ShotFlickAI video editor
Home / Blog / Glossary

Auto Reframe: Turning Horizontal Video Vertical

Glossary · 5 min read · ShotFlick Team

Auto reframe is an editing feature that converts video from one aspect ratio to another, most often horizontal 16:9 to vertical 9:16, by cropping the frame and automatically moving the crop to follow the main subject. It works well for a single speaker who stays roughly in place, and struggles with two people, fast movement, on-screen text and wide shots. For best results, shoot with reframing in mind and check every reframed clip before posting.

Definition

A 9:16 vertical frame cut from a 16:9 horizontal one keeps only about a third of the original width. A fixed center crop works if the subject never moves from the center, which is rare. Auto reframe solves that by analyzing the video, finding what matters in each shot, usually a face, and moving the crop window to keep it in view.

The feature has different names in different tools: auto reframe, smart crop, auto resize, speaker tracking. The idea is the same. Several professional and consumer editors include it, and clip-making tools typically apply it automatically when turning long videos into Shorts.

How tracking works

A vertical crop window follows the speaker as they move across a horizontal frame Frame 1: speaker on the left Frame 2: crop has followed
The crop window moves smoothly with the subject instead of staying fixed in the center.

Under the hood, most reframing follows the same steps:

  1. Detect shots. The software finds cuts so it can decide framing per shot.
  2. Find the subject. Face and body detection locate people; some tools also look for motion or visually important areas.
  3. Plan a path. The crop position is calculated for every frame, then smoothed so the virtual camera glides rather than jitters.
  4. Render. The cropped area is scaled to the output size.

Most tools let you adjust the result: change how closely the crop follows (smooth vs responsive), pick which person to follow, or set keyframes manually where tracking gets it wrong.

Where it fails

Auto reframe is reliable for the simple case and fragile beyond it. Common failures:

  • Two or more people. The crop chooses one, ping-pongs between them, or splits the difference and shows half of each.
  • Fast movement. If the subject moves quickly, the crop either lags behind or swings sharply.
  • On-screen text and slides. Titles, lower thirds and slides designed for 16:9 get cut off at the sides.
  • Demonstrations. If the important thing is your hands or a product on the table, tracking may stay on your face instead.
  • Wide shots. Small subjects in a wide frame lose detail when cropped and scaled.
  • Edge-of-frame subjects. If someone stands at the far edge, the crop can't move far enough without hitting the frame boundary.
  • Low resolution. 1080p footage cropped to vertical becomes noticeably softer.

Always watch the whole reframed clip. A single moment where the crop drifts off your face is enough to look careless.

Shooting for reframe

If you know footage will be reframed later, a few habits at recording time make the result much better:

  • Center the speaker, or at least keep them well away from the frame edges.
  • Leave headroom. Frame slightly wider than usual so a vertical crop still has room above the head and down to the chest.
  • Record in 4K when possible. A 9:16 crop from 4K keeps enough detail for a sharp 1080x1920 export.
  • Keep important things together. If you hold up a product, hold it near your face rather than out at arm's length.
  • Avoid side-placed text in the original, or keep a clean version without text so you can add vertical-friendly text later.
  • Move slowly. Gentle movements track well; sudden steps sideways don't.
  • Think about the "both" frame. Some creators frame their shot so a vertical safe area in the middle works on its own, and treat the sides as bonus space for horizontal.

For a broader comparison of formats, see vertical vs horizontal video and video aspect ratios explained.

Checking a reframed clip

Before posting, watch the reframed version once at full speed on a phone and look for:

  • Moments where the face drifts toward the edge or out of frame.
  • Sudden swings of the crop, especially at cuts or when someone gestures.
  • Text, logos or slides from the original that are now half cut off.
  • Captions or graphics sitting under the platform's buttons.
  • Softness from cropping low-resolution footage.

Most problems take a minute to fix with a manual keyframe or by switching to a smoother tracking setting. If many moments need fixing, the clip probably needs a re-edit rather than a reframe.

Reframe vs re-edit

Reframing changes the shape of an edit. It doesn't rethink the edit. A horizontal video's pacing, text placement and visuals were designed for a wide screen, and vertical viewers on phones often expect something different: bigger captions placed higher, faster starts, visuals designed for a tall frame.

  • Reframe when the content is a single speaker and the edit is simple. It's fast and usually good enough.
  • Re-edit when the original relies on slides, wide B-roll, two-person shots or text placed for horizontal. Rebuild the layout for vertical instead of cropping through it.
  • Edit natively when you can. If you're starting from a raw recording, an editor that outputs the orientation you need avoids the problem entirely. ShotFlick, for instance, edits a raw talking-head clip of up to three minutes into either a vertical (9:16) or horizontal (16:9) video, so captions and visuals are added for the format you chose instead of being cropped out of a finished edit. Record in the shape you plan to publish for the cleanest result.

If you're turning long horizontal content into clips, our guide to repurposing long videos into Shorts and Reels walks through the whole process, and editing podcast clips covers two-speaker layouts in detail.

Key takeaways

  • Auto reframe crops 16:9 to 9:16 and moves the crop to follow the subject.
  • It works best with one centered speaker and struggles with two people, fast motion, slides and wide shots.
  • Cropping costs resolution; record in 4K if you reframe often.
  • Shoot with reframing in mind: centered subject, extra headroom, objects near your face.
  • When the layout depends on width, re-edit for vertical or edit natively in the format you need.

Frequently asked questions

Does reframing lower quality?

Yes, in terms of resolution. A vertical 9:16 crop from a 1920x1080 frame keeps only about 608 pixels of width, which is then scaled up to 1080 for export, so the image gets softer. On a phone it often still looks acceptable, but footage recorded in 4K keeps far more detail after the crop.

Can I reframe two speakers?

Auto reframe usually follows one subject at a time, so with two people it either picks one, swings back and forth, or tries to keep both in a frame that is too narrow. For conversations, cutting between speakers or stacking them in a split screen usually works better. Some tools offer multi-speaker layouts; check the results closely.

Should I shoot in 4K for reframing?

If you know you'll crop horizontal footage into vertical regularly, yes. A 9:16 crop from 4K footage is roughly 1,215 pixels wide, enough for a sharp 1080x1920 export. The cost is bigger files and heavier editing, so if you mainly publish vertical video, recording vertically from the start may be simpler.

Skip the editing timeline

Upload a raw talking-head clip (up to 3 minutes for now) and ShotFlick trims the pauses, adds captions, and builds a themed edit with visuals and motion graphics, in vertical or horizontal. You pay with tokens, only for what you edit.

Try ShotFlickSee token pricing