Original document
GPT Image 2 · en
Create one vertical 3:4 diptych for each uploaded photo, divided into equal top and bottom halves with a seamless join. No divider, frame, or collage. Swiss editorial design, experimental photography, and vintage print aesthetics.
TOP HALF — ORIGINAL PHOTOGRAPH
Preserve the original identity, pose, scene, composition, perspective, lighting, and colors. Allow only subtle color grading and background extension. No text.
BOTTOM HALF — SEVEN EXPERIMENTAL TEXTURE ZONES
Repeat the entire original photo. Overlay seven asymmetrical rectangular zones of different sizes directly onto the image pixels, crossing non-critical details. Maintain visual continuity inside and outside every zone. Never crop, zoom, shift, duplicate, or redraw the image. Keep all important focal features visible.
SEVEN NUMBERED TECHNIQUES
01. Reeded Glass: Vertical optical refraction.
02. Black-and-White Engraving: Fine etched lines and crosshatching.
03. American Stipple: Black-and-white microdot texture.
04. Vintage Halftone: CMYK printing dots.
05. Pixel Mosaic: Square pixel blocks.
06. Transparent Overlay: 2–4 translucent overlapping layers.
07. Duotone Screen Print: Two-color screen-printing effect.
DESIGN DETAILS
Add hairline borders, tiny corner squares, and fine leader lines to identify each rectangular zone. Label every technique with its corresponding number and English name. Keep all areas outside the seven zones photographic and realistic.
STRICT REQUIREMENTS
No logos, watermarks, extra images, or unnecessary decorative elements. Preserve the original photograph's integrity and maintain a clean, sophisticated editorial layout.
OUTPUT: One high-definition 3:4 vertical poster with equal top and bottom halves.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.
