RainDance AI trend: plan a two-person scene

Choose two compatible portraits

Use a separate, clear photo for each consenting adult. Similar lighting and camera angles give a video model fewer conflicting details to resolve. Avoid faces covered by sunglasses, hair or a phone.

Decide who appears in each part of the scene before generating. Keep the two references labeled consistently so their identities do not swap between shots.

Build the story shot by shot

Start with a person waiting at a dock. Follow with a closer performance shot, then bring both characters together at the beach. Treat each location and pairing as a separate shot brief instead of asking for every change in one short clip.

Check faces, horizon, clothing and direction of light across the cuts. An edit can join successful takes without repeating the entire generation.

Choose sound at the publishing step

The RainDance format is associated with the song Raindance by Dave and Tems. Use the music options available for your account and region on the platform where you post. A video prompt does not grant music rights.

Want to start with just one photo?

The subway singing brief on our main page needs one subject and one continuous scene. It is a simpler place to prepare your first AI performance.