9:16 Vertical Video: Framing Tips for Shorts and Reels

Vertical video won. Shorts, Reels, and TikTok all live in the tall 9:16 frame, and the algorithm rewards videos that fill it properly. A horizontal clip with black bars on the sides looks amateur and gets swiped past. Whether you are filming with a phone or generating with AI, framing for 9:16 is its own skill — here is how to get it right.

Why Vertical Matters

People hold their phones upright. A 9:16 video fills the entire screen, which means bigger subjects, stronger immersion, and no wasted pixels. Platforms also measure watch time and completion rate, and a full-screen video simply holds attention better than a letterboxed one. If your destination is Shorts or Reels, compose for vertical from the first frame — never shoot wide and crop later, because cropping throws away resolution and ruins your composition.

Rule of Thirds, Turned Sideways

The rule of thirds still applies, but the grid is tall now. Imagine two horizontal lines dividing the frame into top, middle, and bottom bands. The middle band is prime real estate — that is where eyes land. Place key action along the middle third, use the top third for atmosphere and the bottom third for grounding. Avoid centering everything in the exact middle, which feels static; offset your subject slightly left or right within the middle band for a more dynamic feel.

Keep the Subject Centered (Mostly)

Vertical frames are narrow, so subjects drift out of frame fast. As a default, keep your main subject on the vertical centerline, especially if there is movement. A person walking, a pet running, a product rotating — anchor them centrally and let the background move around them. The exception is when you want tension or style: pushing a subject to one edge with empty space on the other side creates a striking editorial look. But earn it deliberately; accidental edge-cropping just looks like a mistake.

Headroom Mistakes to Avoid

Headroom — the space between the top of a subject’s head and the top of the frame — is where vertical videos most often go wrong. Too much and your subject shrinks into the bottom half; too little and heads get clipped. Frame with a little more headroom than feels natural, because platform UI overlays — progress bars, captions, buttons — eat into the edges. Leave a safety margin at the top and bottom and your subject will survive every interface.

Text-Safe Zones

Every platform covers parts of your video with its own interface: the caption and music label at the bottom, like and comment buttons on the right edge, the progress bar at the very bottom. Keep titles, subtitles, and any important visual detail out of the bottom 15 percent and the right 10 percent of the frame. When writing AI video prompts, you cannot control platform UI — but you can keep the important action in the safe middle band so nothing vital gets covered. If you add your own captions in editing, preview them inside the actual app before posting.

Camera Moves That Work Vertically

Not every camera move survives the tall frame. Lateral moves — pans left and right — feel cramped in 9:16 because there is so little horizontal room. Vertical-friendly moves play to the frame’s strength: height.

Push-in: the camera moves slowly toward the subject. Intimate, dramatic, and perfect for reveals.
Tilt-up: the camera tilts from low to high, great for showing scale — a building, a waterfall, a character reveal.
Tilt-down: the reverse, ideal for grounding a scene or revealing something at the subject’s feet.
Rise/descend: drone-style vertical movement, made for the tall frame.
Slow orbit: circling a centered subject keeps them framed while the world rotates — very effective in vertical.

Avoid fast whip-pans and wide lateral tracking shots; they turn into blurry mush in a narrow frame.

Prompt Phrases That Force Vertical Framing

When generating with AI, the model defaults to whatever it was trained on most — often horizontal. You have to demand vertical explicitly. These phrases, placed in your prompt, reliably force the tall frame:

“vertical 9:16” — the direct instruction; always include it
“smartphone footage” or “shot on a phone” — triggers the vertical association
“portrait orientation” — reinforces the framing
“full-screen vertical” — useful for Shorts and Reels specifically
“subject centered in frame” — keeps the model from drifting sideways

Combine two of them for reliability — for example, end your prompt with “vertical 9:16, smartphone footage, subject centered.” Then check the output: if the model still gives you a wide frame, strengthen the phrasing before changing anything else. Framing is decided in the first seconds of generation, so getting the orientation right in the prompt saves every regeneration after it.