The challenge
Vertical drama – series shot for the phone screen – is growing fast. Existing footage, however, is horizontal. Reframing every shot to 9:16 by hand takes hours per episode.
A simple centre crop doesn’t work: in dialogue scenes the speaker sits at the edge of the frame, and faces must not disappear behind the app’s interface.
The solution: Convertical
Convertical analyses each episode and operates a virtual camera: it follows whoever is speaking, pans smoothly and cuts hard at scene changes. Two-person dialogues can be laid out as split screens automatically.
In the browser editor, editors correct whatever they don’t like – by dragging directly in the frame. Everything runs locally on a Mac; no footage leaves the building.
How it works
-
Shot and face detection
Cuts, faces and people are detected; the same actor is one person across every shot.
-
Active speaker detection
Lip movement and audio reveal who is speaking – with hold times so the camera doesn’t jump around nervously.
-
Automatic camera work
Dead zone, smooth pans, hard cuts – like a human camera operator. Safe zones keep faces clear of app overlays.
-
An editor for professionals
Keyframes, split screens, episode splitting and keyboard shortcuts. Manual changes are preserved.
-
Premiere export & 4K
Export as 4K vertical video or as an editable Premiere Pro project with every keyframe.
Outcome
- Manual reframing becomes review and correction: the AI delivers the first cut, the human the fine-tuning.
- Long episodes are split into mini-episodes at quiet cut points automatically.
- Fully local: no upload times, no cloud costs, no rights issues with third parties.