A melhor versão do prompt dos estádios que nós já vimos

Nos últimos dias, vários influenciadores começaram a aparecer nas transmissões de jogos da NBA, partidas de futebol e até corridas de Fórmula 1. Num primeiro momento, parece que todos eles foram convidados para os mesmos jogos, mas a verdade é que nenhum deles estava lá de verdade.

Todos esses vídeos foram criados com IA usando o passo a passo abaixo.

Passo 1

O processo funciona em duas etapas. Primeiro, você abre o ChatGPT Images, inclui uma foto sua e digita o prompt abaixo:

Ultra-realistic cinematic sports broadcast frame inspired by a FIFA World Cup television transmission, maintaining the recognizable features of the woman from the reference image.

A young Brazilian woman is naturally captured by a live stadium camera during an intense Brazil vs Italy night match. The image should feel like an authentic televised football moment, spontaneous and emotional, rather than a posed portrait or fashion campaign.

She sits among the crowd completely absorbed in the match, unaware of the camera. One leg is folded casually onto the seat, posture relaxed but emotionally tense from the atmosphere of the game. Her expression carries subtle nervous anticipation and emotional involvement, as if reacting to a decisive late-game moment. No direct eye contact with the lens.

Facial realism requirements:

maintain recognizable facial features from the reference image, natural human proportions,

no glamour retouching, no influencer aesthetic, no exaggerated makeup, no enlarged eyes, no skin smoothing, no cosmetic perfection.

Skin and texture should appear authentically human:

visible pores, slight under-eye texture, natural asymmetry, light sweat, humid skin shine from stadium heat, minor imperfections, organic candid appearance.

Wardrobe styling:

casual Brazilian football fan aesthetic, retro Brazil national team jersey or understated neutral clothing with a lightweight jacket, minimal accessories, slightly messy hair affected by humidity, movement, and stadium wind.

Stadium environment:

huge packed crowd during Brazil vs Italy, yellow Brazil shirts mixed with blue Italy jerseys throughout the stands, emotionally reactive fans, people standing, shouting, celebrating, recording with phones, plastic beer cups, flags waving, scarves, security personnel,

constant subtle crowd motion in foreground and background.

Broadcast cinematography:

captured from far across the stadium using a long telephoto broadcast-style lens,

slightly imperfect television framing, soft live-TV focus, realistic broadcast sharpening,

visible compression artifacts, high ISO noise, subtle motion blur, authentic televised sports atmosphere.

Include fictional football broadcast graphics inspired by live sports coverage:

“BRAZIL vs ITALY” scoreboard, match clock around 84:17, tight scoreline suggesting a tense match, small “AO VIVO” live indicator, subtle broadcaster-style logo bug, clean lower-third identifying stadium location.

Lighting and atmosphere:

strong stadium floodlights, humid summer-night air, mixture of cool and warm color temperatures, LED stadium panel reflections on faces, natural contrast and imperfect shadows.

Overall result should feel immersive, candid, emotional, and cinematic, resembling a dramatic televised football moment captured during a major international tournament.”

Passo 2

Depois, você abre o Kling AI, anexa a imagem que o ChatGPT gerou e escreve o prompt abaixo:

Transform the image into a short cinematic multi-shot sequence with realistic live-broadcast energy and natural candid behavior.

Shot 1 — medium broadcast-style framing. She sits casually in the stadium watching the match, posture relaxed and natural. Her eyes subtly follow the action on the field with small realistic head adjustments, gentle breathing, and minimal micro-expressions. The moment should feel completely unposed, like authentic live TV coverage.

Shot 2 — same locked medium framing. Without looking at the camera, she slowly lifts a plastic beer cup to her lips, takes a calm natural sip, then lowers it back down. Her attention casually returns to the game immediately after. Movement should feel effortless and unscripted.

Shot 3 — same medium framing as Shot 1. She continues watching the game quietly, making a slight posture adjustment and subtle eye movement reacting to the match. Calm, candid, absorbed in the moment.

Visual direction for all shots: realistic sports broadcast aesthetic, telephoto TV camera feel, shallow depth of field, softly blurred crowd in the background with subtle ambient spectator movement, consistent stadium lighting, realistic skin texture, natural motion cadence, no exaggerated acting, no cinematic camera movement.

Camera behavior: camera remains completely locked in every shot, no zoom, no handheld shake, no reframing.

Editing style: clean broadcast-style cuts between shots, slow pacing, grounded realism, authentic live sports television atmosphere.

Overall result must feel indistinguishable from genuine candid footage captured during a live televised game.