How to convert horizontal video to vertical for Reels, Shorts, and TikTok

You shot in 16:9 landscape. Reels, TikTok, and YouTube Shorts want 9:16 vertical. What's the least-worst way to bridge that gap?

There are four common approaches, and the right one depends on what's IN your footage. This guide walks through each with the tradeoffs, so you can pick the right one for a given clip.

Why this is harder than it looks

A 1920×1080 landscape frame and a 1080×1920 vertical frame have the same total pixel count, but their shapes are so different that ~40% of the original width is either cropped or letterboxed in the vertical version. What you keep and what you throw away is an editorial choice, not a technical one.

If your subject is centered (single speaker, product shot, centered landscape) any approach works. If your subject moves across the frame, or the important content is at the edges (burned-in text, someone else in frame), the approach matters a lot.

Approach 1 — Fixed center crop

The simplest and most common. You take a 9:16 slice from the center of the frame and throw away the rest.

• Pros: Simple, works with any tool, subject fills the frame at full resolution
• Cons: Anything on the far left or right of the frame gets cropped. Burned-in text (lower thirds, location tags, subtitle strips) gets cut in half.

Use when: your subject stays center-frame the whole clip. Interview / talking-head footage. Product close-ups.

Avoid when: the source has significant graphics/text near the edges, or the subject moves.

How to do it

• CapCut: Import → aspect ratio 9:16 → auto places a centered crop. Move the crop rectangle if needed.
• Premiere / Final Cut / DaVinci Resolve: Duplicate the sequence, set output to 1080×1920, transform each clip's scale to fill.
• ffmpeg (command line): ffmpeg -i in.mp4 -vf "crop=ih*9/16:ih,scale=1080:1920" -c:a copy out.mp4

Approach 2 — Blurred background fill

Instead of cropping the source, you scale it to fit vertically (leaving black bars top and bottom), then fill those bars with a blurred + darkened copy of the same frame. Result: the whole original video is visible, just smaller, and the frame looks aesthetically full.

• Pros: Nothing gets cropped. Original text, logos, graphics all survive. Looks intentional (this is what most TikTok-first meme reposts use).
• Cons:Source video is smaller on screen. Feels less "native" to short-form.

Use when:source has important content near the edges (burned-in text, multiple subjects). Or when you're reposting older long-form content and want to preserve everything.

How to do it (ffmpeg)

The magic is a filter graph with split, one stream gets scaled + blurred for background, the other gets scaled to fit for the foreground, then overlay composites them:

• [0:v]split=2[bg][fg];[bg]scale=1080:1920:force_original_aspect_ratio=increase,crop=1080:1920,boxblur=30:2[bgb];[fg]scale=1080:1920:force_original_aspect_ratio=decrease[fgs];[bgb][fgs]overlay=(W-w)/2:(H-h)/2

CapCut also has this as a preset (called "Background Blur" or "Fill"). Most modern editors have it too.

Approach 3 — Static keyframed reframe ("pan and scan")

You pick a few keyframes throughout the clip and manually move the 9:16 crop window to follow the action. Between keyframes the editor interpolates smoothly.

• Pros: Subject stays in frame even when they move. Looks polished.
• Cons:Manual work. Bad interpolation looks jerky. Doesn't scale to hundreds of clips.

Use when: you have 1–3 hero clips that are worth 20 minutes of manual reframing each.

Approach 4 — AI face-tracking reframe

The 9:16 crop window automatically follows detected faces in the video, smoothly panning between them. When only one face is visible, it centers on that face; when multiple faces appear, it jumps or zooms to include them.

• Pros: Automatic. Handles subject movement. Works at scale.
• Cons:Requires more compute. Occasionally picks wrong subject in complex scenes. Doesn't know that the "subject" might actually be a mountain in the background.

Use when: talking-head vlogs, interviews, speaker-on-stage content — anything face-driven.

Avoid when: landscape / cinematic footage where the subject is the environment, not a person.

Tools with face-tracking reframe include most of the AI clip generators. It's becoming table-stakes in the category.

Quick decision guide

Common mistakes when reframing

• Not exporting at 1080×1920. Some editors default to 720×1280 which looks noticeably softer on modern phones. Always export at 1080×1920 minimum.
• Baking in captions before reframing. Reframe first, add captions after — otherwise captions end up in the wrong position for the vertical crop.
• Forgetting to include padding at the top/bottom. Reels and Shorts UIs overlay text (like/comment/share icons, username) on the bottom third of the video. If your caption or key subject sits in that zone, it gets covered.
• Using different reframe styles across a series. Pick one look for your channel and stick with it. Consistency builds recognizability.

Where Nomads Shorts fits

Nomads Shorts currently uses center crop— the simplest approach — because it's reliable and works well for the talking-head travel content we were designing for. Face-tracking is on the roadmap.

If your content is heavier on landscape/cinematic B-roll or has important edge content, you might get better results with a blurred-fill workflow in CapCut. If it's primarily talking-head footage, center crop (what we do) is usually the right choice.

Try it with 15 free credits to see how our reframing works for your specific footage.

Related reading

• How to turn long YouTube videos into shorts (2026 guide)
• The travel vlogger's guide to YouTube Shorts