Pick a photo that shows the face clearly
Use an evenly lit, front-facing portrait with visible eyes, hairline and facial features. Avoid heavy beauty filters, a face partly hidden by another person, or sunglasses that hide the eyes.
Include enough of the body and clothes to establish the intended appearance. A separate photo for each person is easier to assign than a group image.
Keep identity attached to the correct seat
Photo 1 is the central back-seat singer. Photo 2 is the front-left passenger. Photo 3 is the front-right driver. Optional empty slots keep the reference cast.
The timed instructions track each seat across head turns and reactions. The singer is smaller in the dashboard shot, so a tiny face in the uploaded image makes recognition harder.
What follows the photo, and what follows the reference?
The photo supplies appearance. The selected reference supplies the car interior, camera, singing gestures, front-seat reactions, duration and audio. You do not record a new voice.
The preview is a pet-parody reference. A human singer uses the human photo’s identity in the same back-seat position; a pet uses its own species and markings.
Review the result before posting
Check the opening face, middle head turns and ending, not just the poster. Look for face blending, a swapped seat, unstable hands or inconsistent mouth movement.
AI results can vary. This page explains the input and role controls; it does not claim a measured likeness score or a guarantee of perfect lip-sync.
Last checked .