Leitner's Cinematography Corner

The author at DuArt Film Laboratory in New York in the early 1980s, demonstrating the first lens test projector in New York.

The term “back focus� gets thrown around a lot, mostly by people who little understand it. What, then, is auto back focus?

First let’s cover some lens basics.

Lenses are two-way streets. Light can enter or exit either end. That’s what makes lens test projectors possible—the type you encounter in camera rental houses. They’re basically slide projectors, with a super-precise focus chart deposited on a glass slide. The camera lens takes the place of a projection lens.

What’s great about a lens test projector is that you can fiddle with focus, iris, or focal length (in the case of a zoom) and see with your own eyes what sort of optical gremlins reside within.

For instance, stopping down the iris, even slightly, often crispens the image by choking off residual spherical aberration—which you can easily notice as an improvement in contrast and edge sharpness.

Focus is trickier. Lens test projectors often live on carts that are moved on rails closer to or farther away from the projection screen. Let’s say you set the focus mark of a lens mounted to a test projector to 10ft. You would expect that when the lens is 10ft. from the projection screen, focus would look good. But what if it doesn’t? What if you slide the cart forward and the lens comes into focus at 8ft.?

Richter Cine Equipment Precision Reticle Projector, circa 1982, with Aaton adapter and Zeiss Super Speed. The author’s was serial number 0001.

An interchangeable lens must be mounted to a camera in several precise ways. The optical axis of the lens must be perfectly perpendicular to the sensor (whether film or electronic) and coincide with the precise center of the sensor. If not, one side or the other of the image will tend towards softness, and in the case of zooms, the center of the image will drift during zooming.

And there’s an additional dimension to lens mounting that demands even more precision. The lens must be seated at a very exact distance—accurate to within microns—from the surface of the sensor in order for the focus marks to track accurately. When a lens is mounted sloppily—too close or far from the sensor—two problems arise.

In the case of all lenses, prime or zoom, focus marks grow unreliable. Of course, you can always set focus by eye and often get a passable result. But the focus indications on the lens won’t match tape-measured distances.

A more unnerving problem arises with zooms, which by definition enable a continuous range of focal lengths. They’ll look fine at telephoto focused by eye, but when zoomed wide, it’s blursville. Guaranteed. Why?

Everyone who owns a camera knows about depth of field. It’s apparent to any viewer looking at the resulting image. Some things are in focus and some things aren’t. But as mentioned above, lenses are two-way streets. What is the subject (in front of the lens) and what is the image (behind the lens) depends on what is front and what is back—i.e., which way the light is traveling.

If depth of field exists at one end of the lens, it must exist at the other end too. Which, in fact, it does. On the image side, it’s called depth of focus.

Like depth of field in front of the lens, an image formed by a lens is not truly a “planeâ€? but rather an uneven zone—a miniature “depthâ€?—within which a sharp image will form on the surface of a sensor if the sensor resides precisely inside this tiny zone.

If an interchangeable zoom is precisely positioned by its mounting system at the specified height above the sensor, no problem. All is well with the world. If it is not—if “back focus is offâ€?—degrees of unwanted soft focus are introduced under various conditions.

Stopping down the iris increases both depth of field and depth of focus. If sufficiently stopped down, a zoom or prime with an incorrect back-focus adjustment can seem to focus properly. However, shooting wide open, as most of us do often, virtually showcases back-focus errors (as well as aberrations inherent to the lens design).

Focal length also matters. Long-focal-length lenses produce artful narrow depth of field but maximum depth of focus. (It’s difficult to badly mount them.)

Here’s where the trouble starts.

Short-focal-length lenses, on the contrary, also known as wide-angle lenses, produce endless depth of field (think Gregg Toland’s “deep focus� in Citizen Kane or any Super 8 camera) but dangerously shallow depth-of-focus.

When an improperly mounted zoom is focused at telephoto—a long focal length, that is—the image will appear sharp in the viewfinder. As the zoom is widened to a short focal length, depth of focus will shrink and shift away from the sensor surface, resulting in a soft image.

Now consider this: “Normal� perspective in 35mm Academy is associated with a 50mm lens, while the same perspective in standard 16mm is associated with a 25mm lens. Just how short is the focal length needed for normal perspective in a 1/3in. camcorder? Darn short.

See where this is going?

Interchangeable zooms on 1/3in. camcorders require even more precise mounting than 2/3in. camcorders, or 16mm or 35mm cameras.

In other words, the smaller the format, the more critical the back focus.

Next week: Adventures in Auto Back Focus, Part 2, featuring Sony’s PMW-EX3 and HVR-Z7U.