Make a short song film for a brand or person, where a song written from the brand, brief and references drives the video through audio_reference, with no separate muxing.
woman, hair fastened up, sings to camera through a quiet city in soft overcast light. The same woman in every shot; only her outfit and the place change at each hard cut. Restrained neutral palette, 35mm film grain. 0-4s: She wears shirt, pants, socks and shoes. A street of pale stone buildings. Chest-up shot, her whole face large in frame and turned to camera, camera tracking backward as she walks toward it. lip-syncing precisely to video1. Sings clearly, in an intimate young adult female voice, close-mic and restrained: "Pinstripes walking down the street". Her mouth shapes match each word. 4-8.3s: She wears a dark green turtleneck and a brown suede jacket with jeans. A brick street corner. Chest-up shot, her whole face large in frame and turned to camera. She turns the jacket collar up with one hand. lip-syncing precisely to video1. Sings clearly, in an intimate young adult female voice, close-mic and restrained: "Collar up, the boots in suede". Her mouth shapes match each word. 8.3-13.6s: She wears dress and shoes. A quiet stone courtyard. Chest-up shot, her whole face large in frame, she stands still and faces camera. lip-syncing precisely to video1. Sings clearly, in an intimate young adult female voice, close-mic and restrained: "Maison Urban, stay with me". Her mouth shapes match each word. 13.6-15s: She wears pants, shirt and shoes. A pale empty stairway. Full-body wide shot, she stands still with her mouth closed, holding her gaze on camera. do not add sound-effects match cuts to the beats