Volviendo a mis estilos de chica influencer y proyecto de modelo, jejee estuve haciendo varias pruebas esta semana y salvo efectos de Seedance que son incomparables este video y prompt fue adaptado para hacerlo con WAN3.0 en 480p y de 20 segundos Lo mejor de todo son los cotos un video de 20 segundos cuesta solo 20 creditos en Higgsfield contra uno de Seedance 2.5 de 20 segundos que tiene un costo de 60 creditos. Esto permite hacer 3 videos con WAN contra uno solo con SEEDANCE Dejo enlace con video y prompt para que les sirva de ejemplo. https://www.instagram.com/reel/DfZ4duVO6Th/?mdxt=MTd1NG0yYTUyZDkweQ== En este caso me encanto el outfit y lo utilice como foto de referencia inicial Start exactly from the first-frame image [@Image 1]. Keep the same woman, the same outfit, the same old garage, the same classic black muscle car, the same golden sunset light from the roll-up door and the same pose as in that image — always. Use [@Image 2] and [@Image 3] ONLY for Nina's face and hair. Her body, outfit and the set come ONLY from [@Image 1] — always. They are NOT outfit references: IGNORE ALL CLOTHING in [@Image 2] and [@Image 3] — never copy, never wear, never show any garment visible in them. The reference sheet contains titles, numbers and labels — IGNORE ALL TEXT in it — always. THE ONLY OUTFIT is the one in [@Image 1] — always. ════ FORMAT ════ 20s — VERTICAL 9:16 — hyper-realistic top-model fashion commercial — the signature photoshoot effects: camera-flash pops with a shutter click, extreme slow motion on the hair, speed ramps, a 360-degree orbit, a dutch angle, crane up and crane down, a dolly zoom, a freeze-pop — ONE continuous take, NO hard cuts — always. The video starts in motion at second 0 — always. The camera works ONLY in full-body and three-quarter body shots (head to knees or full length): NO close-ups of the face, NO zoom into the eyes, NO tight framing of the face at any moment — always.