From landscape to vertical: reframe your AI clip

You turn a landscape 16:9 AI video into vertical by reframing it, not by simply cropping it: you shift the crop window with your subject shot by shot, or you generate the clip natively in 9:16 from the start. A fixed center crop looks easy, but it pushes anything that isn't dead center out of frame.

That vertical matters shows in how people watch. According to ScientiaMobile's Mobile Overview Report, 82.5% of smartphone users watch video in portrait orientation, and the few who do rotate wait roughly half a minute before they bother. On Reels, TikTok and Shorts, 9:16 isn't a bonus, it's the norm.

Why cropping throws away so much of the frame

Crop a 16:9 clip to 9:16 while keeping the full height, and you're left with only about a third of the original width, roughly 32%. Nearly two-thirds of your horizontal frame disappears outside the crop.

For an interview or a presenter sitting neatly in the middle, that's fine. But the moment your subject moves left or right, or two people stand side by side, a fixed center crop cuts off heads, hands or half the subject. What was a nice composition in your landscape clip becomes empty sky with a sliced-off shoulder in 9:16.

Reframing: let the crop follow your subject

Reframing means you move the 9:16 crop window with the main action shot by shot, instead of cropping the frame once and leaving it static in the middle. That keeps your subject in shot even when it isn't centered.

• Set the crop so the face or main action sits in the safe center column, with a little room above the head.
• If your subject moves across frame, shift the crop along with a gentle pan so it doesn't stick to the edge.
• AI auto-reframe follows your subject automatically using subject tracking, but it stumbles on fast or unpredictable movement. Always check those shots yourself and correct the crop where needed.

Better still: generate vertical from the start

If you haven't made the clip yet, choose 9:16 while generating instead of cropping afterwards. You keep the full resolution instead of throwing two-thirds of it away, and you compose for the tall frame from the outset.

In the video generator you pick the format up front. If you work with image-to-video, start from a vertical still made in the photo generator, so the motion already happens inside the upright frame. Because you pay per render, it costs nothing extra to work in the right format from the start rather than repairing it later.

How to reframe without losing your subject

1. Decide up front which subject always has to stay in frame.
2. Keep that subject in the middle vertical strip, with room above the head.
3. Shift the crop along with the movement instead of cropping statically.
4. Check the edges for cut-off heads, hands and text.
5. Still unsure? Regenerate the shot natively in 9:16 instead of cropping again.

Frequently asked questions

Can I make a landscape AI video vertical without quality loss?

Cropping always costs frame: you keep about a third of the width and enlarge that crop, which lowers the effective resolution. Generating natively in 9:16 keeps the full quality.

What's the difference between cropping and reframing?

Cropping fixes the frame once, usually in the middle. Reframing shifts the frame with your subject shot by shot, so the action stays in view even when it isn't centered.

Which aspect ratio should I choose for Reels, TikTok and Shorts?

9:16 (vertical). It fills the screen on mobile and matches how most people watch video.

Working vertical is easiest at the source: choose your format before you render, not after. Create an account and generate your next clip straight in 9:16.