What is a oner? The long take, the tracking shot, and how to direct one with AI

A oner is a scene filmed in one continuous shot, with no cuts: the camera starts rolling, moves through the action, and the scene ends before it stops. Film crews also call it a long take, a single take or a "one-shot". The appeal is that the audience experiences time exactly as the characters do, with nothing edited out, and the difficulty is that everything has to go right in the same pass. This guide explains what separates a oner from a long take and a tracking shot, why directors use them, and how to plan one when the camera is an AI video engine rather than a Steadicam.

What is a oner in film?

A oner is a shot that plays a whole scene, or a whole sequence, without a single cut. The word is crew slang for "done in one": one setup, one continuous take, every beat of the scene staged for the camera as it moves.

What makes a oner different from any other shot is that it carries the structure a cut would normally carry. In a conventional scene the editor chooses when you see the wide, the close-up and the reaction. In a oner, the camera move and the blocking of the actors make those choices, live. The close-up happens because the camera pushes in, and the reaction happens because an actor turns into frame.

What is the difference between a oner, a long take and a tracking shot?

The three overlap, which is why they get used interchangeably:

Term	What it describes	Must the camera move?	Must it be a whole scene?
Long take	A shot that runs much longer than the editing around it	No	No
Oner	A scene or sequence played in one uncut shot	Usually, not always	Yes
Tracking shot	A shot where the camera travels with or alongside the subject	Yes	No
One-shot film	A whole film presented as one take, often with hidden cuts	Usually	The whole film

A static wide shot of two people talking for four minutes is a long take and a oner, but not a tracking shot. A three-second dolly beside a runner is a tracking shot, but neither of the others. The famous examples, a camera following a couple through a kitchen into a nightclub, or a soldier running along a trench, are all three at once.

Why do directors shoot oners?

• Tension. Without a cut, nothing relieves the audience. If the scene is dangerous, the absence of a cut says nobody is going to save it in the edit.
• Geography. A camera that walks you from the street, through a door and into a room teaches the space better than any establishing shot.
• Performance. Actors play the scene through, and small moments survive that coverage would have trimmed.
• Spectacle. A long, complex oner is visibly hard, and audiences notice craft.

The cost is control. A oner cannot be tightened in the edit, so pacing has to be designed into the move before the camera rolls.

Can AI video make a oner?

Short ones, yes, and this is where AI video has changed recently. Most generators render a few seconds per render, which is a shot, not a scene. PrismPoster's Premium engine renders one continuous take of any whole second from 4 to 30 seconds, and you direct it with timed beats, so a 30-second oner is one render rather than a stitch of separate shots. The Standard engine renders 4, 6, 8, 10-second shots, which suits conventional coverage cut together on the timeline.

A 30-second take is not a four-minute nightclub entrance. It is the right length for the oners that short-form video actually uses: a product reveal, a walk-through of a room, an opening that follows a character into the story, a music-video verse.

How do you plan a oner with AI?

Plan it exactly as a crew would: as a path for the camera with events placed along it. Write each event as a timed beat, with the camera's action first:

0-5s: street level at dusk, the camera follows a cyclist from behind toward a lit cafe
5-12s: the camera overtakes her and turns to face the cafe door as she reaches it
12-20s: through the door, the camera glides past tables toward the counter
20-26s: the barista looks up and slides a cup across; the camera slows
26-30s: push in on the cup, steam rising, the room soft behind

Four things make an AI oner hold together:

1. Keep one subject to follow. The engine holds a single figure through a move far better than it juggles three.
2. Hand off with motion, not cuts. Get from the street into the room by moving through the door, the way a Steadicam operator would.
3. Change speed on purpose. Oners feel directed when the camera slows for the moment that matters. Say "the camera slows" in that beat.
4. Anchor the start and end. A start frame and an end frame fix where the take begins and lands, and reference images keep the character consistent.

Our guide to image-to-video camera movements lists the moves engines understand by name, and the shot script builder times the beats and checks them against the take limit.

Is a oner always the right choice?

No. A oner that exists only because it can becomes a long wait for the point. If the scene needs a reaction that the camera cannot physically be in position to see, cut. Coverage and a timeline edit are still the better tool for dialogue, comedy timing and anything that needs a choice made after the fact. The Standard engine's short shots and the timeline editor are there for exactly that.

Frequently Asked Questions

What does "oner" mean in film?

It is crew slang for a scene shot in one continuous take with no cuts, from "done in one".

Is a long take the same as a oner?

Nearly. Every oner is a long take, but a long take does not have to cover a whole scene; it only has to run long compared with the cutting around it.

What is a tracking shot?

A shot in which the camera moves with or alongside its subject, on a dolly, a Steadicam, a gimbal, a vehicle or a drone. It can be a few seconds long and does not have to be a oner.

How long can an AI oner be?

On PrismPoster's Premium engine a single take runs up to 30 seconds. Longer sequences are built from several takes on the timeline, with each take continuing from the last frame of the one before.

How do I write a prompt for a one-take shot?

Write timed beats, one camera action per beat, following one subject, and describe transitions as camera movement rather than cuts. Our guide to AI video prompts has more examples.