How to Turn Horizontal Video Into Vertical (Reframe and Crop)

To turn a horizontal video into vertical, you crop a 9:16 window out of the 16:9 frame and move that window so the person or action stays in it. When one window cannot hold what matters, you switch layouts: a split screen for two speakers, or the full frame on a blurred background. Then export at 1080x1920.

This lesson covers why vertical matters, the crop math, four layouts and when to use each, and how to do it in an editor, with a free command-line tool, or with AI speaker tracking.

Why you can’t just upload the horizontal file

A 16:9 video on a phone held upright fills less than a third of the screen. It looks small, and it signals that the video was made for somewhere else.

There is also a rule on YouTube: it categorizes videos with a square or vertical aspect ratio, up to three minutes long, as Shorts. A 16:9 upload stays a regular long-form video, whatever its length. TikTok’s own creative guidance for ads also tells you to orient vertically at 9:16 and keep content inside the UI safe zone, which is a good rule for organic posts too.

The crop math

Going from 16:9 to 9:16 means keeping a narrow vertical slice of the frame. The height stays, the width shrinks.

Two things follow from this:

• About two-thirds of the frame is thrown away. If your subject moves, a fixed crop will lose them. The window has to follow.
• Resolution matters. If you know you will crop, film in 4K. From 1080p the result is usable, but faces and text look softer.

Four layouts and when to use them

Tracked crop

This is the default for talking content. The window sits on the speaker’s face, with the eyes about a third of the way down the frame. In an interview, it jumps to whoever is talking. Jump cuts (a hard switch) usually look cleaner than a slow pan between two people.

Split screen

When both people matter at once, such as a debate or a laugh, stack them. Crop each person to a near-square box and place one above the other, so each takes half the 1080x1920 frame. Put captions on the line between the two boxes.

Blurred background

Scale the full horizontal frame to the width of the vertical frame, center it, and fill the top and bottom with a blurred, zoomed copy of the same video. It is the easiest layout and the only honest one for slides or wide shots, but the picture is small. YouTube Studio’s own Clips and Shorts tool uses a similar look: YouTube says the final Short shows the empty space as a blurred crop.

How to do it in an editor

Most editors work the same way:

1. Create a vertical project at 1080x1920, or change the project or sequence aspect ratio to 9:16.
2. Drop in the horizontal clip and scale it up until it fills the frame (about 178% for a 1080p clip in a 1080p-wide project).
3. Position it so the subject is in the window.
4. Keyframe the position wherever the subject moves or the speaker changes. For an interview, cut the clip at each speaker change and reposition each piece instead.
5. Export at 1080x1920, H.264, at the frame rate you filmed in.

Many editors have an automatic version of steps 3 and 4. In Premiere Pro it is called Auto Reframe, in DaVinci Resolve Studio (the paid version) Smart Reframe, and in Final Cut Pro Smart Conform. CapCut also has an auto reframe option. They all detect the subject and move the window for you; check the result and fix any frames where it drifts.

How to do it free with FFmpeg

If you are comfortable with a terminal, the free tool FFmpeg can reframe without opening an editor. These commands were tested on a 1920x1080 file.

Center crop to full-screen vertical:

ffmpeg -i input.mp4 -vf "crop=ih*9/16:ih,scale=1080:1920" -c:a copy vertical.mp4

Full frame on a blurred background:

ffmpeg -i input.mp4 -filter_complex "[0:v]split[a][b];[a]scale=1080:1920:force_original_aspect_ratio=increase,crop=1080:1920,boxblur=20[bg];[b]scale=1080:-2[fg];[bg][fg]overlay=(W-w)/2:(H-h)/2" -c:a copy blurred.mp4

Two speakers stacked (left half on top, right half below):

ffmpeg -i input.mp4 -filter_complex "[0:v]split[l][r];[l]crop=960:854:0:113,scale=1080:960[top];[r]crop=960:854:960:113,scale=1080:960[bot];[top][bot]vstack" -c:a copy stacked.mp4

A center crop only works if the speaker stays in the middle. For anything that moves, use an editor or a tracking tool.

Speaker tracking with AI

For podcasts and interviews, doing the window by hand means a keyframe or cut at every speaker change, which adds up fast across a batch of clips. AI speaker tracking detects faces, works out who is talking from the audio, and moves the window for you. Short.now does this as part of making each clip, along with the captions.

Whatever tool you use, check these on every clip:

• Does it follow the right person? Tracking can stay on a listener who nods instead of the speaker.
• Is the motion smooth? Small twitches from frame to frame are distracting.
• Does it cut or pan? A fast pan between two people can look like a camera glitch.
• Are hands and props in frame when they matter, such as a product demo?
• What happens on wide shots with three or more people? You may need a split screen or a blurred-background segment there.

Captions and safe zones after reframing

Reframing changes where you can put text. On all three apps, buttons and the caption sit on the right side and the bottom of the screen. Keep captions and important action in the middle area, roughly the center two-thirds of the height, and away from the right edge. The format specs cheat sheet has the safe zones on one page, and the size guides for YouTube Shorts, TikTok and Instagram Reels go into detail.

Common mistakes

• Cutting off the top of the head and the chin by zooming too far. Leave a little room above the head.
• Captions over the face after a speaker change moves the window.
• Using a center crop on a two-shot, which leaves you staring at the gap between two people.
• Exporting at the source resolution, such as 608x1080, instead of 1080x1920.
• Blurred backgrounds on talking heads, where a tracked crop would make the face three times bigger.

What’s next

You now know how to find moments and fit them on a vertical screen. In the next lesson, batch content creation, you will put it all together: film once, edit in one sitting and schedule a month of posts. If you want a refresher on picking moments, go back to how to find clip-worthy moments.