
Original document
GPT Image 2 · en
Scene:
A square outdoor battlefield image shaped by violent weather: slanting rain, flattened reeds, mud spray, charcoal clouds, and wind carving motion through fabric and grass. The palette should stay restrained in charcoal, wet brown, muted steel, and deep moss, with nature overpowering the people within it. Show a wide flooded plain under hard slant rain, with shallow water rippling around the duel ground and one dead branch rising at the edge.
Subject:
The subject is two lone armored duelists facing each other across open ground, bamboo edges, flooded fields, or broken terrain. Their poses should feel precise and ritualized, with tension coming from stance, distance, and weather rather than from explicit impact. Two armored fighters face each other at measured distance, knees soft and blades low, while a crow waits on the branch.
Important details:
The core language is discipline against elemental force. Preserve historically inspired layered armor, soaked cloth, muddy terrain, reeds bent by wind, and the sense that rain itself is a character. Avoid fantasy powers, impossible weapons, or decorative heroic posing. The emotional weight should come from the rain, the silence before motion, and the tiny human scale inside the weather.
Use case:
Rain Wind Samurai Duel Field single study 1/9.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no readable banners, no bright saturated armor, no fantasy monsters, no gore splatter, no stylized comic effects, no broken anatomy, no low-detail background, no accidental text fragments.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.





