Which photos work best for a Rap Duo Video?

Make both faces easy to see

Choose a photo with a sharp face and natural lighting. Avoid sunglasses, hands across the face, extreme angles and strong beauty filters. If the face is tiny in a large group photo, use a clearer portrait instead. The model works from the visual information you provide.

Two photos or one?

Use Two photos when each person has a separate portrait. You can preview and remove each image independently in the homepage studio. The two portraits are combined into one reference image for models using the single-reference workflow.

Use One photo when the picture already shows both people. A simple background and space around both faces help the model distinguish them. Do not assume that an extra person in the background will be ignored.

Match the mood, not every detail

The photos do not need to have been taken in the same place. Similar lighting and camera angles can make the reference easier to interpret, but the scene comes from your prompt and the selected model.

For a Hotel Lobby performance, tell the model to keep both faces visible and describe the microphones, gestures and setting. Avoid asking for a dramatic face transformation while also demanding exact identity preservation.

Preview before uploading

Check that A and B contain the intended people. Replace a blurry image, or remove it and try another. Before signing in, the form displays local previews; choosing a file does not create a video or spend credits.

Use only photos you are authorized to upload. Recognizable faces are a creative goal, not a guaranteed result. If a generation loses important details, simplify the prompt and try a stronger reference rather than adding more competing instructions.