How to focus the view camera

Part 2 of Setting up the view camera

by Q.-Tuan Luong for the Large Format Page


Summary: two visual, iterative methods to focus the view camera on a chosen plane of arbitrary location, assuming that you want to maximize sharpness, as in landscape photography. There are probably as many visual techniques for focusing when employing a Scheimpflug relationship as there are photographers doing so.

In practice, what I personally use is procedure I (focus on far, tilt on near) when photographing mostly flat/planar subjects. When the subject is tall, most of the time I use procedure II, going directly to step 2 (spread focus between far and near). If this results in a very small aperture (f45 or worse), I might try the full procedure II (minimize focus spread with tilt) to get a larger aperture.

I recommend using a metric scale. It is also necessary for the optimal determination of the f-stop.

Principles

The Scheimpflug rule (named after an Austrian Army officer who patented it in 1904), states that the film plane, the subject plane (plane of sharp focus), and lens plane (the plane through the optical center and perpendicular to the lens axis) must converge along a single line. See references.

We all know that when a rigid camera is focused, all the objects which are at a same perpendicular distance from the film plane can be brought in focus at a time. For a rigid camera, the film plane and the lens plane are parallel (or equivalently intersect at infinity). So the subject plane must also be parallel to these two planes.

In a view camera, the geometric relationship between the film plane (the back standard) and the lens plane (the front standard) can be adjusted. This makes it possible to focus on virtually any plane, be it receding or slanted.

To use the Scheimpflug rule, here are a few points to keep in mind:

Which controls to use ?

Procedure I

  1. Estimate the best plane of focus.
    This is the tricky part, and requires judgement. Once this is done, the rest of the procedure is mechanical.

  2. Adjust the tilt and/or swing and focus

    To adjust the tilt, use the following. To adjust the swing, replace "top/bottom" by "left/right".

    Variations of this technique:

  3. (Re-)adjust the focussing point
    Visually, you adjust this point so that the most blurry close point and the most blurry far point would be equally blurry.

    Procedure II

    This method requires that you have a millimeter scale so that you can measure the difference, in millimeters, between the near and the far focus points. By the way, metric units makes calculations much easier. Most monorail cameras come with such a scale, and on several flat-bed cameras (including the Tachihara, Technika, and Canham KBC), it is easy to attach one. You just have to make sure that it doesn't slip during focussing, and that you have a reference point on your focussing rail from which to take measurements. Having a pointer (like on the Technika) helps in making precise measurements. If it is not possible to attach a scale, you can always take measurements with a ruler.

    The procedure is described for tilt adjustment only. Same considerations as before apply for swing. The idea is simple: by successive trial and error, you will determine the tilt which is such that the focus spread is minimized (step 1). Then you determine the optimal focus point (step 2).

    The idea behind step 1 can be used without applying the full procedure. For instance, you have a distant landscape with some tall trees in the foreground. Since this is not planar, you'll have to stop down. If you shoot without movements, focussing somewhere behind the trees, the trees and the horizon will be out of focus and require stopping down. If you tilt the lens, focussing somewhere on 2/3 the height of the trees, the bottom and top of the trees will be out of focus and require stopping down. Which of the two alternatives is the best ? The answer is given by measuring the focus spread for each of them, and seeing which one is the smallest.

    1. Ajust the tilt
      • Read the initial total focus spread.
        • Focus on the nearest point. Note the position A of focussing rail which corresponds to the maximum extension for any point in the image.
        • Focus on the furthest point. Note the position B of focussing rail which corresponds to the minimum extension for any point in the image.
        • Read the difference D between (A,B).
      • Make a guess about the amount of tilt needed, and apply it. One possibility, now that your camera is focussed on the far, is to tilt until the near is sharp.
      • [F]Read the new total focus spread.
        • Focus on the closest point (highest point above the plane of focus). Note the position A of focussing rail which corresponds to the maximum extension for any point in the image (you might have to try several points and note the maximum value).
        • Focus on the furthest point (lowest point below the plane of focus). Note the position B of focussing rail which corresponds to the minimum extension for any point of in the image (you might have to try several points and note the minimum value). In the rock/mountain example, this is now the base of the mountain.
        • Read the new difference D between (A,B).
      • [T] Adjust the tilt by a small amount.
        • If the difference D has decreased, add tilt, repeat [F]
        • If the difference D has increased, remove tilt, repeat [F]
        • If the difference D remains the same, you are done. Note that if the subject is planar, D will be zero.

    2. Adjust the focussing point
      • Focus on the closest point (highest above plane of focus), which you want to make sharp. In the rock/mountain example previously used, this would be the top of the rock. Note the position A of focussing rail.
      • Focus on the furthest point (lowest below plane of focus) which you want to make sharp. In the rock/mountain example previously used, this would be the base of the mountain. Note the position B of focussing rail.
      • Focussing at the median of (A,B) will make the closest point and the furthest point equally sharp.
        Note that this works regardless of camera movements (tilts/swings, translations), focal lengths, and formats. To be exact, the general result (referenced in Paul Hansma's article, and detailed by Leonard Evens, see references) is that you focus on the point C such as the ratio of distances d(C,A)/d(C,B) is (1+MB)/(1+MA), where MA and MB are the magnifications associated with the close and far object. This point is always closer to A than to B. However, the only case when the median rule is not a good approximation is when the two magnifications are significantly different and at least one of them is comparable to 1 which is only the case for close-ups with very wide lenses, so in practice one needs not worry about this technicality.
        • Making he closest point and the furthest point equally sharp is optimal if the far point is not at the horizon. If you had applied movements, this would always be the case.
        • If you have maintained the standards parallel and the far point is the horizon, you should instead focus closer to infinity than to the near point (the two thirds of (A,B) is practical for determining the f-stop to use). This is because perceptually, a higher resolution is required for the horizon to appear acceptably sharp than for the foreground. Note however that Ansel Adams recommends, if sharpness has to be compromised that the nearest point be sharper.

      Alternatives methods and references

      The two methods that I have described are visual and iterative. There are other approaches which try to nail down the correct tilt directly. All of them require a way to measure tilt angles.

      One class of methods are based on measurements on the camera and computations. They require a calculator such as the Rodenstock tool or Bob Wheeler's Vade Mecum which runs on palm devices. Both are detailed in Bob Wheeler's Photographer's Aids: A survey. In particular, Bob Wheeler gives a in Notes on view camera geometry a practical rule for use in the field. Wheeler's rule states that the angle of tilt is 60*delta_focus/delta_GG, where:

      • delta_focus is the focus spread, the distance on the focussing rail from A to B
      • delta_GG is the perpendicular distance between the images of the near point and the far point used to define the plane of focus, imaged on GG before applying tilt to the camera.

      There is also a method developed by Harold Merklinger in his book "Focusing the View Camera", and well summarized on his web site. This site in particular has a few brilliant animations which are very helpful in visualizing the Scheimpflug. Merklinger points out to a rule complementary to the Scheimpflug, called the "hinge rule". His method calls for the measurement of the distance J from the lens to the required plane of sharp focus, measured parallel to the film, and the use of a table to get the tilt value. Personally I think this a great geometric analysis, but that in practice it is not easy to apply because J is not that easy to estimate. For a discussion on the practical feasibility of this approach, see this QA forum thread and this one. If you want to try to use this approach, this Excel file might help.

      For the historically inclined, Harold Merklinger's page has the 51-page Theodor Scheimpflug's 1904 British Patent. For the geometrically inclined, I sketched an elegant (I think) geometric proof of the Scheimpflug's principle using Desargues theorem. An essentially identical derivation is given in Bob Wheeler's excellent Notes on view camera geometry. Generally speaking, these notes are some of the best exposition of the math behind view camera operation that I have seen.

      Subsequently, Emmanuel Bigler sent me a new derivation (Scheimpflug's rule: a simple ray-tracing for high school ?) which requires only high-school level (well, French high school from the 60s) geometry. He then goes on to show geometrically, using an absolute minimum of algebra how to derive the position of the slanted planes defining the DOF area (Depth of field and Scheimpflug's rule: a minimalist geometrical approach).

      A number of great technical/mathematical notes can also be found on this page by professional mathematician and photographer Leonard Evens. Of particular interest is the thorough View Camera Geometry. Originally intended as a relatively short article for the American Math Monthly, it soon got too large for that, approaching book length at 105 pages (in the spirit of Wheeler's notes). It does address a whole lot of issues that arise in the use of a view camera from an analytic, mathematical point of view. That page has also the more focused, but formula-free (and therefore more accessible to those familiar with geometry) Depth of Field for the Tilted Lens.

      Howard Bond describes with plenty of details his well-proven method in his article, Setting Up the View Camera (reproduced on this site). The idea of biasing the focus towards infinity in landscape photography is discussed by Harold Merklinger's The INs and OUTs of FOCUS, and in the article by Joe Englander, Apparent Depth of Field: Practical Use in Landscape Photography.


      View or add comments