GPT Image 2.0 提示词 - 奢华法证式产品悬疑广告

AZIZ | AI 🇸🇦 @aziz4ai · 11小时 Use the uploaded product image as the single source of truth. Create one premium 16:9 campaign image where the product is the only suspect in a clean forensic-style scene. Build a refined editorial composition where the product sits calmly as the hero, while 3–5 carefully controlled physical clues around it all point to ONE clear product benefit, behavior, or experience. Do not show the product in action. Do not explain the clues. Let the viewer solve the ad. The scene should feel like a luxury visual investigation: clean, minimal, precise, cinematic, and highly intentional. Every clue must be physically believable and connected to the same invisible event caused by the product. Possible clues may include residue, droplets, crumbs, displaced objects, impressions, scratches, condensation, fibers, folds, liquid traces, reflections, steam marks, dust, pressure marks, or other realistic evidence depending on the product. Preserve the product exactly: shape, proportions, packaging, logo, label, typography, colors, materials, textures, and all visible details. Do not redesign, recolor, distort, duplicate, or invent features. Silently generate multiple evidence-based concepts, evaluate them for visual hook, clarity, originality, realism, and brand fit, then produce only the strongest one. Add the authentic brand logo naturally and create one original headline of two words maximum that becomes more meaningful once the viewer understands the clues. Use bold editorial framing, refined negative space, cinematic lighting, tactile material detail, clean hierarchy, and premium commercial realism. Output one standalone image only. 16:9. No grid, no collage, no split screen, no miniature world, no duplicated product, no police clichés, no evidence-number cards, no arrows, no infographic elements, no mockup presentation, and no random text.