Technique · AI content creation

Vertical video with AI.

Vertical video works when it is composed for a tall frame: subjects close and central, action that moves up and down rather than across, and text kept in the middle of the screen. A straight crop from a 16:9 master to 9:16 keeps only about a third of the original width, so good vertical versions are reframed, extended or regenerated rather than simply cut down.

Composing for a tall frame

A tall frame suits one person, one product, standing figures and movement on the vertical axis: a pour, a fall, a rise, a reveal from top to bottom. It struggles with wide landscapes, groups and two people talking side by side. Get closer than you would for a landscape frame, put the eyes in the upper third, and build depth front to back instead of width left to right. When a scene genuinely needs width, split it into two shots rather than squeezing it into one.

Three ways to reframe a horizontal master

A tracking crop follows the subject through the 16:9 frame and is quick, but it loses most of the scene and can feel cramped. Outpainting keeps the original shot and generates new image above and below it to fill the tall frame; the added areas must continue the perspective, light and horizon convincingly, so check edges and repeated patterns. Regenerating the shot vertically from the same references gives the best composition, at the price of continuity checks against the master. Most campaigns use a mix: crops for close-ups, extension for wides, regeneration for hero shots.

Text, captions and the interface

Every vertical platform lays its own interface over the video: captions and audio at the bottom, buttons on the right, account details near the top. Keep logos, prices and legal lines in the central area, design to the strictest platform you deliver to, and check each cut in a real phone preview. Burned-in captions should be large enough to read at arm's length, one short line per beat. Text belongs on a separate layer from the footage, so a new market or a new platform means changing type, not regenerating video.

One master, many vertical outputs

Reels, TikTok, Shorts, Stories and Snapchat all take 9:16 video, usually at 1080x1920, but they differ in maximum length, interface layout and what openings work. Plan one vertical master per idea, then cut platform versions from it: a faster open here, a different end card there, captions rewritten per market. The Synthetic White video editor cuts, captions and exports in every ratio, so the vertical, square and horizontal versions of a campaign are made in one place.

Updated 26 October 2026

Questions

Vertical video with AI, answered.

Can AI turn horizontal video into vertical video?

Yes, in three ways: a crop that tracks the subject, outpainting that extends the frame above and below, or regenerating the shot vertically from the same references. Crops are fastest, regeneration gives the best composition, and outpainting keeps the original shot intact.

What resolution should vertical video be?

1080x1920 pixels is the common delivery size for vertical feeds, at a 9:16 ratio. Some platforms accept higher resolutions, but everything is recompressed on upload, so clean framing and readable text matter more than extra pixels.

Should vertical ads be edited differently from landscape ones?

Yes. Vertical viewers scroll quickly with the sound often off, so open on the product or the payoff, use closer shots and larger type, and keep each beat short. A landscape edit squeezed into a tall frame usually feels slow and distant.

Get started

See your AI studio generating this week.

Thirty minutes. A live studio in your colours, your brand hub loaded, a real campaign brief.

See plans