Hotel Lobby AI動画ジェネレーター

Edit the performance in @Video1 using @Image1 for the LEFT performer and @Image2 for the RIGHT performer. Image 1 is the LEFT subject, Image 2 is the RIGHT subject; these images are separate identity references, NOT first or last frames. Replace exactly these two performers completely: face, hair, skin tone, body appearance, clothing and accessories must come from their respective images. Two adult friends, each retaining the face, hair, clothing and body proportions of their own reference image. @Video1 controls the background, microphone, lighting, framing, camera movement, poses, gestures, interactions and performance timing. Follow its original action order and turn-taking; do not invent an equal-half split, new choreography, a different backdrop or a wider full-body composition. Preserve the reference composition and original camera behavior, whether horizontal or vertical. Keep both identities assigned to the same left/right roles, with natural anatomy, stable clothing and believable contact with the floor. Output 11 seconds to cover the reference clip; for any rounding remainder, hold the final pose without restarting the performance. No merged faces, identity swapping, extra people, duplicate limbs, wardrobe flicker, added titles, logos, cuts or background changes. Source audio policy: If @Video1 contains an audio track, use that original track as the sound reference and preserve its voices, music, timing and ambience alongside the visual edit. Do not mute it or replace it with unrelated audio. If the source has no audio, keep it silent rather than inventing a soundtrack.