What Is Auto-Reframe?

When you shrink a wide frame into a tall one, you throw away most of the width. The whole problem is deciding *which* slice to keep. A manual crop picks one fixed slice; auto-reframe picks the *right* slice, moment to moment, by following the subject.

How auto-reframe works

The software analyzes each frame to locate the subject — usually via face and body tracking — then positions the vertical crop window to keep that subject framed. As the subject moves, the crop moves with it. Good implementations add stabilization so the movement is smooth, like a camera operator panning, rather than a jittery per-frame jump.

Auto-reframe vs a manual crop

• Manual crop — one fixed rectangle; works only if the subject never moves.
• Auto-reframe — the crop follows the subject, so nobody gets cut off.
• Manual — you keyframe every movement by hand (slow, tedious).
• Auto — the tracking is automatic; you just review the result.

What good auto-reframe handles

The hard cases are where implementations differ: a subject who walks across frame, multiple speakers in an interview, moments with no face at all (b-roll or screens). Strong auto-reframe follows the active speaker, splits the frame when needed, and avoids over-cropping when there's no clear subject.

Why it matters for short-form

Nearly all long-form video is shot in 16:9; nearly all short-form is watched in 9:16. Auto-reframe is the bridge — and doing it well is the difference between a clip that looks shot for vertical and one that looks hacked down from something wider.

In short: auto-reframe converts wide to vertical automatically by tracking the subject, so you get correctly-framed clips without manual keyframing. See smart cropping explained for how the decisions are made.