Rack focus in AI video: shift focus on purpose

A rack focus is a shift of sharpness within one continuous shot: focus travels from one subject to another while the rest goes soft. In AI video you don't steer it with a focus ring, you steer it with words. You state where the sharpness starts, where it ends, and that the background stays blurred. Below is how to get that shift into your prompt reliably.

What exactly is a rack focus?

A rack focus, also called a focus pull, moves the plane of focus from one object to another during the take. On a film set this is a dedicated job: the focus puller, usually the first assistant camera (1st AC), turns the focus ring live during the shot. It's a classic device. Cinematographer Conrad L. Hall used it in films such as Road to Perdition (2002), for which he won a posthumous Academy Award for Best Cinematography.

In AI video there is no focus ring. The model decides sharpness based on your description. So you name the shift explicitly instead of trusting the model to add it on its own. Modern video models know the term: prompts for Veo 3 and Kling respond to phrasing like "slight rack focus onto the subject's face", and Runway offers fairly precise focus control.

Why a rack focus strengthens your story

A rack focus guides the viewer's eye to exactly where you want it. By moving the sharpness you say, without words: look here now.

• You connect two elements in the same shot, for example a hand holding a key and the face reacting to it.
• You reveal something: first you see the foreground, then the person in the background snaps into focus.
• You add rhythm to a static shot without moving the camera.

It's subtler than a zoom or a quick cut, and that's exactly why it feels cinematic.

How to describe a rack focus in your prompt

Describe a rack focus as a movement with a beginning and an end. Name the starting point of the sharpness, the end point, and keep the background soft. A workable structure:

1. Start and end focus: "sharp on the coffee cup in the foreground, then the focus shifts to the face in the background".
2. Direction: from foreground to background, or the other way around.
3. Pace: a slow, gradual shift reads calmer than a fast one.
4. Shallow depth of field: name a shallow depth of field or a portrait lens with a wide aperture, so the contrast between sharp and soft is strong enough.

A compact prompt line: "rack focus, shallow depth of field, focus slowly shifts from the hands in the foreground to the face in the background, 85mm look". Test variants like this in the video generator and change one thing at a time. Not sure how to phrase it cleanly? Let the prompt generator write the sentence for you.

Why you need a shallow depth of field

A rack focus only stands out if the rest of the frame blurs enough. That calls for a shallow depth of field, and in photography that means a wide aperture: a low f-number. According to Digital Photography School, shooting at f/2.8 or f/1.4 produces the most blur, and a low f-number between roughly 1.4 and 5.6 gives you a shallow depth of field. With a deep depth of field everything stays sharp and shifting focus barely registers.

You steer this in AI by adding terms like "shallow depth of field", "f/1.8" or "85mm portrait lens". Also make sure your scene has depth layers: a clear foreground and background give the focus something to travel between. If needed, build that depth first in a start image with the photo generator and then bring the image to life.

Common mistakes and how to avoid them

• Everything equally sharp. Without a shallow depth of field there is nothing to shift toward. Name the aperture or the lens.
• No depth layers. If foreground and background sit at the same distance, the shift disappears. Place your subjects at deliberately different distances.
• Too fast or too often. One calm shift per shot reads better than three jittery ones.
• Too much at once. Don't combine a rack focus with a big camera move in the same short clip; the model gets confused.

Because you pay per render, you can cheaply test a few variants on a budget model and only finish the best one on a premium model.

Frequently asked questions

What is the difference between a rack focus and a dolly zoom?

With a rack focus only the sharpness moves: the camera stays still and the composition stays the same. With a dolly zoom the camera and the lens move at once, which warps the background. A rack focus is subtler and easier to generate reliably.

Can I make a rack focus from a single photo?

Yes, with image-to-video. Pick a start image with a clear foreground and background and a shallow depth of field, and describe in the prompt which plane the focus should move to. Without depth layers in the start image, the shift has little to work with.

Why does my focus shift fail?

Usually because the frame is too flat (too much in focus) or the shift moves too fast. Add a shallow depth of field, place your subjects at different distances, and ask for a slow, gradual pull.

A rack focus is one of the cheapest ways to make your AI video feel more cinematic: one line in your prompt, no extra editing. Want to try it? Create an account and test your first focus shift.