The reference scene
The vertical reference shows two performers in a teal studio with a hanging silver microphone. It opens with both performers, cuts to the right performer, moves to a wide shot and returns to the duet. Your Left and Right photos keep their original identity assignments through these camera changes.
The two input photos
Both sample adults were created with GPT Image. The left character has red hair, a cobalt-blue coat and yellow clothing. The right character has silver hair, a purple metallic jacket and black clothing. These photos guide appearance and outfits; the supplied reference guides the studio, actions and timing.
The generated example
The displayed result is a real 14-second Standard 480p video generated with Alibaba Wan 3.0 using these two photos and the supplied original studio video. The soundtrack was added separately. The example represents Standard resolution; results can vary by photo.
What can vary?
AI generation can change facial details, hands, clothing and motion. A 4-second preview helps inspect the opening, but cannot guarantee the later shots of a separately generated full video. Watch your complete result before downloading and sharing.
Try the same workflow
Sign in, upload one adult photo to each of Left and Right, then choose a preview or full video. For a full video, select 480p or 720p and set BGM on or off. Confirm the age and photo-permission requirements, then generate. Follow progress and download the completed MP4 in My videos.