Корейский Уличный Портрет Высокой Моды

Промпт: Корейский Уличный Портрет Высокой Моды

Готовый промпт на основном языке: Korean High Fashion Street Portrait prompt.

Please generate an intense Korean high-fashion street portrait in a 3:4 vertical aspect ratio.

The subject is a young streetwear model looking directly into the camera on a rainy Seoul night street. Avoid making them look like a specific celebrity or idol; instead, emphasize a natural face, pores, fine skin texture, slight pigmentation differences, and wet hair details.

The outfit consists of {argument name="outfit" default="an oversized silver and gray techwear jacket, black layered pants, and glossy waterproof fabric"}. Raindrops and reflections from cold lights should be visible on the clothing surface.

The scene is a city street like {argument name="location" default="Seongsu-dong, Seoul"} after the rain. Include glass storefronts, wet asphalt, reflecting puddles, chrome railings, cold white signage light, and pale blue neon reflections. Do not include legible text or brand logos.

The composition is a high-angle shot looking down. The camera is about 55 degrees above the subject, who is clearly fixed in the center. Surroundings like people, umbrellas, and scooters should be blurred with a diagonal motion blur, as if captured with a slow shutter.

The lens gives the feel of 35mm fashion editorial photography. The face is in sharp focus, while the background is blurred by shallow depth of field and motion. Let glass panels and reflective surfaces frame the subject's face.

The lighting features {argument name="lighting" default="cold white street lamps, silver rim light from glass shops, and blue reflections from the wet ground"}. The overall atmosphere should be halfway between a Korean fashion magazine cover and viral AI street photography—dramatic, luxurious, and realistic.

Do not include text, watermarks, brand logos, celebrity faces, idol-lookalike faces, anime style, distorted faces, extra fingers, plastic-like skin, or complex backgrounds.