A big thanks for these instructions.
I wasn't suggesting to the original poster that I knew how to test a lens myself,
so I'm glad you have taken the trouble of explaining it all here.
Kind regards.
You're welcome! But I wouldn't say that's all... It's just a reasonable start. I can offer a bit more about my favorite two ways to find the correct AF Fine Tune value for a lens (btw, sorry I interchange "fine tune" and "micro adjust," because I can never remember which term is Canon and which is Nikon, but I mean same thing by it).
First, when you're in a hurry, this is the quick and dirty way to get an OK fine tune value:
1. Put camera on tripod, point at an easy target, a target that complies with the camera manual warnings about AF and would make Leonard happy (choosing a target that Leonard would approve of is the most important thing, because otherwise Leonard will declare your entire process invalid! ;-)).
2. Put the camera in Live View, and use Contrast Detect AF to focus on the target. The logic here is that Contrast Detect when on a tripod is generally the most accurate thing of which the camera is capable, and, more importantly, Contrast Detect AF is absolutely exempt from the type of calibration error that you're trying to offset in the phase detect system.
3. Turn off Live View, and, while looking at the distance/focus scale on the lens, auto focus the camera the normal way. If the focus scale does not move, it means there was no difference between Contrast Detect and regular AF (aka phase detect), which means you're done! If however it moved, that means the regular AF was off, and you need to enter a value. You can tell whether the value needs to be postive or negative based on the direction the focus scale moved.
4. Enter a value based on the movement you saw, and repeat these steps until there is no movement between contrast detect and phase detect.
The reason that is the "quick" method is that it eliminates having to retreat to a computer to evaluate the files, which is method 2. Also with the quick method, lenses that have very short focus throw might move a slight amount without you noticing.
So, the slower, better way is this:
1. Step 1 same as above.
2. Turn on AF Fine Tune, and set to -20.
3. Take 5 shots, using the manual focus ring to defocus after each shot, so that you know for sure each shot is a new focus attempt on part of the camera. For best results, shoot NEF.
4. Move your AF Fine Tune setting to -15, and repeat step 3, and so forth in 5 point increments until you get to AF Fine Tune setting of 20. This should give you 45 shots per lens tested.
5. On a computer, evaluate the NEF files. Be sure to actually render them as NEF, and do not view them using the embedded preview JPEGs, because the embedded preview in a NEF file is lossy to save space. As an alternative to viewing them as rendered NEF, you could convert them to hi-rez JPEG or TIFF to get the benefit of whatever steps you normally do during conversion... Now look for two things: Look for the setting at which the image quality is best. Because the variations can be very slight, it is entirely possible that you might find a tie, which would indicate the best value actually lying between those two 5-point intervals. If you do this to all your lenses, depending on which you have, you may also see that some lenses react more strongly to fine tune on a per-point-of-adjustment basis than do others; this is normal... The second thing you want to look for in my opinion gives you a more accurate read on the correct setting for a lens: Look at the entire range of shots, and try to find the outer point at which the shots are
equally bad . In other words, does -10 suck equally with +20? If so, this might help point you to a value of +5 for the lens; by looking for the center of the lenses adjustment continuum on a wider basis, you can corroborate what you think is the absolute best value. This also helps average out the effect of shot-to-shot focus variations, and can help point to odd values like +3 that don't fall on your 5-point test interval. Some lenses may demand retesting at single-point intervals to get the absolute best setting, but in my opinion that is rare, with 5 points being more than adequate.
--
David Hill
http://www.bayareaweddingphotographer.com
San Francisco & San Jose, CA | Austin, TX
Wedding Photographer and Apparent Gearhead
