Nikon AF Fine Tune: what am I doing wrong?

[michael_tennant]

<p>After spending most of the morning fiddling with my D4S, LensAlign and two Nikon lenses, I thought I was getting consistent results. Specifying an AF Fine Tune of +3 for the AF-S 50 1.4G produced spot on focus on the 0 mark on the LensAlign ruler. The AF-S 24-70 f/2.8 zoom was tougher to calibrate, but specifying an AF Fine Tune of +10 put me spot on at 60mm and slightly front focused at 28mm.</p>

<p>I took many shots with these +3 and +10 figures, saw consistent results, and figured I had successfully configured AF Fine Tune. I did my experimenting during the day in diffuse natural light, at F8 (as Michael Tapes suggested for LensAlign, or at least my version of LensAlign). I set up the LensAlign carefully, to center the red LensAlign 'bullseye,' defocused the lenses between shots and made sure to take test shots with the LensAlign sight gate closed. </p>

<p>This evening, I slipped on an SB-900 flash and took some casual snaps in my living room of a telescope with a long line of embossed letters on the telescope shaft. I'd just calibrated my lenses with AF Fine Tune, and figured I'd nail focus at the aimed-for letter with both 1.4 on the AF-S 50 and 2.8 on the 24-70. This time, though, I was hand holding the camera at 1/60 or 1/80, shooting with iTTL flash and shooting with the lenses wide open at 1.4 and 2.8, respectively, rather than F8. </p>

<p>Well, I <strong>didn't</strong> nail focus. The 24-70 came close, but the AF-S 50 1.4G didn't consistently focus on the money until I changed the +3 (for mild front focus) to -20 (!!) for severe back focus. </p>

<p>I'm nearly new to AF Fine Tune. (As you can probably tell.) I could load some shots into Photoshop, draw in little arrows to show what I was trying to focus on and so forth, but would rather just hear some off-the-cuff thoughts from those more experienced than I am with Fine Tune. Does this post make you think "equipment problem," or have you seen similar variations in different lighting situations, particularly when shooting wide open with a lens as fast as the 1.4? </p>

<p>(Or, maybe I'm trusting those Nikon focus points more than I should, if this vintage Luminous Landscape article still holds today: http://www.luminous-landscape.com/tutorials/nikon-focus.shtml )</p>

<p> </p>

[michael_tennant]

<p>I should have mentioned this article, too:</p>

<p>http://www.dslrbodies.com/newsviews/nikon-focus-followup.html</p>

<p> </p>

[CvhKaar]

<p>This evening, I slipped on an SB-900 flash and took some casual snaps in my living room of a telescope with a long line of embossed letters on the telescope shaft</p>

<p>Can you post the result, so we can understand how you took this shot ?</p>

<p>An aperture is in a lot of cases not realy favourable unless you realy need a shallow DoF. so maybe this shot would work better at a smaller aperture ?</p>

[ilkka_nissila]

<p>Focus fine tuning is a bit tricky but once you've figured out the right settings for your camera and lens, it does result in improved results in many cases. What distance is your target when you're doing your fine tuning measurements, and how is it lit? I recommend having a target at 50x the focal length of the lens. This is really important that you don't use a too short a distance for the fine tune testing, as this can result in a value that doesn't work well over long distances if the original test distance is too close.The target needs to be big enough to cover the whole sensitive area of the active focus point and preferably have some text; I've found black text on white cardboard to work really well for the AF.</p>

<p>I've moved from capturing a series of images to using the focus confirmation dot to do initial fine tuning without having to capture a single shot. First I focus the lens on the target using live view. If CDAF in live view seems to work correctly then I use that as a reference and turn off live view, then I go through a range of fine tune settings, pressing the shutter button half way to see the focus confirmation lights (I have the camera set to autofocus only on pressing AF-ON, but if you have AF activated from the shutter button, then you need to switch the lens or camera to manual focus) and see where the focus confirmation dot is solid, and where it blinks with arrow either indicating front or backfocus. I then take the average of these limits where it turns into a solid dot and set that as my fine tune value. Then I perform real world testing. I think the test with the ruler and repeated AF operations is perhaps more precise than the method I use, but it takes more time and requires firing the shutter and analyzing the data on the computer. In the end the real world conditions may be different from the focus fine tuning conditions so I have to finalize the value I use based on real world shooting anyway. I find that with some lenses my focus confirmation dot technique yields a value that is perfect for real world use, and with some lenses I need to adjust it a bit for real world use. For zooms, the 70-200/2.8 II focuses with all the cameras I've used with minimal focus fine tune and appears consistent across focal lengths. The 24-70/2.8 by contrast shows a large difference in optimal focus fine tune setting between the two ends of the zoom, so in practice I have to measure both ends of the range and then use an intermediate value. BTW with some lenses the contrast-detect AF in live view doesn't yield a good focus for use as a reference, this is especially the case with some screwdriver AF lenses e.g. the DC Nikkors. In this case I focus the lens manually using live view and then proceed to test the AF function by observing the focus confirmation dot and arrows when the fine tune setting is varied. If in doubt, repeat the procedure and get more data and figure the limits of where the results vary. Due to the finite precision and reproducibility of focus using autofocus lenses (that are not based on stepper motors), there can be some variability in test results. </p>

<p>I recommend testing in bright enough light where possible. BTW I believe Nikon recommends testing autofocus with the lens sligthly stopped down if it is a fast prime. With some lenses the focus can be shifted when the lens is stopped down, so it may be best to use an aperture which is typical for your use. I test wide open because that's where the focus is most critical and where I usually use my fast primes. But if you stop them down typically e.g. to f/2.8 then it is perhaps best to test at that aperture, or test at both f/2.8 and wide open and see if there is a difference in the result. </p>

[arthur_smith1]

<p>Really a function designed for prime lenses, not zooms. </p>

[john_curlett]

<p>"Really a function designed for prime lenses, not zooms."<br>

A more correct statement would be that it is designed for all lenses but often works better on prime lenses. Errors in phase detect focus can vary with focal length, focus distance and aperture. The results can be better with primes because one of the variables have been removed. Canon allows micro-focus values to be saved at wide and telephoto settings of zoom lenses which should significantly improve the effectiveness with zoom lenses. </p>

[al_derickson]

<p>My AFS 50mm f1.4 back focuses wide open at close distances but is right on at infinity. I used AF fine tune on my D7100 to correct it because I know the chances are my subject will be close if I'm using it wide open. I'll be stopped down at far distances and DOF will cover any focusing error.</p>

<p>Al</p>

[rodeo_joe1]

<p>Almost every f/1.4 lens ever made exhibits some focus shift on stopping down. Residual spherical aberration is the culprit, and even Aspherical designs suffer from it.</p>

<p>You can see this in action in Photozone's review of the 50mm f/1.4 G Nikkor here: http://www.photozone.de/Reviews/413-nikkor_50_14g?start=1<br>

Scroll down to the "Bokeh fringing/LoCA" section and mouseover the various apertures. The focus can easily be seen to creep backwards as the lens is stopped down, and is one of the worst cases I can recollect seeing on Photozone's site.</p>

<p>Sorry, but unless you set a different fine tune for nearly every aperture, this is an effect you'll have to live with.</p>

[michael_tennant]

<p>Boy, those are some great responses! Gentleman, thank you very, very much for contributing. Ilkka, I read another one of your posts about AF Fine Tune, and had hoped you'd have thoughts about this. (I did my testing at about four and a half feet. I should have mentioned that in the first post.)</p>

<p>Rodeo Joe, I have a feeling you nailed the issue right on the head. The example cited at the bottom of the linked page is convincing. (Convincing and a little amazing; I'd never guessed stopping down could have that much of an effect.)</p>

<p>Looks like I'm in for another session with LensAlign.</p>

<h1> </h1>