I resisted vertical video for about as long as it was possible to resist it professionally, on the general grounds that it was wrong and I did not like it. Then a client who paid well wanted everything delivered vertical, and I discovered that a lot of what I thought was compositional skill was actually just familiarity with one shape.
The rules do not transfer
Rule of thirds, headroom, lead room, the whole standard vocabulary. It all assumes a frame wider than it is tall, and most of it either inverts or stops applying.
Lead room is the clearest example. In a horizontal frame you leave space in the direction a subject is looking or moving, because there is horizontal space to leave. In a vertical frame there is barely any horizontal space at all. A subject looking off to one side in a vertical frame reads as cramped no matter where you put them, because there is nowhere for the look to go.
What replaces it is vertical space. Movement up and down, and the relationship between what is above the subject and what is below. I found myself thinking about vertical layering in a way I never had before — foreground at the bottom, subject in the middle, something meaningful in the upper third.
Wide shots basically stop working
This was the hardest adjustment. A wide establishing shot in vertical is a narrow strip of a room with a lot of ceiling and floor. It does not establish anything.
The way people solve it is to move closer and use the vertical extent for depth rather than breadth. Instead of showing the width of a space, you show a path through it — something near the camera at the bottom, something distant at the top.
Staircases, corridors, streets seen along their length. Anything with recession into the distance works in vertical because the recession maps onto the long axis of the frame.
Anything wide and flat does not. A landscape that is spectacular horizontally is a strip of ground, a strip of scenery and a lot of sky. There is no vertical framing that saves it.
Two people in frame is a genuine problem
A conversation between two people, shot horizontally, is straightforward. Vertically it is not, and this is where I spent most of my early frustration.
Side by side does not fit. One behind the other works but flattens the relationship. Stacking them, one seated and one standing, works but only for specific situations.
What most people do is cut between singles, which is fine and which changes the rhythm of the edit considerably. You cut more. Whether that is a loss depends on the material, and for a lot of conversational content it genuinely is a loss — you lose the ability to hold both faces and let a reaction happen in the same shot as the line.
Text and interface elements eat the frame
Something I did not anticipate at all. Platforms overlay their own interface on vertical video, and captions, buttons and account information occupy specific regions.
The bottom quarter and the right edge are essentially unusable for anything you need people to see. Which means the actual composable area is smaller than the frame, and it is not centred.
I now shoot with those regions blocked out on the monitor, because otherwise you compose beautifully and then discover the important part is underneath a caption.
This also affects where you put a subject's face. Centred vertically looks natural on an empty frame and sits awkwardly once the interface appears, because the visible area's centre is higher than the frame's centre.
Shooting horizontal and cropping is a trap
The obvious workaround is to shoot horizontal, deliver horizontal, and crop a vertical version. I tried this on a project and would not do it again for anything that matters.
The crops are always compromised. You are choosing between subjects rather than composing with them. Any camera movement designed for the wide frame reads as arbitrary in the crop. And the resolution loss is real if you are not shooting with substantial headroom.
There is a version of this that works, which is shooting at high resolution specifically planning to crop, with the vertical safe area marked out during the shoot and the blocking arranged for it. That is not cropping a horizontal video, that is shooting vertical on a horizontal sensor, and it is a different discipline.
What I actually gained
Some things are better in vertical and I did not expect to say that.
Anything about a single person is stronger. The frame is the shape of a standing human, and a person filling a vertical frame has a presence that the same person in a wide frame does not.
Anything with a strong vertical element — architecture, trees, a fall of water, a staircase — is better served.
And the intimacy is real. A vertical frame held close to a face at reading distance is a different kind of address than a horizontal frame across a room. It suits direct speech to camera in a way that horizontal never quite did.
Where I have landed
I stopped thinking of it as a constraint on real filmmaking and started treating it as its own format with its own grammar, which sounds like the sort of thing you say to make peace with something you did not want.
But the practical result is better work. When I shoot vertical now I plan vertical from the location scout onwards, I look for depth rather than breadth, I block for singles, and I keep the interface zones clear.
The work I did in the first six months, cropping and hoping, is noticeably worse than the work since. The problem was never the format.