The photo is the main creative input to a Damon Salvatore church scene AI video. A strong image helps your appearance stay recognizable through wide movement, medium shots, and closer reactions.
The good news: your photo does not have to be square. A vertical portrait can be a better source because it preserves the full person. The uploader shows the image with a contain fit, so you can inspect its complete frame before generating the square video.
The best all-purpose input
For most people, the best first choice is a vertical waist-up portrait with one person, a clearly visible face, and simple lighting. Include the top of the head, both shoulders, and some torso. That balance preserves facial detail while keeping the body information needed by the entrance shot.
If you have a clean full-body portrait, that can also work well. Make sure the person is not so small that the face becomes a few pixels wide.
Portrait, waist-up, or full-body?
Close portrait
A close portrait offers the most facial detail. It can help the reaction shots, but it shows less information about clothing and body shape.
Use it when the face is the priority and the image is sharp.
Waist-up portrait
This is the most balanced option. The face remains large while shoulders, hair length, outfit, and posture are still visible.
Use it for your first generation when several photo types are available.
Full-body photo
A full-body image provides the most context for the entrance. It works best when the person occupies a substantial part of the frame and the original file has enough resolution for the face to stay clear.
Use it when the outfit or silhouette is important.
A five-point photo check
Before uploading, zoom in and check the following.
- One person: group photos make the target identity ambiguous.
- Visible features: eyes, nose, lips, and jaw should be readable.
- Natural light: avoid a face that is half lost in darkness or blown out by flash.
- Clean edges: heavy motion blur, hair across the whole face, or a hand covering the chin removes useful detail.
- Useful scale: the face should remain recognizable when the image is shown at phone size.
Does the expression matter?
A neutral, confident, or lightly smiling expression is usually easier to adapt than an extreme expression. The reference includes its own reaction moment, so a relaxed face works better than a frozen exaggerated pose.
Direct eye contact is not mandatory. A slight three-quarter angle can work, but avoid a full profile for the first attempt because the reference contains changing angles.
Clothing and background
Visible clothing can influence the color and silhouette of the result. For a clean, coherent look, choose an image where the outfit is not hidden behind a large object.
The source background matters less than the person, because the church setting comes from the reference. A simple background still helps by creating a clean edge around hair and shoulders.
Common photo problems and the fastest fix
- The face changes between cuts: use a sharper, more front-facing portrait.
- The body looks generic: try a waist-up or full-body image with a visible outfit.
- Hair shape is inconsistent: choose a photo with clear contrast around the hair.
- The wrong person dominates: crop the source so only one person remains.
- A vertical photo looks cut off in preview: replace or reload it only if the original itself is cropped; the site preview is designed to show the full frame.
A simple test workflow
Upload the strongest waist-up photo first and make a 540p draft. Watch the complete 16.97-second result. If the identity holds across the cuts, generate the 720p final version. If it does not, change the source image before spending more credits on the same one.
Ready to test a portrait? Upload it to the church-scene generator. For the rest of the process, see the complete creation guide and one-photo face-swap explanation.