The scene and movement
The setting is the Royal Ballroom Waltz reference scene. Formal partners turn across a palace ballroom. Your generation uses the prepared opening 5 seconds, so later moments in a full-length example may fall outside your result.
This preset replaces 2 designated roles with 2 separate reference images. Match the labels to the original clip and keep supporting subjects unchanged.
The role your photo replaces
For this template, your photo defines image 1: partner in the tailcoat; image 2: partner in the gown. Use the editor preview to locate that role before uploading. The photo guides appearance, hair or fur, body proportions, and visible clothing; it does not record a new voice.
Use one photo per labeled role. Subjects without an upload slot keep their reference appearance.
Prepare a useful appearance reference
Upload 2 separate photos, one clear subject per image, in the labeled order. image 1: partner in the tailcoat; image 2: partner in the gown. Do not combine several people into one photo.
Keep one subject clearly separated from the background. For Royal Ballroom Waltz, pay particular attention to foot placement, hands crossing the face, costume edges and spacing from supporting dancers. Include the body and outfit details that will be visible in those moments; avoid a collage or a group photo.
What follows the prepared reference
The camera path, scene order, supporting cast and props come from the prepared clip. The instructions preserve the Royal Ballroom Waltz reference scene and the movement described above. Replacing the appearance does not add separate controls for a new location, a different prop or another choreography.
Review these moments in the result
Play the entire result at normal speed, then pause around foot placement, hands crossing the face, costume edges and spacing from supporting dancers. Check the same identity before and after a turn or overlap. A strong first frame does not guarantee that every later frame is consistent.
Length, credits and sound
The prepared generation is 5 seconds and costs 1 credits. The collected full reference is about 5 seconds. Previews can be longer and are free to play. The quote on Generate applies to the prepared clip, not to every second in the example gallery. Short-template estimates on this guide use 720p; choose 480p in the studio to see its lower quote. Long-template estimates use fixed 480p pricing.
This reference has no original audio track. When no original soundtrack is available, the video is generated with new audio.


