Imagen 4 Ultra E-commerce Product Photography: Perfume Bottle Lighting Practice

A set of product photography prompt examples for e-commerce image generation scenarios, completed using Imagen 4 Ultra's text-to-image capability. The work focuses on "translating studio language into prompts": subject materials (frosted glass, metal cap), light form (softbox key light plus rim light), background and grounding (gradient gray backdrop and soft cast shadow), and lens feel (85mm macro perspective, shallow depth of field) are each described with reusable phrases rather than piled-up adjectives. The input structure is lightweight: prompt carries the full scene description, aspect_ratio is fixed at 1:1 to match e-commerce main image specs, negative_prompt is used to exclude blown-out highlights, cluttered backgrounds, and extraneous text, and seed makes it easy to reproduce and compare the same composition. According to the Imagen 4 documentation, the model accepts only text input (limit of 480 tokens) and can output 1 to 4 images per request, making it suitable to first test compositions with a lower-cost version and then produce final assets with Ultra; although text accuracy within images has improved, manual verification is still recommended. Use cases: e-commerce main images and detail page visuals, brand social media square images, and prompt baselines when A/B testing different lighting schemes.