AI / ML
Auto Reframe
An AI feature that converts a horizontal (16:9) source video to vertical (9:16) or square (1:1) by tracking the active speaker's face frame-by-frame and cropping to keep them centred.
Auto reframe is the AI process of converting a horizontal video to vertical or square output by frame-by-frame face tracking and intelligent cropping. The challenge is multi-speaker content: a static centre crop loses the face when the conversation shifts to a guest off-centre. Modern reframe systems use active-speaker detection (matching audio energy to face mouth movement) to pick the right speaker per moment, then smooth the crop transitions to avoid jarring jumps. Reframe quality is one of the largest practical differentiators between content-repurposing tools — bad reframe produces unwatchable vertical output with off-centre faces or visible jump cuts.
Related terms