Plain generation draws a new person each time, but a series of clips, a brand or a blog needs one and the same hero. The “Characters” section trains a model on 4-30 photos of one face, and then you shoot that person in any scene.
A trained hero keeps the features across poses, angles and settings, so the frames add up to a coherent series.
A frame with your hero can be animated or passed to “Avatar” so that the person speaks in a clip.
Train on your own photos or on pictures of people who agreed. Other people and public figures must not be used without consent.
From 4 to 30 pictures of one face. The more varied the angles and light, the better the result.
Fast is cheaper, portrait catches facial features more precisely. Both results are used in the same way.
No, the service word is added automatically and is not shown to the user.