In short: one person who moves → Smart reframe. Wide shot or text on screen → Fit with a blurred background. Centred and static → a plain Fill crop is enough.
The maths of going vertical
A 16:9 frame is 1.78 times wider than tall; a 9:16 frame is 0.56. To fill a vertical frame from a landscape one, you keep a slice only 32% of the original width. From 1920 × 1080 footage that slice is about 608 pixels wide, which is then scaled up to 1080 — so vertical crops from 1080p footage are softer than you might expect. Shooting in 4K gives you a slice of about 1215 pixels, enough for a sharp 1080 × 1920 result.
Three options
| Option | How in VDOAPP | Best for | Watch out for |
|---|---|---|---|
| Fill (crop) | Canvas → 9:16 → Fill (crop) | Centred, static subjects | Anything off-centre is cut off |
| Fit + background | Canvas → 9:16 → Fit, Background: Blur or Colour | Wide shots, two people, slides, screen recordings | Small picture; add large captions above/below |
| Smart reframe | Fill, then AI → Smart reframe | One person who moves around | Groups — it follows the largest face |
How smart reframe works
Smart reframe samples the clip about three times a second. At each sample it looks for the largest face; if there's none (the person turned away, say), it uses the centre of the person mask from the segmentation model; failing that, it keeps the last known position. That path is then smoothed so the virtual camera glides rather than jitters, and written into the clip as position keyframes — so you can see it in the Clip tab and nudge any point by hand.
The result behaves like a camera operator who keeps the speaker in frame. It works on any canvas shape that's narrower than your footage — 9:16, 4:5 or 1:1 from 16:9.
- Set the canvas to 9:16 (or 4:5 for Instagram feed) with Fill (crop).
- Select a clip and press Smart reframe in the AI tab's Camera section.
- Repeat for each clip. Undo reframe removes it from a clip.
Reframing by hand with keyframes
When you want a specific framing — cutting from a speaker to the object they're pointing at — use keyframes in the Clip tab instead. Move the playhead, set Position X so the right part is in view, and a keyframe is recorded. The editor animates smoothly between keyframes. Two or three keyframes per clip is usually enough; more makes the movement feel nervous.
Vertical safe zones
Short-video apps cover parts of your frame with their interface: the bottom 20% or so with the caption and music line, the right edge with like and comment buttons, and a strip at the top. Keep faces and text in the central area. That's why VDOAPP's caption Middle position is the right choice for vertical.
Tip: when filming new content you know will be vertical, shoot vertically — or shoot landscape with the subject centred and generous headroom so every crop works.
Frequently asked questions
Does smart reframe work with two people?
It follows the largest face, so it favours whoever is closest to the camera. For a two-person conversation, Fit with a blurred background, or split the clip and reframe each speaker's turn, works better.
Can I reframe to square?
Yes — set the canvas to 1:1. Smart reframe works whenever the canvas is narrower than the footage.
Should I stabilise before reframing?
Yes. Stabilise first so the tracker follows the subject, not the shake.