Short-Form Video4 min readLast updated September 2026

How to Turn Horizontal Video Into Vertical Without Bad Cropping

Face tracking, subject centering, smart aspect reframing, and safe zone guidelines.

Quick Takeaway

Convert 16:9 landscape to 9:16 vertical using AI face tracking so speakers stay centered in frame without awkward head cutoffs.

1. Why a fixed crop fails

Cropping a 16:9 recording to 9:16 throws away about three quarters of the horizontal image. If the crop window is fixed to the centre, anything that happens outside it is simply gone — which is fine while the speaker sits still and disastrous the moment they lean, gesture or move. The classic result is a video where the subject's head slides to the edge of frame and then out of it, and it is entirely a framing problem rather than a resolution one.

  • A centred crop assumes the subject stays centred. Most real footage does not.
  • Gestures, whiteboards and screen content on one side are lost entirely.
  • Two-person footage almost never survives a fixed crop.

2. Three ways to fill a vertical frame

There is more than one right answer, and the choice depends on what is in the shot. Filling crops to cover the frame, which is what most short-form wants. Fitting keeps the entire image and pads around it, which nobody should use full-screen but is occasionally right. Framing places the whole shot inside an inset card, which is a deliberate look rather than a compromise. And when the subject moves, reframing tracks them so the crop follows.

  • Fill: crops to cover. The default for full-bleed vertical short-form.
  • Reframe: the crop follows the subject, which is what a moving talking head needs.
  • Frame: the whole shot sits in a rounded inset card — a styled look, often paired with a headline above it.
  • Fit: keeps everything, pads the rest. Use sparingly.
Fill plus subject tracking
"Convert this to 9:16 and keep the speaker in frame even when they move."

3. Mobile safe zones

Every vertical platform draws its own interface over your video: a caption and handle along the bottom, action buttons up the right side, and a progress bar at the very bottom edge. Anything you place in those regions is either covered or hard to read. Keep faces, captions and any on-screen text within the central band of the frame, well clear of the bottom strip and the right-hand column.

  • Bottom: reserved for the platform's own caption, handle and progress bar.
  • Right edge: action buttons on TikTok, Reels and Shorts.
  • Top: usually safer, but the very top edge can be cropped in some viewers.
  • Captions belong in the middle band, not near the bottom edge where desktop habits put them.
Important

This is the most common reason a vertical clip looks broken in the feed but fine in the editor. Check the middle band, not the full frame.

4. Two people in one vertical frame

Interview and podcast footage is the hardest case, because both speakers rarely fit in a 9:16 crop. The usual solutions are to follow whoever is talking, so the frame cuts between them as the conversation moves, or to stack the two speakers vertically so both are visible throughout. Following the speaker reads more like television; stacking reads more like a podcast clip, and both are better than a wide crop where neither person is legible.

Active speaker framing
"This is a two-person interview — keep the frame on whoever is speaking."

5. Resolution and quality

Cropping into a frame means using fewer of the original pixels, so the quality of a vertical conversion depends on what you started with. A 4K widescreen source has plenty of headroom for a vertical crop; a 1080p source has enough for a normal crop but not much left over for an aggressive punch-in on top of it. If you know a recording is destined for vertical, shooting at a higher resolution is the cheapest possible fix.

  • 4K source: comfortable for vertical crops with room for zooms.
  • 1080p source: fine for a straight vertical crop, tight if you also punch in.
  • Shoot wider than you need when you know a clip is going vertical later.
Common Mistakes to Avoid
✖ Adding black bars instead of reframing

Why it fails: Letterboxed video wastes most of the screen and reads instantly as a cross-post rather than native content.

✔ Better approach: Crop to fill the vertical frame and let the crop follow the subject.

✖ Putting captions at the bottom of the frame

Why it fails: The platform's own interface sits there, so your captions get covered or compete with it.

✔ Better approach: Place captions in the central band, comfortably above the bottom strip.

✖ Cropping a two-person shot down the middle

Why it fails: You end up with half of each person and a legible view of neither.

✔ Better approach: Follow the active speaker, or stack the two speakers vertically.

Frequently Asked Questions

Common questions around this editing workflow.

No. The framing is stored as part of the timeline, so you can change it, adjust it or revert it at any point before export.
Using Cubix?

Relevant Product Documentation

Want to edit this way without doing everything manually?

Try Cubix conversational AI video editor free for 7 days.