D500 Auto AF Fine Tune - Poorly Implemented

[James Farrell]

I've had my D500 for 10 days. Due to other personal issues, I've not had an opportunity  to get out "into the field" and put the D500 through its paces yet. With that said, I really, really like this camera and its features (and I plan now to buy a second one). The 350 or so shots I've taken are of the test variety - mostly trying to fine tune my lenses (e.g. 200-400/4 VR-2; 300/2.8 VR-1; 80-400G AF-S VR and others)

The only issue that really, really disappoints me is the auto AF fine tune procedure (AFFT). What Nikon should have done or could do (via firmware) is to add a step in which one would select "AF Test Mode" or similar in the setup menu (or someplace) where you would select LV AF Fine Tune or similar. Then one could complete an AFFT action using just a remote shutter release. In other words, you select the AFFT test mode (from a menu or My Menu, go into Live View, and, using your remote, auto focus on the target, The press the button on your remote again to reach the pop-up screen that prompts you to save the AF fine tune value.

As the procedure stands now, after going into LV and focusing on the target, you have to basically grab the camera with two hands to push the AF mode and Movie Record button to get to the screen prompt. This physical action of "mauling " the camera seems to defeat the whole notion of putting the camera on a tripod, locked down, having a good target and carefully focusing  on a good target. The whole process should be able to be done without ever touching the camera after entering Live View (except to maybe zoom in on the target BEFORE triggering the remote to double check focus before another remote shutter press to get the screen prompt.)

For example, I had set up my Lens Align target (aligned parallel to the sensor plane) with ruler at about 40x focal length. My D500 and 300 f/2.8 was set up on a Series 4 Gitzo tripod and Wimberley head - all locked down and with a remote shutter release. You all know (or should know) the drill for this kind of testing. After taking about a dozen shots, it was clear, in examining the images of the Lens ALign target/ruler (but not using the Focus Tune Software) that my 300 f/2.8 (at f/2.8 ) AF fine tune should be +1. Test results were consistent.

But then, just for kicks I tried a few shots using the same Lens Align target properly set up, but using the AFFT procedure - following it by the book as has been well documented by some of you and on other websites. The AFFT results were all over the map - +6, +8,-15, -10 and so forth. I gave up instead of diving into the nebulous rabbit hole of AFFT abyss.

Disappointed - I admit that the first new feature that immediately caught my eye when the D500 was announced on January 5 was the LV Auto Focus Fine Tune Procedure as the whole AF fine tuning procedure consumes an enormous amount of time and can be frustrating. However, as the procedure is presently structured, it does not produce consistent, useable results. I'll stick to using Lens Align target/ruler and verifying the results by shooting real objects at reasonable distances.

All I am saying is that, FOR ME, the AFFT procedure is a waste of time that produces conflicting test results. But Nikon could easily make major improvements to this interesting concept.

[chris dees]

If you're not happy with the new procedure, just use the old one.

[Peter Connan]

Perhaps my expectations are skewed by the fact that my gimbal head is a cheap chinese rather than a real Wimberley-McCoy, but i would never attempt such a procedure on a gimbal head...

Taking this further, i would wonder whether Nikon's notoriously floppy lens feet are sturdy enough for this operation in the first place.

So on the one hand, it does seem a pity that Nikon didn't make it possible to perform the setting without physical contact, either by cable or wifi, or simply with a time delay, but on the other hand why not just put the camera on multiple sandbags?

[gryphon1911]

Once the AF is locked, it probably won't matter if you bump the camera little.  Locked is locked.  I think you are overthinking things a little.

[afx]

But then, just for kicks I tried a few shots using the same Lens Align target properly set up, but using the AFFT procedure - following it by the book as has been well documented by some of you and on other websites. The AFFT results were all over the map - +6, +8,-15, -10 and so forth. I gave up instead of diving into the nebulous rabbit hole of AFFT abyss.

This echos what the author of FoCal recently wrote.
http://www.reikan.co.uk/focalweb/index.php/2016/04/nikon-d500-automatic-af-fine-tune/
While his post could be seen as pushing his product, your results support his findings.

cheers
afx

[James Farrell]

This echos what the author of FoCal recently wrote.
http://www.reikan.co.uk/focalweb/index.php/2016/04/nikon-d500-automatic-af-fine-tune/
While his post could be seen as pushing his product, your results support his findings.

I don't know anything about FoCal. So, I cannot comment on anything about what it does or how. What I am saying is that the Nikon D5/D500 Live View Auto Fine Tune process yields wild, unpredictable results. The procedure as presently structured is flawed. It requires you, the user to push the OK button on the camera (inducing some shake and potential misalignment from your focus on your target.) The actual measurement of the focus accuracy is done at the exact instant when you press the OK button on the Nikon D5/D500 body after the prompt you see when had pushed and held the AF Mode button and the Movie Record button for a couple of seconds. So you introduce some movement to the body, even being careful there is some movement, at the exact instant that the body should not be moved as it reads focus in Live View to determine an AF tune value. That makes ZERO sense.

What Nikon should have done (or maybe could do via firmware update) is to make the whole process one that can be done using a remote shutter release cable. I outlined that in my original post. One knowledgable poster on DPreview had a terrific suggestion which states that if the process could be done by remote, and when one presses, say, a shutter release cable, the camera would take a series of focus readings and create an average - thus yielding a potentially more meaningful AF fine number adjustment.

As it stands now, the most accurate method FOR ME is to use a Lens Align target and slanted ruler - shooting a series of shots at different AF fine tune settings and to look at the images at 100% to arrive at what I believe is the best settings. Then take some shots at a distance and evaluate them. For me this works well. Yes, I'd love to use (i.e., depend upon) the Live View Fine Tune process, but it simply does not produce consistent, meaningful results. It's almost laughable to read some of the posts on DPreview where folks are scratching their heads about why, for example, their Nikon 200-500 lens at -15 or something like that, just doesn't seem to focus well, having used the AFFT process.   

[Ilkka Nissilä]

I don't have a camera with auto AF fine tune yet but a few things come to mind.

Try a longer distance so that there is more depth of field. This way touching the camera would likely have less effect.

Ideally the tripod should be such that when you touch the camera lightly, nothing happens. I realize the tripod collar on that lens may make it unrealistic. But for shorter lenses and sufficiently long distances the effect of touching should be reduced.

Another thing is to make sure there is enough light. In dim light the AF sensor gives more noisy data.