RipGuide
How do I turn a horizontal video into a vertical one (9:16)?
Updated
In short: there are three ways. Crop the picture and follow the subject, stack two crops in a split screen (for example a facecam over gameplay), or fit the whole frame in the middle and fill the bars. Which one works depends on what the moment is about. A 1920 x 1080 video cropped to fill a 1080 x 1920 screen keeps only a 608 x 1080 strip of the source, about 32% of its width, which is then enlarged about 1.8 times. YouTube counts a video as a Short when it is square or vertical, up to three minutes (YouTube help, read 18 October 2026).
The three ways
| Method | What it does | Works when | Falls down when |
|---|---|---|---|
| Crop and follow | Cuts a tall strip out of the wide frame and moves it to keep the subject in view. | One person talking to camera, a podcast host, a reaction, a single object of interest. | Two people at opposite edges, important text or a game HUD at the sides, very fast movement. |
| Split screen | Crops two parts of the source and stacks them: facecam on top, gameplay below. | The reaction and the action both matter, as in most game streams. | The cam is tiny or low quality, because it gets enlarged a lot. |
| Fit with bars | Shrinks the whole 16:9 frame to the screen width and fills the top and bottom with a blurred copy, a colour or a graphic. | Nothing may be cut: a scoreboard, a map, a wide shot, a screen recording. | You want the picture to feel big. The video takes about a third of the screen height. |
For a longer look at the facecam layouts, see how to make a vertical stream clip and keep your facecam.
The resolution maths
9:16 means the width is 9/16 of the height. For a frame that is 1080 pixels tall, the width is 607.5, which is rounded to 608 because video sizes should be even numbers. A standard vertical frame is 1080 x 1920, and its pixel count (2,073,600) happens to equal that of a 1920 x 1080 frame. A centre crop that fills the vertical frame therefore keeps one third of the source pixels and stretches them over the whole screen.
| Source | Crop kept at full height | Share of source width | Scaling to 1080 x 1920 |
|---|---|---|---|
| 1920 x 1080 | 608 x 1080 | 31.7% | Enlarged about 1.78 times. Visibly softer. |
| 2560 x 1440 | 810 x 1440 | 31.6% | Enlarged about 1.33 times. Slightly soft. |
| 3840 x 2160 | 1215 x 2160 | 31.6% | Reduced about 0.89 times. Sharp. |
| 1920 x 1080, fit with bars | whole frame, shrunk to 1080 x 608 | 100% | Reduced 0.56 times. Sharp, but only 32% of the screen height. |
What it means in practice: a 1080p stream cropped to fill the screen is enlarged by almost double, and stream compression is already in the picture. It is usually fine on a phone for a face or a big scene, and it shows on fine detail such as small text. A source recorded at a higher resolution survives the crop. A split screen is gentler on the gameplay, since a pane that is 1080 wide and 960 tall takes a 1215 x 1080 piece of a 1080p frame and shrinks it slightly. A facecam is the weak point: a 480 x 270 cam shown across the 1080-pixel width is enlarged 2.25 times.
Exporting at a higher resolution afterwards does not add detail that was not in the source. The export size only decides how the enlargement is done.
What auto reframe gets wrong
Automatic reframing picks a point of interest in each frame, usually a face, and moves the crop to it. It is good at one speaker. It struggles in predictable places:
- Several people. It has to pick one, and it may switch between them or settle on the wrong one. A split or fit layout is often better.
- On-screen text. Titles, lower thirds, scoreboards and game HUDs sit near the edges, which is the part a crop throws away.
- Fast motion. If the subject moves quickly, the crop either lags behind or shakes. Smoother follow settings help, but a very fast scene is better shown whole.
- Subject at the edge, or leaving the frame. The crop can only go as far as the picture goes.
- Game footage. There is often no face to find, so the tool guesses, or sits in the centre. Check the framing yourself.
Whatever tool you use, watch the whole clip once on a phone-sized preview before you export.
Steps
- Decide what the moment is about. A reaction, the gameplay, or both.
- Pick the layout. Crop for one person, split for cam plus action, fit with bars when nothing can be cut.
- Set the canvas to 1080 x 1920. Use the highest source resolution you have.
- Frame the subject. Keep faces in the upper-middle and leave room at the bottom for captions.
- Check what the crop removed. Text, HUD and the second person are the usual losses.
- Preview full screen on a phone. Look at sharpness and at the app's buttons covering anything.
How to do it in RipClip
RipClip makes every clip vertical 9:16 and frames it on the speaker with a face-centred crop on every plan. On Premium there is a facecam split layout, and on a clip in the editor "Auto reframe and Tracking" is also Premium. Premium features can be used by anyone while editing, but exporting them needs a paid plan. You can export a clip in 8 shapes or at a custom size, and exports go up to 4K at 60 fps: 4K on the Pro and Studio plans, 1080p on Free and Creator. A 4K export of a 1080p source is still a 1080p picture, so the maths above applies.
RipClip clips Kick and Twitch VODs, highlights and clips, Google Drive, Dropbox and OneDrive share links, direct video links and uploads (MP4, MOV, MKV, WebM, AVI, M4V). It does not take YouTube links; upload the file instead.
Once the clip is vertical, keep the captions clear of the app's buttons, and see how long a stream clip should be.
If you already edit in a desktop editor that has its own subject tracking, or you only need one clip reframed once, that is a fine route too. The layouts above work in any editor.