How to convert landscape video to vertical
A full-height 9:16 crop of a 1920x1080 frame is 607 pixels wide. Which third of the picture you keep, and whether it is allowed to move, is the entire job.
8 minute read
Start from what the crop costs you
Take a standard 1920x1080 landscape frame and cut a full-height 9:16 window out of it. That window is 607 pixels wide. You are keeping under a third of the picture and discarding the rest before you have made a single creative decision, which is why this feels harder than it should. Everything below is about which third you keep and whether it is allowed to move. It is also worth asking whether you should convert at all. B2Linked tested aspect ratios across four LinkedIn Ads accounts and more than 140,000 dollars of spend and found 16:9 beat 9:16 clearly on view rate and cost per completion, with the honest caveat that they were testing polished produced ads rather than talking-head footage. Vertical is the default for short-form feeds, not a universal law.
The static centre crop, and when it is genuinely fine
The simplest method is also the most maligned. Set a 9:16 window over the middle of the frame and leave it there. For one person, framed centre, on a camera that does not move, this is completely fine, and anything you add on top of it is wasted work. A solo talking-head recording, a webcam, a lecture shot square-on: the subject is already where the crop is, the shot does not change, and a moving crop would introduce drift that was never in the original. If your footage is one person who stays put, do the static crop, put the time you saved into the first three seconds, and skip the rest of the reframing literature.
The shots where a centre crop falls apart
It breaks the moment the frame holds more than one point of interest. Two people either side of a table is the classic failure, because the middle of the frame is the gap between them: the crop lands on the space, and clips both faces at the edges. A three or four person panel is worse. So is any shot where the speaker is deliberately off-centre, like an interview subject framed to one side with looking room, or a presenter standing beside a screen. Anything carrying burned-in graphics, lower thirds, or a shared screen loses exactly the part that held the information. In each case the fix is not a better centre point. It is a crop that moves, or a layout that stops pretending one window can hold the shot.
Crop by hand, and move only on cuts
Manual reframing is the same three steps in any editor: drop the landscape clip onto a vertical timeline, scale it until its height fills the frame, then animate horizontal position. What separates good manual work from obviously-cropped work is when you move. Hold the crop still through a shot and jump to the new position on a cut that already exists, so the movement hides inside an edit the viewer expects. Where there is no cut to hide behind, move decisively rather than drifting — a slow creep across the frame reads as a mistake, a quick push reads as a camera move. Two keyframes and a short ease is usually the whole job. Put them on the cut, not spread around it.
Auto reframe in the editor you already own
All three major editors ship a version of this, and they are not equivalent. Premiere Pro's Auto Reframe, introduced in 2019, applies as an effect to a clip or across a whole sequence and uses motion tracking to follow the action, with a setting that controls how hard the crop chases movement. DaVinci Resolve has Smart Reframe, which detects the subject and holds it in the new frame. Final Cut Pro's Smart Conform is the outlier: editors comparing the three have pointed out that it conforms clips to a new ratio without tracking a subject the way Auto Reframe does. The property that matters more than any of that is shared — each writes keyframes you can then edit. Treat the automatic pass as a first draft and hand-fix the shots it misreads.
Sometimes the answer is a layout, not a crop
Some shots should not be cropped at all. Two people on opposite sides of a wide two-shot often work better stacked, one above the other in their own windows, because you keep both faces and both reactions instead of choosing between them every few seconds. A group shot around a table can be dropped in whole at full width, with the space above and below carrying a headline and captions, which reads as a deliberate design rather than a failed crop. Screen shares and slides nearly always want this, since a vertical slice of a slide is unreadable. The test is blunt: if no single 9:16 window contains the thing the viewer needs to see, stop hunting for one.
Decide the safe area before you commit keyframes
Do this first, because it changes where faces belong. Every app draws its own interface over your video — captions, handle, buttons, progress bar — across the top and bottom of the frame and down the right edge. If you compose the crop to centre a face vertically, your burned-in captions land on the chin, or the platform's own caption bar covers your text. The habit that fixes it is to put the subject's eyes in the upper third and reserve the lower third for captions from the start, then check the finished clip against each platform's layout before you publish rather than after. Framing and captioning are one decision, not two sequential ones, and treating them separately is what produces clips that look fine in the editor and wrong in the feed.
Convert once, at the source
The workflow trap is converting per platform. Shorts, Reels, TikTok, Facebook Reels and vertical LinkedIn all take 9:16 at 1080x1920, so you need one vertical master rather than five conversions, and what changes between destinations is the caption and the cover frame, not the crop. The reason people end up redoing the work is volume. One clip is a pleasant twenty minutes of keyframing; twelve clips out of one recording is a lost afternoon, and that is the point where automated reframing earns its place — provided you keep the ability to fix what it misjudges. FrameOS cuts long recordings into vertical captioned clips with framing that follows whoever is speaking, and the fair way to evaluate it is to feed it your own hardest two-person footage. 300 credits for 3 days · no card.
FAQ
How do I convert a 16:9 video to 9:16 without cutting off the speaker?
Scale the landscape clip to fill the vertical frame's height, then animate horizontal position so the crop sits on whoever is speaking, moving on existing cuts rather than drifting. On single-speaker footage a static centre crop achieves the same thing with no keyframes. On two-person footage you either need a moving crop or a stacked layout that keeps both people.
Is a static centre crop good enough?
For one person on a fixed camera, yes, and adding motion would only make it worse. It fails predictably on two-person interviews, panels, deliberately off-centre framing, and anything with a shared screen or lower thirds, because the centre of the frame is not where the subject is. Judge it on your own hardest shot, not your easiest.
How much of the picture do I lose converting landscape to vertical?
From a 1920x1080 source, a full-height 9:16 crop is 607 pixels wide, so you keep roughly 32 percent of the frame. That is why the choice of crop position matters so much: two-thirds of the shot is gone regardless, and the only decision left is which third survives.
Should I add black bars instead of cropping?
Only when the content genuinely needs the full width, like a slide or a wide group shot. Letterboxing a talking head wastes most of the screen in a full-screen feed. If you do keep the full frame, fill the empty space above and below with a headline and captions so it reads as a designed layout rather than an unedited upload.
Sources
Related pages
Keep reading
More Video Basics →Social media video specs for every platform
One 1080x1920 export covers most of short-form. This is where that stops being true — the length caps, file limits, and the platforms where the number changes depending on how you upload.
8 minute readSocial media image sizes for every platform
Instagram's help page now allows 3:4, YouTube's recommended thumbnail is 4K, and TikTok publishes almost nothing. Here are the numbers each platform actually documents, and where they leave you guessing.
7 minute readHow to crop video for Instagram
Your reel plays at 9:16 and previews on your profile inside a 3:4 rectangle. Anything you centred for one gets cut by the other, which is why so many covers lose their own headline.
7 minute readTurn one long video into a week of posts
300 credits for 3 days · no card.
Start free