4 scifi books visualized, photorealistic cinem

4 scifi books visualized, photorealistic cinematic diorama, GPT Image 2 vs Nano Banana Pro vs Reve vs Mai Image 2.6

3:4, do this for a non-copyrighted scifi book INPUT ::= [TOPIC] Create a photorealistic cinematic diorama emerging physically from an open book. Infer automatically: HERO_ENTITY TITLE WORLD TITLE_MATERIAL FOREGROUND_EVENT MINIATURE_ACTORS PROP_SET LIGHT_DNA FX_DNA Composition: - open hardcover book occupies the lower majority of the frame - book pages physically transform into WORLD rather than remaining clean paper - TITLE becomes massive freestanding 3D typography embedded within the terrain - TITLE material and construction are inferred from TOPIC - HERO_ENTITY rises behind, above, through, or around the typography at monumental relative scale - FOREGROUND_EVENT unfolds across the front page as a miniature narrative scene - miniature actors, vehicles, tools, vegetation, debris, structures, or props are inferred only when semantically useful - page edges remain visible so the illusion of “world emerging from book” is unmistakable Lighting: cinematic practical lighting inferred from TOPIC plus believable environmental illumination. Rendering: macro practical-effects photography, real book fibers, physical miniature materials, contact shadows, dust and surface imperfections, atmospheric depth, high-end handcrafted diorama realism. No hard-coded title, character, environment, material, action, vehicle, or visual effect.