Ideal inputs
Upload one subject image with the foreground visible and explain the desired background, horizon, depth, lighting, and whether the subject must remain centered.
Create a new setting around an uploaded subject while describing the environment, perspective, light, and amount of space around it.
Upload one subject image with the foreground visible and explain the desired background, horizon, depth, lighting, and whether the subject must remain centered.
Choose a source with a readable silhouette and explain which parts of the original scene must remain untouched.
Name the location, surface, time of day, color temperature, perspective, and negative space needed for the composition.
Review the subject edges, scale, contact shadow, reflected light, and background perspective before use.
GPT Image 2 generates a new background direction around the uploaded subject, useful for creative compositing and layout exploration.
The workflow is not a pixel-locked background replacement. Hair, transparent objects, reflections, and subject lighting can change during the edit.
The prompt can request preservation, but image editing may still alter edges, lighting, or facial details. Compare the result with the source.
Mention the location, camera viewpoint, light direction, depth, surfaces, and open space instead of relying on a single style word.