AI video comes with a lot of jargon. Here’s what each term means, in one or two lines, with what it looks like in Rift. For the pair people mix up most, see motion transfer vs face swap.
Motion transfer
Movement from one video, put on a different person, pet or character. A dance from a TikTok, done by you from a photo. In Rift: share the video, add a photo, “same moves, it’s me”.
Face swap
Only the face changes; the body, hair and clothes stay the original person’s. Fast, but it often looks pasted on.
Character swap
The whole person is replaced, so face, body, hair and outfit all match. In Rift: “replace the dancer with me”.
Image to video
A still photo starts to move: a smile, a turn, wind, a camera push. In Rift: add a photo, “turn this into a video”.
Text to video
A clip made from words alone. In Rift: tell Spark the scene.
Reference video or reference photo
An example that shows the look, movement or camera you want, without being the thing you remake.
Remake
A new version of a video you saw: same moves, camera and timing, new people or setting. Rift’s main use.
Consistent character
A person or creature that stays the same across videos. In Rift: a character saved from a few photos.
Restyle
The same video or photo in a new look: anime, clay, film, oil paint.
Draft
In Rift, Spark’s plan for your video, with the price on the Make button, before anything is made.