Rack Focus

What is a rack focus?

A rack focus is a change of focus made on camera, during the take, that the audience is meant to see. At the start of the shot one plane is sharp, say a face close to the lens, and a second plane is soft. The focus puller turns the focus ring, the sharp zone travels through the scene, and the shot ends with the second plane sharp and the first dissolved into blur.

Two conditions make it work. First, the subjects sit at clearly different distances from the camera, which is why rack focus lives in shots built in layers (foreground, midground and background). Second, the depth of field is shallow enough that only one of those layers can be sharp at a time. On a wide lens stopped down to f/11 everything is already in focus and a rack is invisible.

It is a lens technique, not a camera move. The frame usually stays exactly where it is, and nothing changes but the plane of sharpness. One static setup does the work of a shot, a cut and a reaction. It belongs to the family of camera focus decisions covered across the lenses and focus hub, and it is the moving version of the choice every shallow depth of field frame makes.

Why do filmmakers use a rack focus?

The eye goes to what is sharp. A rack focus exploits that reflex to direct the viewer's eye with more precision than blocking or light alone.

• Reveal. Something has been sitting in the blur all along: a figure in a doorway, a weapon on a table, a face in the crowd. When focus lands on it, the audience discovers it at the same instant as the character, or before. In The Graduate (1967), focus shifts from Elaine to her mother standing behind her, and Elaine understands who Benjamin's affair was with in a single beat, without a word.
• Connection. Racking between two people in one frame links them: a speaker and the listener who reacts, a character and the person watching them. The In the Mood for Love frame below keeps both lovers in shot and lets focus decide whose face we read.
• Cause and effect. Rack from an object to a face, or from a face to an object, and you have an insert shot and a reaction in one take. The wall clock and the man in the hat in the Taking of Pelham One Two Three frame below show the pairing.

Against a cut, the rack keeps real space and real time intact. The audience watches attention travel instead of being told where it went.

How to shoot a rack focus

Build the depth first. Put the two subjects at clearly different distances: something at 0.6–1.5 m from the lens and something at 3–6 m works on most lenses. The larger the ratio between the two distances, the more dramatic the rack.

Pick a lens and stop that make the difference visible. Longer focal lengths and wider apertures shrink depth of field. An 85mm at T2 focused at 2 m on a Super 35 sensor holds only about 5–6 cm in focus, so a background at 4 m is thoroughly soft. A 50mm at T2.8 at 3 m holds about half a meter. A telephoto lens gives the most pronounced racks; a 24mm at T5.6 barely racks at all. Open the aperture and control exposure with ND filters rather than closing down.

Measure and mark. The 1st AC (focus puller) measures the distance to each subject with a tape or a laser rangefinder and writes both marks on the lens ring or on the white marking disk of the follow focus. Actors hit floor marks so those distances stay true take to take.

Time it to a cue. A rack should be motivated: a line of dialogue, a door opening, a glance, a sound. Rehearse it with the actors so the move starts on the cue, not before. Speed carries meaning. A 1–3 second rack reads as a natural shift of attention; a snap of a few frames reads as a jolt; a slow drift over 5 seconds or more feels dreamy or ominous.

Mind the lens. Many lenses change image size slightly as focus moves (focus breathing), so the frame seems to push or pull during the rack. Cinema primes are designed to breathe less.

Light both planes. The background subject has to be exposed and lit for its moment in focus, not treated as bokeh. Check the end frame on its own: it should hold as a composed shot.

Racking focus, pulling focus and follow focus: what's the difference?

Racking focus and rack focus mean the same thing. "Rack focus" names the shot, "racking focus" the action ("we rack to her on the line"). Crews also say focus pull, pull focus or focus shift. The word "rack" is usually traced to the rack-and-pinion gear that moved the lens on early cameras.

Pulling focus is the broader job. Any shot where the camera-to-subject distance changes needs focus pulled, usually invisibly.

Follow focus is that invisible version: keeping a moving actor sharp as they walk toward the lens or the camera tracks with them. The term also names the geared device mounted beside the lens that the focus puller turns.

A rack focus is the pull made visible: focus leaves one subject on purpose and lands on another. In a single dolly shot you may have both: follow focus on the actor during the move, then a deliberate rack to a detail behind them at the end.

Common types of rack focus

• Foreground to background. The classic reveal. A near face or object is sharp, then focus travels past it to someone arriving, watching or waiting deeper in the frame.
• Background to foreground. Something enters the near plane: a hand, a gun, a glass set down. The rack pulls the audience's attention forward, often abruptly.
• Object to face. An insert and a reaction in one move: the letter, then the person reading it; the ringing phone, then the person deciding whether to answer.
• Between two characters. In an over-the-shoulder shot or a two shot staged in depth, focus racks from speaker to listener, letting the edit rest while the scene keeps shifting.
• Rack in or out of blur. Starting a scene completely soft and racking into focus (or ending on total blur) works as a soft transition, often for waking, fainting or memory.

Rack focus vs. split diopter

Both techniques solve the same problem, a near subject and a far subject in one frame, in opposite ways.

• A rack focus keeps only one plane sharp at a time and moves between them. The audience reads the two subjects in sequence, and the order is part of the meaning.
• A split diopter is a half-lens attachment that lets one side of the frame focus close and the other far, so both subjects are sharp at once, with a soft seam where the two halves meet. Brian De Palma made it a signature in films such as Blow Out (1981).
• Deep focus does it with optics and light: a wide lens and a small aperture hold everything from near to far sharp, as Gregg Toland did for Citizen Kane (1941).

Choose the rack when the discovery matters. Choose the split diopter or deep focus when both people must be read in the same moment and the tension lives in the distance between them.

Each clip below is one continuous shot built around a focus pull (two pulls in Murder in the First and Munich). Watch the blur rather than the actors: see which plane is sharp in the first second, when the softness starts to travel, and what the newly sharp plane reveals.

Image models can draw either end of a rack focus, as long as the prompt names which plane is sharp and how far the other one sits behind it. A video model then has to change only the focus, which it will do only if the prompt locks the camera and describes the blur moving. Below, Nano Banana 2 Lite builds the start frame with focus on a foreground object, GPT Image 2.5 draws the landing frame of a rack between two characters, and MiniMax H3 Max Turbo animates the first frame into the pull. In FlashBoards, keep the film references, both stills and the video side by side on one board. The fourth prompt re-shoots a real film shot from a text description alone, and the original plays next to the result so you can compare.

Common mistakes

• Everything is sharp. Image models default to deep focus. Name the lens and stop (85mm at T2), give distances for both planes and say which one is sharp and which is soft.
• Both planes blurry. Too much blur language and the model softens everything. Put crisp sharp focus on the subject that should be sharp, and soft only on the other plane.
• A push-in instead of a rack. Video models often read focus on the man as move toward the man. Lock the camera (static tripod shot, the frame stays still) and describe only the sharpness changing.
• The rack happens too early or never. State the order and speed in plain words (a slow rack from the key to the man, taking about two seconds) and tie it to an action cue like a door opening.
• A new face appears when focus lands. If the blurred figure is not described in the start frame, the model invents one as it sharpens. Describe the background person fully in recipe 1.
• The wrong plane blurs. Some takes swap the planes. Generate several and keep the correct end frame next to them on the board as a check.

Further reading

Steven D. Katz, Film Directing Shot by Shot (Michael Wiese Productions). · Blain Brown, Cinematography: Theory and Practice (Routledge). · American Cinematographer Manual (ASC Press), sections on depth of field and focus.