The short answer
Upload one clear, front-facing photo of one adult who agreed to be in the video. Face and shoulders in frame, good light, no heavy filter. JPEG, PNG or WebP up to 10 MB.
The studio already knows the room, the band and the camera moves. The only thing it learns from your photo is the lead's face, so the photo decides most of the result.
What the studio takes from your photo
The desk-concert scene is saved in the preset: the office band, the red keyboard, the headphones and boom microphones, the camera cuts and the new sound. You do not write a prompt and you cannot change the room. Your photo answers one question: whose face is in the lead spot.
That face has to work at two very different sizes. In the wide band shot the lead is small, one person among several players, so the model needs the overall shape of the face and hair to stay recognizable. In the laughing close-up the face fills much of the vertical frame, so the eyes, nose, mouth and skin tone have to hold up at full size. A photo that is fine for one of those shots and weak for the other usually gives a clip where the lead looks like a different person halfway through.
What works and what does not
| Photo | Works? | Why |
|---|---|---|
| Close portrait, facing the camera, even light | Best | The close-up needs a clear face to carry |
| Slightly turned, both eyes visible | Usually | The model can still read the face |
| Sunglasses, hat brim over the eyes | Weak | The eyes get guessed |
| Full body far away | Weak | The face is too small for the close-up |
| Group photo or a couple | Risky | The studio may pick the wrong face for the lead |
| Dark, blurry or heavily filtered | Weak | Skin tone and features get guessed |
A plain studio-style portrait, like the upload example on the homepage, is ideal. It does not need to look like a concert, and the background of your photo does not end up in the clip.
Light, angle and distance
Soft daylight from the front is the easiest light to work with: stand facing a window, not with the window behind you. A bright lamp right above the head throws shadows under the eyes that the model may read as part of the face.
Hold the camera at eye level. A selfie at arm's length with a wide front camera stretches the nose and shrinks the ears; step back a little, or ask someone else to take it, and the proportions come out closer to how the person really looks.
Crop so the face takes up a good part of the picture, roughly from the top of the head to the shoulders. Do not crop so tight that the chin or the hairline is cut off. A neutral face and a smile both work, because the clip animates the expression for the laughing close-up anyway.
Glasses, hair and hats
Everyday clear glasses are usually fine. Tinted or mirrored lenses hide the eyes, and the model then has to invent them, which is where a lead most often stops looking like themselves.
Hair that covers one eye or most of the forehead has the same effect on a smaller scale. A cap is fine if the brim does not shade the eyes; a wide hat or a hood is not. If the person always wears glasses or a hat, a photo with them on is still better than one where the face is half in shadow.
Why only one person
Each spot in the band takes its own photo. A group photo in one spot does not put your friends in the band; it only makes the studio choose one face, and it may choose the wrong one. Crop to one person before you upload.
If you want several friends in the meme, tap their spots on the frame (keys, sax or the back row) and give each spot its own photo. The price stays the same.
File type and size
The upload takes JPEG, PNG or WebP files up to 10 MB. If your phone saves photos as HEIC, share or export the picture as JPEG first. A screenshot of a photo works if it is sharp, but a screenshot of a video call or a photo of a printed picture usually loses too much detail.
Skip beauty filters and AI upscalers before uploading. They smooth the skin and reshape features, and the clip then carries those changes into every shot.
Your right to use it
Upload yourself or an adult who agreed to be in the video. Do not upload a child, a stranger, a coach or a celebrity from an image search, even though the coaches video did. The acceptable-use page has the full rules.
A weak photo costs the same
A clip from a weak photo is still a delivered clip, and it costs the same $9.99 at 480p or $12.99 at 720p as a clip from a good one. Only a task that fails before delivery releases its credits or goes to refund review. A minute spent on the photo is the cheapest way to get a better result.
Before you upload
A quick check saves a paid retry:
- One face per photo
- Eyes, nose and mouth visible
- No heavy beauty filter or tinted glasses
- At least the face and shoulders, nothing cut off
- JPEG, PNG or WebP under 10 MB
- The person agreed to be in the video
The photo
One person, face visible, readable light.
