. Output one finished photograph requiring no retouching.

[Final output format]
Output exactly one finished 2:3 portrait image, 1024x1536 pixels. Keep all important text inside the central 73% of image height: leave the top 12% and bottom 15% free of essential lettering.
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.