What is AI reframe?
AI reframe turns landscape footage into vertical clips without losing the subject. Here's how active-speaker tracking works — and why center-crop fails on real podcasts.
What AI reframe does
AI reframe automatically recomposes a landscape (16:9) video into a vertical format like 9:16, choosing the crop so the important subject — usually whoever is speaking — stays in frame. Instead of keyframing a crop by hand for every shot, the tool tracks the subject through the clip and moves the crop to follow. It's the step that makes a horizontal recording look native in a vertical feed rather than letterboxed or awkwardly cut.
Why a center crop breaks on real footage
The naive approach is to crop to the middle of the frame and hope the subject stays there. That works for a single, stationary talking head and falls apart everywhere else: two-person podcasts where speakers sit on opposite sides, panels, wide interview shots, or a facecam over a screen share. The result is a clip framed on empty space between two people, or one that cuts off the person actually talking. On multi-speaker content, a static center crop is the most common reason auto-generated clips look amateur.
Active-speaker tracking
Better reframing follows the conversation. Active-speaker tracking detects who is talking and keeps the crop on them, moving between people as they trade off — the way a human editor would cut between cameras. In a two-person interview, the framing lands on the host for the question and the guest for the answer, so the viewer is always looking at the person speaking instead of a fixed midpoint.
Caption-safe composition
Reframing and captions have to be decided together. A good crop leaves room for burned-in captions and keeps faces away from the frame edge and platform overlays — TikTok's caption bar and side buttons, for instance, cover part of the frame. If the crop ignores this, captions end up under a button or a chin gets clipped. Composing for the caption-safe area is what makes the finished vertical read cleanly on every app.
What to test before you trust a reframe tool
Don't judge reframing on a vendor's chosen demo. Run your own hardest footage through it: a two-person conversation, imperfect lighting, some movement. Then watch the output on a phone and ask three things — does the crop stay on whoever is speaking, do faces and captions stay inside the safe area, and can you re-render a different crop without re-uploading? Reframe quality is the most visible quality signal in an auto-clip, so it's worth the five-minute test.
FAQ
What is AI reframe?
AI reframe automatically crops a landscape video into a vertical format like 9:16, keeping the important subject — usually the active speaker — in frame, so horizontal footage works in a vertical feed without manual keyframing.
Does AI reframe work on two-person podcasts?
Good reframing does — that's its main strength. Active-speaker tracking moves the crop between people as they talk, which is exactly where a static center crop fails. It's worth testing on your own two-shot footage before trusting any tool.
What aspect ratios can AI reframe output?
Vertical 9:16 for Shorts, Reels, and TikTok is the default. The crop is composed to keep the speaker and captions inside each platform's safe area.