How to Convert Landscape Video to Vertical Without Ruining the Shot

A practical guide to converting 16:9 footage to 9:16 for Shorts, Reels and TikTok — why centre-cropping fails, when to use split-screen, and how to reframe a whole recording automatically.

Clipo AI August 12, 2026 9 min read

A widescreen video being reframed into a vertical 9:16 crop

Almost every recording worth repurposing was shot wide. Podcasts, webinars, interviews, conference talks, customer calls — all 16:9. And almost every feed worth publishing to is vertical. That gap is where most repurposing projects quietly die: the footage exists, the vertical version looks bad, and nobody wants to publish it.

The good news is that the failure is almost always mechanical rather than aesthetic. Understanding why the obvious approach breaks tells you exactly what to do instead.

Why centre-cropping fails

The default instinct is to crop the middle of the frame and export. It works for a single speaker sitting dead centre, and fails for nearly everything else.

In a two-person interview, the centre of a wide frame is usually the gap between the guests. Crop it and you get a vertical video of a wall, with two half-faces pinned to the edges. Crop to one side and it looks correct until the other person starts talking — at which point you're showing a silent listener while the audio comes from off-screen.

Two-shots — the centre is the space between people, not a person
Stage and webinar footage — the speaker is often off to one side of a wide frame
Screen-share recordings — cropping cuts off half the slide
Anything with movement — a fixed crop can't follow someone who moves

The three approaches that actually work

There are only three real strategies, and choosing the right one per clip matters more than any export setting.

1. Speaker tracking

The crop follows whoever is speaking, moving across the frame as the conversation moves. This is the right default for interviews and panels: it keeps the person talking centred without you keyframing anything.

It works best when there's a little room around each speaker in the original. Guests sitting hard against the left and right edges leave nothing to crop around.

2. Split-screen

Two speakers stacked vertically, both visible the whole time. This is the better choice whenever guests interrupt each other, because constant cuts between speakers are more distracting than a static two-up.

Rule of thumb: if you'd describe the conversation as 'lively', use split-screen.

3. Manual framing

Sometimes you want the crop somewhere the audio doesn't suggest — on a reaction rather than the speaker, or on a product rather than a person. Setting the centre point and zoom by hand for those segments is worth the time, but it should be the exception, not the workflow.

Don't forget the safe zones

Every vertical platform covers part of the frame with its own interface. A clip that looks perfectly framed in an editor can have the speaker's chin behind a caption and the CTA behind a username.

Keep faces and key text out of the bottom quarter — that's where captions, handles and descriptions sit
Keep the top ~10% clear of anything critical
The right edge carries the action buttons on most platforms
Check a real post, not just the editor preview — the overlays differ per platform

Captions are part of the conversion

A vertical clip without captions is a half-finished clip. Most short-form is watched muted, and reframing a video for a feed while leaving it unwatchable without sound defeats the exercise.

Treat captions as part of the same pass as the crop rather than a separate job afterwards: it's fewer round-trips, and it means the caption placement is chosen against the actual vertical framing rather than guessed.

Doing it at volume

One clip is a five-minute job by hand. Twenty clips a week isn't — and that's the cadence short-form actually rewards. This is where automatic reframing stops being a convenience and becomes the only viable approach.

The workflow that scales looks like this: upload the wide recording once, let clip detection propose the moments worth cutting, let speaker tracking handle the crop, apply one caption style from a brand kit, and review only the switch points where the framing changes. Reviewing the transitions rather than watching every clip end-to-end is the single biggest time saving available.

Export settings that matter

Most of the settings people worry about don't matter much. Two do.

Resolution: export at 1080×1920. Upscaling a 720p crop from a 1080p source looks soft, so start from the highest-resolution original you have
Aspect ratio: 9:16 for Shorts, Reels and TikTok; 1:1 is still useful for in-feed posts; keep the 16:9 original for YouTube

Screen shares and slides need a different answer

Speaker tracking solves faces. It does nothing for a slide, a spreadsheet or a screen recording, because there is no subject to follow and cropping to 9:16 throws away most of the information. A slide cropped to vertical is usually illegible, which makes it worse than not posting it.

There are three workable options, in rough order of quality.

Stack the slide above the speaker

Put the full 16:9 slide in the upper portion of the vertical frame and the speaker below it. Nothing is cropped, the slide stays readable, and the arrangement reads naturally in a vertical feed. This is the default worth reaching for.

Cut between them

Show the slide full-width when it is being introduced, then cut to the speaker for the explanation. More editing effort, but it produces the tightest clip when a single slide carries the point.

Rebuild the slide as an overlay

If a slide is one number or one sentence, do not crop it — retype it as text over the speaker. A 40-point number on screen beats a shrunken screenshot of a chart every time.

Multi-camera and B-roll footage

Multi-cam recordings are the easiest source to reframe well and the easiest to reframe badly. If your recording already cuts between a wide two-shot and individual close-ups, the close-ups reframe almost perfectly — they are already framed on one person — while the wide shots are the ones that need tracking or a split.

The practical move is to prefer close-up segments when choosing where to cut a clip. A clip built from close-ups needs almost no reframing work, which is a good reason to let the source footage influence which moments you pick.

B-roll is the opposite case. Cutaways of a product, a screen or a room have no speaker to track, so a static crop chosen by hand is usually correct. Keep those segments short — B-roll cropped to vertical loses context fast, and a long vertical cutaway reads as filler.

The artefacts to look for

Automatic reframing fails in recognisable ways. Knowing the four common ones turns review from watching everything into checking for specific things.

Crop jitter — the frame drifting or twitching during a single speaker turn, usually caused by background movement being read as a subject. Fix by locking the framing manually for that segment
Cut-off gestures — the speaker hands leaving frame at exactly the moment they are making a point with them. Zoom out slightly rather than recentring
Late switches — the crop arriving on the new speaker a beat after they start talking. Almost always fine; only worth fixing on the clip opening line, where it costs you the hook
Wrong-subject lock — tracking settling on the listener rather than the speaker in a noisy room. This is the one artefact that genuinely ruins a clip, so it is worth a specific look on any clip with more than two people

When not to reframe at all

Not every clip should be vertical, and forcing it is a common mistake. Some footage carries its meaning in the width of the frame.

Anything where the composition is the point — a wide landscape, a group shot, a stage with visible scale
Dense screen recordings where even a stacked layout leaves the content unreadable
Side-by-side comparisons — the whole idea is two things in one frame
Footage you will post to YouTube proper, where 16:9 remains the native format and a vertical crop is a downgrade

A five-point check before publishing

Run this on each clip. It takes about twenty seconds and catches nearly everything that would make you delete a post later.

Is the speaker face clear of the bottom quarter?
Does the framing hold through the first two seconds — the part that decides whether anyone watches?
Are the captions readable over the busiest frame in the clip, not just the first one?
Does the crop stay on the right person at every speaker change?
Watch it muted. If it still makes sense, it is ready

Working through an existing archive

Most people arrive at vertical video with a back catalogue already sitting there — years of webinars, episodes and talks, all 16:9. Reframing that archive is a genuinely different job from handling this week upload, and treating it the same way is why it usually never gets done.

The instinct is to start at the beginning and work forward. Do not. Archive footage has wildly uneven value, and chronological order guarantees you spend your first week on your oldest and least relevant material.

Sort by evergreen value, not by date

Start with the recordings whose subject has not moved: the explainer that still holds, the interview with someone still notable, the talk you still get asked about. A two-year-old clip on a topic that has not changed performs like a new one, because the feed has no idea how old the footage is.

Check the resolution before committing

A 9:16 crop keeps roughly a third of the frame width, so a 720p original yields something closer to 240p wide. Old recordings are the most likely to be low-resolution, and it is worth confirming what you actually have before planning a month of content around it. If a recording is too soft to crop, it may still work as a stacked layout where the footage occupies less of the frame.

Set the brand look once, at the start

Archive work is batch work, and batch work is where inconsistent styling shows up worst. Decide the caption style and framing conventions before the first clip rather than during, or you will end up with a feed that visibly changes appearance halfway through your own back catalogue.

Key takeaways

Centre-cropping fails on two-shots because the centre of the frame is the gap between people.
Use speaker tracking as the default, split-screen for crosstalk, manual framing as the exception.
Keep faces and text out of the bottom quarter — platform UI covers it.
Caption in the same pass as the crop, not afterwards.

Frequently asked questions

Can you convert 16:9 to 9:16 without losing quality?

You'll always lose pixels — 9:16 keeps roughly a third of a 16:9 frame's width — so the aim is to lose the right ones. Start from the highest-resolution original you have and export at 1080×1920; a crop from a 4K or 1080p source holds up well, while a crop from a 720p source will look soft no matter what you do.

What's the best way to reframe a two-person podcast?

Split-screen if your guests talk over each other, speaker tracking if they take clear turns. The mistake to avoid is a fixed centre crop, which on a two-shot frames the empty space between them.

Do I need to reframe manually for the best results?

Rarely. Manual framing is worth it when you want the crop somewhere the audio doesn't suggest — on a reaction, or on a product. For ordinary talking-head and interview footage, speaker tracking plus a check of the switch points gets you there far faster.

How long does it take to reframe a full episode?

By hand, roughly five minutes per clip once you include checking the framing. Automatically, the reframing itself is not the bottleneck — reviewing the clips is, and that's minutes for a whole batch if you review the transition points rather than watching each clip through.

How do I make a screen share or slide deck work in vertical?

Do not crop it. Stack the full 16:9 slide in the top of the vertical frame with the speaker below, or cut between a full-width slide and the speaker. If a slide is really just one number or sentence, retype it as text over the speaker — a large clean number beats a shrunken screenshot.

Why does the crop jitter or drift during my clips?

Usually background movement being read as a subject — someone walking behind the speaker, or a screen changing. Lock the framing manually for that segment rather than trying to tune the tracking; it is a few seconds of work and only affects the clips where it actually happens.

Should every clip be vertical?

No. Keep 16:9 when the composition carries the meaning — wide landscapes, group shots, side-by-side comparisons, or dense screen recordings that stay unreadable even stacked. And for YouTube proper, 16:9 is still native; a vertical crop there is a downgrade rather than an optimisation.

Keep reading

Turn one recording into a week of content

Clipo finds the clips, reframes them vertical, captions them, and drafts the posts and carousels that go with them — from a single upload.

Try Clipo Free