Here is what I’m trying to accomplish with this test regime.
Fairly easily performable by non-experts. We’ll see how well I’ve accomplished this goal. I’ve been surprised and disappointed by the number of people who have been unable to follow the simpler instructions in my Simple Decentering Test, so I’m not totally confident. Nevertheless, I have hopes. The exposures can be made in a five minutes, once the target and camera are set up. Postprocessing for a set of images takes another ten minutes, using Lightroom.
High tolerance for setup departures from perfection. This is really another aspect of ease of performing the test. Alignment is not critical. Target distance is not critical once a minimum is attained (and, truth be told, even that minimum distance can be shaded without much penalty). Exposure is not critical; all that is required is that no part of the target be clipped.
Setup errors manifest visually. One of the biggest problems with using full frame flat targets is that setup errors can produce the same visual effects as lens misalignment. In this test, many setup errors are detectable from their effect on the captures. Misfocusing, for example, will show up with good lenses as an absence of aliasing and false color in the target-centered capture. One error which will not stand out as such and can lead to false positives is a shift in lens focus distance during a single set of captures. For this reason, I recommend that the test be repeated if the results point towards a misaligned lens. It is highly unlikely that any jarring of the lens will result in the same focus shift pattern in the second test.
No calculations need to be performed by the tester. It is true that there are some calculations involved in the test, but I’ve performed them in advance for a variety of cameras and posted the results here. All the tested need do is find the right graph, and read off the minimum distance for the chosen camera, f-stop, and focal length to be tested.
One target can test most lenses. If the target is large enough, you only need the one. But a target for a very fast, very short lens can get large and unwieldy. For most focal lengths and f-stops, a single 17-inch Siemens Star target should suffice. For the more demanding cases, I have provided graphs that allow the tester to see how big the target needs to be.
Visual analysis of results. There are many lens testing regimes that incorporate a computer analysis of the captured images. They have advantages, the chief one being quantitative results. However, they require that the tester learn to use a computer program as well as gain an understanding of some moderately-sophisticated optical, engineering, and mathematical concepts. This test is different. You can see what’s going on just by looking at the images with the proper mindset. That makes it doable by a far greater range of photographers.
Detect defects with some specificity. This test is capable of detecting field flatness errors, field tilt, astigmatism, coma, spherical aberration, and other lens defects that result in degradation of off-axis sharpness. If any of these is suspected after the first test, it is sometimes possible to construct related tests to figure out what’s going on. For example, corner softness could be the result of field curvature or field tilt, or something else. Running another test with focusing in the corners instead of the center can tell you which of those three things you’re up against. This part of the test is a work in progress. With the help of Brandon Dube, a lens design expert, I hope to attain more precision here.
Work with almost any raw development workflow. Learning a new raw converter or a new image-editing suite is a lot of trouble. I wanted a test that would work with whatever software the user is comfortable with, and I think I have one.