Keep the whole person visible without losing the face
The chair sequence contains both distant body views and singing close-ups. A photo that includes the shoes but reduces the face to a few blurry pixels gives the model little facial information. Use a sufficiently detailed original image, with the person filling most of the frame and a small margin around the body.
The minimum pixel dimensions are an upload threshold, not a promise of likeness. Check that the eyes, mouth, jawline and hairline remain readable when you view the image at its normal size. Do not enlarge an already blurred screenshot and expect new facial detail.
Show the clothes you want the edit to follow
The prompt asks the performer to keep the photographed person’s own appearance, visible clothes and proportions. Show the neckline, sleeves, trousers or skirt and shoes clearly if those details matter to your version.
A headshot supplies little information about the body or lower outfit. Hidden details must be inferred by the model during the wide shots. You do not need to sit on an office chair or imitate the backward fall in the uploaded image.
Avoid ambiguity around the face and body
Use even light, without a strong filter changing skin color or shadows hiding the eyes. Keep hands, phones, sunglasses and loose objects from covering the face. Separate the person from other people and a busy background.
A plain background makes the subject easier to identify, but it does not replace the template’s office. The prepared edit keeps the green room; a different backdrop in your photo is an appearance-reference background, not a custom scene request.
| Photo issue | Better reference |
|---|---|
| Tiny or blurred face | A sharper original with a larger readable subject |
| Body or shoes cut off | A full-body image with space around the person |
| Another person overlaps the subject | One clearly isolated person |
| Eyes hidden by hands or dark glasses | Visible eyes and an unobstructed facial outline |
Check the angles that the photo cannot show
One photo cannot establish every overhead, profile and distant view. After generation, compare a frontal singing close-up, an overhead close-up and a wide chair tip. Watch for changes in hairstyle, face shape, outfit and limb proportions at the cuts.
Changing to 720p increases output resolution but does not add missing views to the reference photo. Improve the upload first, review the quote and judge the complete result rather than only its opening frame.
Use the scene-specific identity checklist · Follow the creation steps
Reviewed against the EasyVideoAI studio on October 30, 2026. Check the live studio for current availability and the final credit quote.