Documento original
GPT Image 2 · en
Create a hyper-realistic vertical smartphone photo of {argument name="character description" default="a young Asian woman"} sitting courtside/in the lower audience section at a packed Los Angeles Lakers NBA game, framed waist-up in a close candid closeup. She is centered in the foreground, seated on a black arena chair, wearing {argument name="team hoodie" default="a black cropped Los Angeles Lakers hoodie with purple-and-gold sleeve stripes and a large number 23 on the front"}, black leather-look pants, and gold accessories including exactly 2 visible jewelry pieces: a thin gold necklace with a round pendant and a dangling gold earring. Her hair is dark, slightly messy, pulled back into a sleek ponytail with loose strands around the ears. Show her posture as calm and focused, hands resting low near her lap, watching the game action off-camera. Include exactly 3 prominent foreground spectators: the main woman in the center, a young man on the left in a purple Lakers hoodie looking left, and a young man on the right in a black Lakers zip hoodie holding a phone; obscure exactly 2 faces with flat rectangular skin-tone privacy blocks, one over the main woman's face and one over the right spectator's face. The background is a densely crowded basketball arena with hundreds of fans, upper seating tiers, bright stadium lights, LED ribbon boards, and readable arena signage such as {argument name="arena signage" default="LOS ANGELES LAKERS"} and a Toyota logo near the scoreboard. Use realistic indoor arena lighting, natural phone-camera perspective, high detail fabric texture, shallow-to-moderate depth of field, slight wide-angle distortion, authentic candid sports-event atmosphere, no glamour studio lighting, no watermark.Cómo usar
- Copia el prompt original completo, incluidos los parámetros.
- Adapta el tema, el texto y la composición a tus necesidades.
- Usa una herramienta externa compatible; comprueba el acceso y el resultado.
Antes de empezar
- Comprueba si necesita una imagen de referencia. La compatibilidad no demuestra qué modelo generó el ejemplo.
- Revisa nombres, cifras y textos; el resultado puede variar entre versiones del modelo.
