aibbywow.comThe best photo for a BbY WOW dance video
A practical checklist for lighting, framing, clothing and one-person photos.
A clear face is more useful than a dramatic pose
A good input gives the model enough information to recognize one person. Use a front-facing or gently turned face, natural skin detail and soft light. A heavy filter changes the features that you want the video to preserve. An extreme side profile leaves the other side of the face to be guessed.
The upload form accepts JPG, PNG and WebP, up to 10 MB and at least 240 × 240 pixels. That minimum is a file check, not a quality recommendation. Prefer an original photo with a clearly visible face over a small screenshot enlarged to pass the size limit.
Frame for the template
For Cowboy, the reference supplies the western outfit. Your photo does not need a cowboy hat. Keep the face unobstructed so the model has a useful identity reference.
For Original Walk-In, show the clothes you want to keep. A head-to-knee or full-body photo supplies more information than a face crop. Avoid a busy foreground or objects covering the torso.
For Bucket Hat, keep the eyes clear. The example's light-blue hat and party outfit describe that reference's look; they are not a guarantee about every output. For Muscle, the body shape changes deliberately, so evaluate face likeness separately from body proportions.
Leave these photos out of the first attempt
- Group photos with several equally prominent faces.
- Faces hidden by sunglasses, hair, hands or deep shadow.
- Screenshots with stickers, captions or another face in a corner.
- Very blurred or heavily compressed social-media downloads.
Change one thing at a time
If the result loses the face during a turn, compare two photos of the same adult while keeping the template and settings fixed. If you change the person, model, prompt and resolution together, you cannot tell what helped. Keep a short note of the input, settings and specific defect. A second attempt is a new generation, so choose a comparison that can answer a real question.
The Dreamina guide also emphasizes a readable identity and body. Those are useful criteria for reviewing an output; they do not promise a successful result from every image.