Restrained photographic diffusion

Start with the right photo.

A sharp internally owned portrait with distinct facial features.

What to expect

Apply restrained optical-diffusion softness to this same portrait, giving bright hair edges and background highlights a gentle warm bloom while keeping the eyes, eyebrows, lips, freckles and hand boundaries readable.

Keep an eye on

Subtle diffusion is intentional. It cannot repair source blur and should not obscure facial identity.

Make it yours, step by step.

1. Adjust the recipe to describe your subject and preferred background.
2. Copy the complete recipe, then open ChatGPT.
3. Upload your photo in ChatGPT and paste the recipe. The short code alone is not an official command.
4. Inspect identity, text and proportions. Ask for a targeted correction if needed.

Choose the change deliberately.

Diffusion should bloom around highlights while leaving the eyes and important edges readable. A gentle treatment can soften the mood without changing facial geometry.

Review and refine the result

If the face becomes blurred, reduce diffusion in the focal area rather than adding global sharpening. Compare freckles, lips and hairline. Glow cannot restore detail missing from an out-of-focus source.

A related direction, and why

/BACKLIGHT — A directional rim creates separation through light placement rather than optical softness throughout highlights.

These are editing suggestions, not additional test observations. Read the practical guide or use the before/after checklist.

Where this effect helps — and its limits.

Apply restrained optical-diffusion softness to this same portrait, giving bright hair edges and background highlights a gentle warm bloom while keeping the eyes, eyebrows, lips, freckles and hand boundaries readable. Subtle diffusion is intentional. It cannot repair source blur and should not obscure facial identity.

Input advice and use cases are editorial guidance, not additional test results.

Notes from recorded tests.

These notes describe the named runs, which may use a different subject from the comparison above. They do not guarantee results on another input or a future model. See our testing method.