>Could you explain how do you manage these bodies, properly >adjusted, to poorly focus on the test chart using the >procedure I described? That way, I can maybe workout a better >one . It would be great if you could post an example,maybe >in the original thread not to pollute the current one...
It's easy to cause a mis-focus. A large percentage of the time, the focus point that a tester has positioned on a 45 degree target sheet is not actually locking on at that chosen focus point. The camera may signal a confirmation (with the focus dot), but the resulting shot appears mis-/front-/back-focused because the lock signal was an approximation. This has been demonstrated many, many times in other threads on Nikonians - most or all of which predate your other thread on the subject. Nothing new here. Once I've done somewhere between 40-50 shots with a single setup, the variability begins to show up in the results charting. Once I've done upwards of 100 test shots with a single setup, the AF system variability is incontrovertible. If someone wants to go to those lengths to find an AF fine tune setting that can improve overall performance, go for it. But test shot sessions of 10, 20 or even 30 shots with a single setup usually don't tell the photographer anything useful. In the middle of testing on a very hot or very cold day, open a window to either heat up or cool down the room and watch the variability range shift every so slightly.
Considering that a moderately well used Nikon DSLR in the hands of an active amateur photographer might be used to shoot a thousand photos a month, AF fine tune targeting results based on a minimum of 100 shots with a specific set up is a basic drop in the bucket. But it's enough to get a real sense of how well or how poorly a particular AF system is performing. Most people who do such extensive AF test target shooting using parallel/plane setups, usually then back away from any AF fine tune adjustment because they're suddenly more cognizant of the inherent variability of the AF system in the first place.
I hesitate to contribute samples to the discussion of a calibration methodology with which I fundamentally disagree. The reliance on focus point positional confirmation in software such as Capture NX/NX2 is also problematic. I'm not suggesting for one minute that our cameras and the post-processing software we use is totally problem ridden - far from it actually. I'm just suggesting that I haven't found a testably duplicable AF system problem in something like 85 different camera bodies (or more, if I also take into account Sony, Olympus and Panasonic bodies). I've just found the occasional broken camera plus a couple of DOA units.
Maybe we're talking about two different standards here. I'm satisfied with all of the photos I publish in my gallery on Nikonians and on Photo.Net. Some photographers may feel that my published photos could have been even more sharply or more accurately focused, and maybe I should have expected a to get a higher percentage of keepers out of the cameras I use most often. I accept those arguments, especially if the degreee to which such photographers examine focus accuracy and sharpness is criticial to their photography enjoyment and satisfaction. For photographers with the latter approach, there may be a determined intent to always eke every last bit of accuracy from their chosen cameras and lenses. Can't argue with that either - the philosophy and the approach is perfectly sound.
The only thing I'm questioning is the use of test methods (not just yours - the commercial retail test kits are totally suspect IMO, for the reasons mentioned above) containing so many set up variables of lighting, ISO, chart positioning compromises, chart/target-to-camera-frame ratio, chart target point/edge size/dimension vs. camera focus sensor size and response characteristics, chart target point edge definition, horizontal/vertical/cross-type sensor differentiations, camera positioning, chart angle, sensor plane, lens choice, test result repeatability and so on. Then there's the inherent mechanical and optical variability of the particular AF system and chosen lenses which grossly affects testing result repeatability. Again, none of this information is new.
In fully controlled conditions, all I've ever been able to find is uncompensatable variability in an AF system with a given lens. The variability is only reflective of the tolerance range of the particular AF system design. High numbers of sample shots also reveal not only AF system variability but the variability of the lens mechanism as well.
Regression to the mean certainly says that sooner or later I'm going to be using a camera and lens combination which can benefit from AF fine tune. But I won't base a final decision on the results gleaned from tests done with a 45 degree target. Such targets are bascially compromises that are relatively easy to produce, package and sell. By contrast, setting up a series of absolutely parallel/plane targets in a home testing environment (on a parallel fore & aft rail system, with the camera mounted on a lateral positioning rail system), identical lighting for each target position, etc., etc., is expensive, space consuming, time consuming and usually highly impractical even for the most ardent hobbyists, enthusiasts and pros. And even if someone sets up such an elaborate testing system, a left-focus-sensor miscalibration such as that which occured in certain series of D800 bodies is still not remediable with the use of AF fine tune. That last part is a bit of a tangent - sorry.
The only time I've ever had to recalibrate any particular camera body and lens combination for focus accuracy is after the pair suffered damage on a research trip or from my own carelessness. Trip to Nikon Canada (in Mississauga, Ontario - just west of Toronto - very handy), payment of $250, and a perfect combination about 10 days later. A very rare occurence though because I try not to make a habit of dropping, hammering, banging around or abusing my gear. It happens anyway.
Somebody straighten me out on all this. I think Nikon added AF fine tune mainly as part of endless feature creep. I also know that several Nikonians have demonstrated successful AF fine tune testing and adjustments. But I also think the actual need for AF fine tune is so rare that the feature might as well not be there and that almost every Nikonian who has tackled AF fine tune (successfully or not) has had to battle the difficulties in focus target setup. At least one AF fine tune session (by a Nikon shooter who is local to me) resulted in a pin/tack/razor sharp zoom lens from 30-45mm but terrible results from 24-30mm and worse still from 45-70mm. If he had made the (I strongly believe) requisite number of shots sufficient to reveal the AF system variability in the camera/lens combo, he'd still have a very good zoom lens that (was) very sharp through its focal range. With that particular lens he was delighted with the improvement from 30-45mm because that's the focal range in which he spends most of his shooting time. But he rendered the rest of the lens' capabilities messed up. Seems counterproductive.