人像转半调旅行拼贴海报提示词

Saul Goodman @Goodmanprotocol原推发布于 2026.09.13

在 X 上打开原推

📸 Upload your photo → paste the prompt → let AI transform it into a stunning poster. GPT Image 2.5 Prompt: Turn this portrait photo into a single high-end halftone travel-collage poster. Output one poster per uploaded photo — never merge multiple photos into one frame. FORMAT 3:4 vertical canvas. Split into two roughly equal zones — real photo on top, printed collage below — divided by one clean, sharp horizontal line. Do not stretch, warp, or reposition the figure to force the proportions. TOP HALF — THE REAL SHOT Preserve the original photo faithfully: same identity, face, expression, pose, hands, layered clothing, headphones/books/devices/personal items, background structure, natural light and shadow, original color mood. Add only a light high-end color grade and the faintest film grain. Do not redraw or reinterpret the person. BOTTOM HALF — THE PRINTED RECONSTRUCTION Same warm gray old-paper background. Rebuild the same figure as a printed collage: torn photographic slices + fixed-width ASCII character fields + photocopy halftone dots. - Keep face, hands, hairstyle, headphones, layered clothing, and one signature item recognizable — identity must survive the breakdown. - Torn-edge photo fragments carry the face and key structural lines; fabric, shadow, and background dissolve outward into ASCII characters, black halftone dots, and fractured print grain. - Texture: rough white paper-fiber edges, slight print misregistration, old paper stains, a couple of crop marks, one restrained red calibration bar. - Color: warm gray paper + black ink + the photo's own muted original colors + the single red accent. Never fully desaturate. - Figure fills roughly 60–88% of the frame width (scale to whatever reads best for this image) — but always leave 22–38% of the paper untouched and empty, so it reads as archival print, not a packed terminal screen. - Add one readable fixed-width title pulled from the photo's theme, plus the line "REC. STUDY 01" and one short observational caption. A few loose ASCII glyphs can trace the figure's edge as texture only — never used to fake body detail or invent specs. MOOD Early digital systems meets analog photocopy meets underground zine meets modern travel editorial. The figure should feel like it's surfacing — half photograph, half character, half print — restrained, archival, a little experimental. AVOID Solid black terminal backgrounds, walls of dense code, cyberpunk neon, fully hidden faces, random symbols with no meaning, invented technical readouts, garbled fake text, author credits, brand logos, QR codes, watermarks.