2x2 grid, 16:9, 4 AI inferred dystopian subjects,topics: // Canvas grid 600x800 (vertical panel). Each cell encodes ink level + style ownership. // 1. Photo ink-ification: // - For each cell, compute luminance from [Location Plate] photo. // - L > 200 -> paper-white (0). L < 80 -> solid black (1). Else -> mid-tone (2). // - Mid-tone cells are patterned with a halftone dot array at declared LPI. // - Edge detection: cells where luminance gradient exceeds threshold become mechanical line (3) — uniform thickness, no taper. // 2. Character rendering: // - Using [Char Seed], place a small figure (under 25% cell area) in the scene. // - The figure cells are drawn in the dominant style: line owned by dominant, flats by dominant, light by dominant. // - Subordinate style may appear in background elements (props, textures) but never on character face cells (defined by a mask). If dissolved, subordinate elements are redrawn with dominant line. If declared, boundary cells between styles are explicitly a different color. If oscillating, subordinate only in depth planes behind the character (assigned by depth map). // - Silhouette anchor: hair shape + accessory preserved in pure black at any scale. // - Color key: accent color appears only in eye cells and one prop cell cluster. // 3. Interaction: // - For tourist: character's contact cells have a shadow cluster offset according to sun vector (e.g., shadow cells 5 pixels down-right). // - For fugitive: mid-tone clutter cells from the photo are placed in front of the character partially (overlap). // - For haunting: no shadow, no interaction. // 4. Grain-and-artifact: a layer of noise cells (JPEG artifacts, dust) is added only to photo-derived cells, not character. // 5. Lens distortion: at edges of the canvas, character cells are slightly displaced to simulate barrel distortion matching the photo. // Render final cell grid as high-res ink-ified image.