Ai character generation japanese books consistency guide

2x2 grid, 16:9, do this for 4 AI inferred characters from non-copyrighted Japanese books:  V₁  CONCEPT COHERENCE (master)   every channel — palette, costume, props, emblem, environment, lighting, expression — 
      derives from and reinforces ONE concept/culture; the emblem echoes the palette; the environment echoes the

      culture/era; the time‑of‑day echoes the mood.   ¬V₁ = arbitrary clashing elements / generic mismatch.

V₂  CHARACTER READ + HIERARCHY   the character is the dominant focal (low‑angle heroic, centered, vertical poster);

      expression conveys the mood; signature face/hair/ornaments recognizable.   ¬V₂ = weak character / off‑mood / lost hierarchy.

V₃  COSTUME + PROP TRUTH   costume layered & culture/era‑faithful, highly detailed (fabric folds, lacquer, metal plates,

      mechanical+ceremonial mix); props authentic to archetype, correctly carried/wielded.   ¬V₃ = generic costume / floating props.

V₄  LIGHT + ATMOSPHERE   lighting (key sunlight + hard rim + glossy highlights + dramatic shadows + lens flare) consistent

      with time‑of‑day and mood; high contrast.   ¬V₄ = inconsistent / flat lighting.

V₅  SYMBOLIC BACKDROP + COMPOSITION   the emblem (rising sun / moon / halo) sits behind the character as a graphic anchor;

      the environment (ruins + landscape) frames and echoes culture/era + age; dynamic perspective.   ¬V₅ = missing/clashing emblem / unrelated environment.

η   HONEST RENDER   painterly cel‑shade + sharp linework + detailed fabric/metal; key‑visual polish (the style kernel).

FORBIDDEN = generic/clashing elements · weak character · wrong props · inconsistent light · missing emblem · off‑style