Bulin 47 AI 提示词指南

Use @Video1 as the exact performance reference. Replace ONLY the lead rapper, identified by sunglasses and an open light-pink shirt behind the speaker holder, with the person in @Image1. The source outfit identifies the target only. @Image1 defines the new face, hairstyle, skin tone, body shape, clothing, and accessories. SAME IDENTITY throughout, including turns and reappearance after occlusion. Follow the source gestures, expressions, mouth movements, body rhythm, action order, and timing. Preserve Kevin Wai, the speaker holder, crowd, portable speaker, lighting, setting, framing, and camera motion. Keep the source shot order and foreground occlusions. NO ROLE SWAP. NO DUPLICATE. NO NEW SCENES. No face drift, extra limbs, distorted hands, clothing flicker, or objects merging with bodies. Use a supported 5–10-second segment and 9:16 framing. Handle audio separately unless source-audio retention is supported.

Use @Image1 as the complete appearance reference and @Video1 as the performance reference. Replace only the lead rapper in sunglasses and an open light-pink shirt; keep the foreground speaker holder and crowd unchanged. Transfer the subject’s face, hairstyle, skin tone, proportions, clothing, and visible accessories from @Image1. Do not copy the original pink shirt or sunglasses unless they appear in @Image1. SAME IDENTITY and stable wardrobe throughout turns, arm movements, and occlusions. Follow the source mouth movements, rap gestures, expressions, body rhythm, positions, and timing. Preserve the source lighting, location, portable speaker, framing, camera motion, and shot order. NO ROLE SWAP. NO DUPLICATE. No NEW SCENES. No added props, outfit changes, distorted fingers, clothing flicker, or background drift. Use a supported 5–10-second segment in 9:16; retain or add suitable audio only through supported tools.

Replace ONLY the lead rapper in @Video1 with the original humanoid character from @Image1. The source target wears sunglasses and an open light-pink shirt behind the speaker holder. @Image1 defines the character’s face, hairstyle, silhouette, colors, proportions, costume, and distinctive features. SAME IDENTITY across every frame and after occlusion. Adapt the source rap gestures, expressions, mouth movements, posture, and body rhythm naturally to the character’s anatomy without changing action order or timing. Keep Kevin Wai, the crowd, portable speaker, source location, lighting, camera motion, framing, and shot sequence unchanged. Replace every appearance of the lead, including close views. NO ROLE SWAP. NO DUPLICATE. NO NEW SCENES. No character fusion, extra subjects, distorted hands, costume flicker, or object-body merging. Match a supported 5–10-second excerpt and 9:16 framing. Handle audio separately unless supported.