How face tracking and reframing work
Why the frame follows the speaker instead of cropping the middle.
Turning 16:9 into 9:16 means throwing away most of the width. A centre crop throws away the wrong part whenever the speaker is not dead centre — which, on almost any real footage, is most of the time.
ClipSpeedAI tracks the speaker and moves the frame to follow them. The movement is smooth and continuous: one frame sliding, never snapping between positions. A frame that jumps is more distracting than a slightly imperfect crop.
On multi-speaker footage the frame follows whoever is talking, or splits the screen so both are readable.
Common questions
What if there is no face in the shot?
The frame follows the subject rather than a face.
Does the frame jump between speakers?
No. Movement is smooth by design.
Related
Choosing a layoutFour layouts, and when each one is the right call.Multi-speaker split screenTwo people, both on screen, switching to whoever is talking.Changing the aspect ratio9:16, 1:1 and 16:9 out of the same source.
Still want to just try it?
Every run starts on the $1 trial. Three days, card up front.
Start for $1 →