A double-edged sword in my life is the quantity of cameras and lenses that I purchase, receive, evaluate, use, review ... I love getting to use this gear, but the logistics definitely get complicated - and my credit line is generally not happy with me as nearly every camera and lens reviewed on this site to date has been purchased retail. Owning each camera and lens outright leaves me personally on-the-line financially for verifying the operation of each within the retailer return period. Or at least within the manufacturer warranty period. Important shots missed due to equipment problems are also potentially costly to me.
Upon receiving a new camera or lens and after a visual inspection, the first test I perform is a focus calibration test. While I seldom have a problem with Canon or Nikon cameras and lenses, focus calibration issues are the primary defect I find - especially with third party AF lenses. If a camera or lens does not focus where desired, the images are often not usable. So, an out-of-calibration defect is a serious issue to have. Therefore, I test for it.
How do you focus test a camera and lens combo? I'll interrupt my question to emphasize the point that this is a camera and lens combination test. Variations can occur between various lenses and camera bodies. Case in point is that my Canon EF 400mm f/2.8 L IS USM Lens and my Canon EOS 5D Mark II DSLR do not play nicely with each other - back-focusing is the result. The same lens works great on all of my other bodies and the camera works great with the many other lenses I've used with it.
OK - back to the question: How do you focus test a camera and lens?
You first need a good focus target. I want a target that is flat and large enough to more-than-cover the selected AF point (I usually test with the center AF point) at the distance I am calibrating at. This distance is typically the distance I expect to use the lens at - but not near an infinity focus distance. The target needs to have good contrasting colors to provide to the camera's phase-detection AF system. I also prefer that the focus target stands up by itself. Along with a focus target, a measurement scale must be running through the frame near the target - it must be visible from foreground to background. A tape measure, sidewalk, carpet ... anything with a little detail in it will work for this requirement.
Good shooting technique must then be employed. This means a stable camera position (a tripod-based setup with mirror lockup enabled and a remote release or self-timer is best) with the sensor plane parallel to the focus target. A shutter speed fast enough to stop any remaining camera movement is required and strong lighting levels should be present (direct sunlight is a good choice, office lighting brightness is the minimum). One Shot AF mode should be selected. ISO 100 is preferred. A wide open aperture will provide the shallowest DOF (Depth of Field) and is what I typically select for focus calibration testing. Keep in mind that a lens that exhibits focus shift will show different autofocus distances at different apertures.
A zoom lens should be tested at both ends of the focal length range and perhaps at a mid-focal length.
Shoot the focus test. From a slightly defocused setting, allow the camera to focus on the target and take the picture. Then analyze the results. Contrast and sharpness can be increased to emphasize the plane of sharp focus. Properly calibrated, a camera and lens combo should focus accurately on the focus target - not in front of or behind it.
If necessary, the manufacturer can make focus adjustments in either the camera or the lens (or both). The easiest adjustments are those made using the AF microadjustment feature found in today's better DSLR cameras. In-camera focus calibration generally should be done at the longest focal length in a zoom lens. If the wide end of the focal length range remains mis-calibrated, the manufacturer's repair service should be employed for the adjustment.
To determine if it is the camera or the lens that needs adjustment, try other lenses on the suspect camera (or try the same lens on other cameras). To be completely sure your kit is properly calibrated, send both the camera and the lens to Canon Service.
Here is a real life example. For some time now, I've scientifically employed an ice cream cone box for my focus calibration testing. I simply put it somewhere and shoot it. Here is an example:
This test was shot with the Canon EOS 1Ds Mark III and Canon EF 70-200mm f/4 L I USM Lens at 200mm, f/4. I had just received the camera back from service and needed to verify that AF was properly calibrated. This image is from the actual test I performed for that verification. Looks perfect to me - the plane of sharp focus is aligned with the front of the box. The plane of sharp focus would be in front of or behind the box if focus calibration errors were present.
But this test is not exactly perfect. Notice that, even though I focused at the bottom of the box, there is a tiny amount of angular error in this specific test. Better would have been to place the box on a table and add a tape measure or other calibration scale to the setup. The rule would give me a more precise measurement device than the carpet fibers
Enter the Datacolor SpyderLensCal - or as Datacolor calls it, "The smarter focus tool".
A few months before writing this review, I received the Datacolor SpyderLensCal press release along with an offer to receive an evaluation sample. Well, the SpyderLensCal was a far more elegant and precise solution to my focus calibration testing than the ice cream cone box, so I of course said "Sure!".
The compact, lightweight, folded plastic Datacolor SpyderLensCal unfolds into a base, a focus target and a calibration scale that gravity-hooks at a 45° angle onto the focus target in its 90° position.
The base has a metal 1/4-20 thread tripod insert and an integrated level for precise and solid setup. The flat target has an ideal-for-focusing, sharp, high contrast, black and white design. The evenly black and white design also aids in getting a correct autoexposure from the camera.
The calibration scale has six numbered lines on each side of the 0 position with 10 shorter marks between each numbered lines and an even shorter mark between each of these. At a 45° in-use angle, the distance between the numbered marks from the camera's perspective is 10mm. The distance between the shorter marks is 1mm and the distance between the smallest marks is .5mm.
Setup is fast. Simply open the device and place it on a flat surface or on a tripod, light stand or other similar stand. Then shoot it in the lens focus calibration test methodology I described earlier. The results are clear and accurate as demonstrated below.
The above example was shot with the Canon EOS 5D Mark II and Canon EF 200mm f/2 L IS USM Lens at f/2 using tungsten studio lights. You obviously want the line at "0" to be the most-sharp in your test. Note: these are reduced-size images - results are easier to see at 100% resolution as shown below.
If you have a DSLR with AF microadjustment capabilities and focus adjustment is necessary, adjust your camera and try the test again. Repeat if necessary. Repeat for all of your lenses - the improvement in their performance may surprise you.
If you camera does not have AF microadjustment capabilities and focus adjustment is necessary, a call to the camera and/or lens manufacturer's service center is likely necessary and is your best option.
I answer a very large volume of email and one of the most common requests I receive is for help in determining why images are not sharp. If they are using a good quality lens, focus calibration testing is one of the tasks I typically have them perform (as long as their shooting parameters and technique are sound). The Datacolor SpyderLensCal makes this task easy - and the results are precise.
The Datacolor SpyderLensCal obviously costs more than my ice cream cone box (and its contents are not nearly as tasty) and owners of a single lens may not be able to justify the cost of this tool. But those who own multiple lenses and those who want the ultimate in image sharpness may want to consider adding a SpyderLensCal to their kit for the improved image quality their accurately calibrated lenses will give them. This is definitely the tool I will be using for my focus calibration testing.