Smalltown Boy original avatar duo
The Original avatars setup is for two illustrated characters you created or are allowed to process. Add one complete character reference per slot. The cast instructions ask for their style and costume to remain recognizable; generated results have not yet been validated.
Try the photo setupPrepare an unambiguous character image
Use a clean image with one character and visible facial features. Include the silhouette and costume if those details matter. A model sheet with several poses or expressions gives the model multiple possible subjects in one slot.
Use your own originals rather than treating a search result as permission. Selecting an avatar does not establish rights to a recognizable franchise, character or artist’s work.
Check the background and file
A plain background makes the reference easier for you to inspect. Transparent image regions are flattened onto white during photo preparation, so check the outline against a light surface before adding the file.
The same JPG, PNG and WebP limits apply to drawings: 10 MB per image, at least 300 pixels on the shorter side, and no more than 25 megapixels. A vector source should first be exported to an accepted image format.
- Export one character per image.
- Check lines, colors and the costume at thumbnail size.
- Place the initial reacting character in A and dancer in B.
- Select Original avatars and confirm the two roles.
Use a consistent visual intention
The instructions preserve each reference’s own drawing style. Mixing a flat cartoon with a realistic render is a deliberate experiment; the setup does not promise that both styles will survive in one scene.
Keep extra direction specific, for example a confident final expression. A paragraph introducing a new style or costume can compete with the references you are trying to preserve.
Inspect identity beyond the face
Look at color blocks, silhouette, accessories and costume throughout the switch. Illustrated eyes or mouths can be simplified as the camera moves. A character may look familiar in one frame while losing its defining details later.
This is generative interpretation, not a rigged animation project. You cannot use the studio to place keyframes or lock an exact pose.
Keep the result’s status clear
Photo setup is available; a generation-enabled button and quoted cost will be shown only when the service is ready. There is no verified avatar output to promise at this stage.
If you later generate and share a result, review it against both original drawings. Keep source files and permission records separately from the temporary browser draft.
Your questions, answered
Can I use a character sheet?
Export a single complete character per slot instead of several poses in one image.
Does transparency stay transparent?
The photo preparation step flattens transparent regions onto white.
Is this a rigged animation editor?
No. This workflow uses two image references and a performance video reference.