Focus Fine Tune

Started May 13, 2014 | Discussions thread
stevo23 Forum Pro • Posts: 17,640
Re: Focus Fine Tune

mosswings wrote:

richiebee wrote:

Interesting. The implication here and the suggestion in others posts is that this is somehow semi-destructive. On my camera, all lenses start with a value of zero, and you can offset each lens (which gets recognized by the camera when its attached) using a single parameter slider. Resetting the value to zero for each lens (or I suspect resetting the camera itself will put all lenses back to default) is all that is needed to return it to its original position so I don't get the need to be so careful here. You're not taking a screwdriver to anything.

Is it done differently on other cameras?


ajscullard wrote:

As darklamp says, unless you really know what you are doing and you have sufficient experience to analyse your results well enough to ascertain correctly why a shot is not sharp enough, then you need to think carefully about dong any MA. You may end up making things worse so be sure you have some way of resetting the lens(es) to the original settings.

So here's my take on the subject. Darklamp is correct that properly setting the AFFT coefficents requires patience, thoroughness, and the right setup. It doesn't help that there are sites on the web that purport to help but use wrongly designed targets (the infamous 45 degree piece of paper) that can induce further errors. It also is a process that requires really a statistical approach to get absolutely right, because there's an inherent spread to the results that is a function of lighting level, type of target, and where the lens was focused previously. You need to test multiple times at each AFFT coefficient value and at each FL you plan to use. So we're talking 50-100 measurements per FL. Oh, and it needs to be on tripod, with careful attention to the orthonormality of the target and optical axis, and sources of vibration need to be carefully removed; i.e., make sure that there's a goodly delay between mirror-up and shutter release. Then you get to process the results and see where the statistics take you.

AFFT is not useless for zooms, but if you only have a single AFFT coefficient per lens, you're definitely going to have to compromise or optimize at a particular FL. Some manufacturers provide 2 or more AFFT coefficents and an interpolation feature, allowing you to at least approximate the full focal-plane vs. focal-length curve. But most don't. AFFT works most completely with primes, but then only if they're at the correct subject-camera distance. AF error is not constant with subject-camera distance but asymptotes as you increase distance.

That being said, AFFT is commonplace on Pro bodies because pros demand the best focusing performance and know how to calibrate it. It is becoming more common on lower end bodies both as a marketing feature differentiation but also because the high resolution of current bodies will reveal focusing errors more readily, and because manufacturing QC standards are permitting more variability in lens and body performance. It is true that slower lenses with their greater minimum depth of field are more tolerant of focusing errors than fast lenses, so as long as you're using an f4-f5.6 kit lens not having AFFT is a reasonable thing; but when you put that cheapo (and good) 35mm f1.8 on it, you might start seeing things.

Where I have found AFFT useful on my D7100 is in using telezooms. The 70-300 Nikon is not as sharp as I might like it at 300mm, but that's where one tends to use telezooms, and mine, for example had a +9 correction at 300mm but a +3 correction at 70mm. In this case, I'd use +9, and I will see differences in sharpness when I do so, even handheld, but again with as many sources of vibration as possible eliminated. My midrange zoom, on the other hand, is more problematic. There, I have to use an average value.

How long did it take me to set my lenses up? About a day (a bright, sunny day), and FoCal, and 300+ measurements. At least I won't have to send my lens and body back to Nikon Service for basic lens-body matching (which they actually don't do, they just calibrate the lens and body separately to the centers of their individual specifications to allow for the maximum number of lens-body combinations). But if there's something seriously wrong with the body AF or the lens calibration, the factory calibration uses far more parameters than are available to the AFFT user - far more even than Sigma does with its USB dock and lens tuning program.

I don't find it to be all that difficult or complex as you describe. I've managed to find critical focus fine tuning for a given lens in an hour or so. It's a lot of shooting and re-shooting to get it right, but if it takes a whole day, I think something else is wrong. Seriously, it's not and can't be a perfect science because lenses don't behave linearly.

I also think flat targets are useless. 3 dimensional subjects are best as that's what we shoot. Trying to determine focus from a flat, black and white target won't tell you what the lens will do in real life. You need all the colors and all three dimensions as that is what your lens will be seeing when you shoot a photo.

 stevo23's gear list:stevo23's gear list
Fujifilm X-Pro2 Fujifilm XF 60mm F2.4 R Macro Fujifilm XF 14mm F2.8 R Fujifilm XF 27mm F2.8 Fujifilm XF 23mm F1.4 R +3 more
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow