E-commerce Product Photography: gpt-image-2 Prompt Breakdown

Use gpt-image-2 to run through a complete "e-commerce-grade product image" generation workflow: break down the control elements of commercial product photography into a reusable prompt structure — subject and material, lens and viewpoint, lighting position and lighting ratio, background and surface, color management and negative space — and reserve usable layout space within the same image so that price and selling-point copy can be overlaid later. The example subject is a frosted glass fragrance bottle, emphasizing highlight transitions, layered transmission and reflection, and the realism of the contact shadow at the base; the image is generated in a 4:5 vertical composition, suited to e-commerce detail pages and mobile placements. The prompt deliberately stays "clearly constrained but not wordy": first declare the overall style of a studio-grade product image, then specify material, lighting position, background, and depth of field one by one, and finally use the negative-space position to explain the space reserved for later layout. This model family performs outstandingly in photographic realism and text rendering, making it suitable for rapid multi-version iteration of key advertising visuals, SKU main images, packaging, and layout-oriented materials; note that high resolution significantly increases time and cost, and manual review of material details and text layout is still recommended before official launch. Applicable scenarios: e-commerce main images and detail pages, social media placement assets, packaging concept drafts, and exploration of product photography style directions.