Pixel Dissolve Landscape

Delivery

• Use the source photo only as a reference for subject, structure, space, direction, and color. Stylize the entire final canvas; do not show the original photo or create split-screen or before-and-after layouts.
• Generate one standalone 3:4 vertical poster per source image. Process multiple images separately; never create a collage.
• If the user asks only for a prompt, return a ready-to-use generation prompt. Otherwise, use the available image generation or editing capability directly. Ask for an upload only when no reference image is available.
• Create a manually selected, color-limited, recomposed pixel landscape—not a mosaic filter or globally downsampled photo.

Fixed visual specification

On warm off-white or pale ivory paper, choose one highly recognizable visual anchor from the source, such as a bridge, facade, dome, vehicle, boat, gate, or roof group. Retain only the minimum elements required to communicate subject identity, place, and spatial relationships.

Preserve subject identity, count, pose, architectural proportions, road direction, occlusions, viewpoint, and perspective. Remove repetitive windows, tile texture, tiny signage, tourist clutter, dense cables, foliage noise, and photographic noise when they are nonessential. Never alter defining structures or invent landmarks, people, vehicles, or scene elements.

• The subject occupies roughly 30%–50% of the canvas width and height, positioned in the lower-middle and offset according to the source direction. Preserve at least 50%–65% complete paper white; imply the background with only two or three sparse pixel layers.
• Rebuild the scene with crisp square or near-square discrete blocks. Use medium blocks for contour, mass, light-dark structure, and major color fields; use smaller, denser blocks at identifying details such as eaves, spires, domes, windows, arches, poses, wheels, or bows.
• Keep the subject's interior relatively complete. Toward architecture edges, vegetation, roads, water, reflections, and distance, progressively reduce, separate, misalign, and remove blocks.
• Dissolution must come only from decreasing the count and density of independent blocks. A few blocks may drift along perspective, water flow, terrain, or motion, but never become random glitch effects.

Extract and compress the source into 6–12 principal color steps. Preserve its 1–3 most identifying colors and use related values to build volume. Lower environmental saturation toward source-appropriate natural gray, limestone, gray-blue, gray-green, wood brown, or architectural tones. One genuinely present saturated detail may act as a visual jump. Paper white must remain the largest color field.

Before generation, write the exact English text from visible facts: line one is a 1–3-word title; line two is a 5–10-word observation about the visible place, subject, direction, or action. Place it on an intact blank edge far from the pixel subject in a clear monospaced or light sans serif, optionally with a slight pixel stair-step. Text may occupy no more than 3%–5% of the canvas. Avoid generic travel slogans and invented places, dates, or coordinates.

Quality boundaries

The subject must remain immediately traceable to the source while its edges disperse block by block, like a travel memory receding from paper. Paper white must exceed the pixel area. Avoid visible source photography, collages, photo mosaics, global downsampling, blurry pixelation, transparency gradients, stretched pixels, smooth digital illustration, 8-bit game interfaces, voxel 3D, Minecraft, cyber glitches, halftone dots, cross-stitch, diamond painting, full-bleed pixel backgrounds, generic travel icons, oversized titles, names, logos, QR codes, URLs, watermarks, and garbled text.