GPT Image 2.0 Prompt - Série d’affiches cinématographiques d’objets low-poly

{ "style_rules": { "illustration_style": "low-poly faceted geometric render — every surface broken into flat angular planes, heavy film grain/noise overlay", "lighting": "single diagonal cinematic light split — deep indigo/blue dominant, sharp crimson-red beam cutting across at an angle, faceted highlight on every plane", "subject": "one mundane everyday object or partial human gesture (hand, silhouette) rendered with outsized cinematic weight — no full visible human face", "frame": "full-bleed black rounded-corner border around entire canvas", "header_footer_bars": "identical top and bottom text bars: [LEFT LABEL] ... [CENTER LABEL] ... [NO. XX] — small caps yellow text", "convention_block": "small yellow text block: 'they say:' followed by 2-4 short clichéd instructions, one per line — must sound reasonable/correct", "divider_glyph": "one small typographic symbol (× / * +) between convention block and title", "title": "ultra-heavy condensed yellow sans-serif, 2 lines stacked, left third of frame — title DIRECTLY CONTRADICTS the 'they say' block above it", "corner_annotation": "optional bracket-framed micro-copy in a corner: series number + 2 short imperative phrases", "color_system": "indigo/navy + crimson red + signal yellow only — no other hues", "format": "16:9 horizontal" }, "negative": "photorealistic rendering, full visible human face, warm/bright palette, headline that agrees with the setup text, flat even lighting, more than 3 colors", "variables": { "object_or_scene": "[Mundane Object/Moment]", "convention_lines": "[Convention Lines]", "title": "[Title]", "top_bottom_labels": "[Top Bottom Labels]", "divider_glyph": "[Divider Glyph]" } }