Question about AFMA on 7D/5D/etc.

[mmodica]

So I have a simple question about autofocus microadjustment on the bodies that have it.... the values that you dial in to calibrate lenses, what do those actually refer to? They aren't just arbitrary numbers, are they? I thought I read somewhere that they were percents of focal length, but that doesn't seem to make much sense...

[Joel Eade]

The AF system looks mostly at contrast and moves then lens to the sharpest point and then stops. When you perform AFMA you are telling it to stop either in front of or behind where it normally would to achieve focus. The numbers are related to depth of field. Each number being equal to 1/8 the depth of field that lens would have when shot wide open.

Check this out: http://www.learn.usa.canon.com/resources/articles/2011/af_microadjustment_article.shtml

[neuroanatomist]

Minor correction - not depth of field, but rather each AFMA unit is 1/8 the depth of focus, which is the image plane counterpart to depth of field (measured in microns at the AF sensor/image sensor). At normal (non macro) subject distances, changes in focal length and subject have significant effects on depth of field, but minimal effects on depth of focus.

[mmodica]

Minor correction - not depth of field, but rather each AFMA unit is 1/8 the depth of focus, which is the image plane counterpart to depth of field (measured in microns at the AF sensor/image sensor). At normal (non macro) subject distances, changes in focal length and subject have significant effects on depth of field, but minimal effects on depth of focus.

That's the answer I was looking for, thanks! Although one other quick tidbit: since depth of focus does change with aperture, I'm assuming it always measures from full aperture? That would make the most sense I think...

Oh, and greetings from another Bostonion :D

[neuroanatomist]

Correct - autofocus (and AFMA) is always done with the lens wide open. It gets stopped down to the selected aperture just before the shutter is opened.

[Dr Croubie]

Correct - autofocus (and AFMA) is always done with the lens wide open. It gets stopped down to the selected aperture just before the shutter is opened.

Which is why AFMA will not help you correct for focus shift at different apertures, as some certain pesky lenses seem to have (even expensive ones).

Still, we've gone from 1 AFMA value for a whole zoom lens to 2 values, long and wide end, and from per lens-type to per-serial-number.
In the future I don't see it impossible that you could dial in different values for different apertures too, it's just a matter of canon getting off their butts and coding it in (and maybe a bit more nvram for a huge lookup table), and a lot more work for us testing at every possible combination...

[Sean Setters]

I was really impressed by the Sigma Optimization Pro Software that allows you to dial in specific AFMA values based on different subject distance ranges. I have the Sigma 35mm f/1.4 - and even though I'm very happy with the lens as it is, my usb dock is pre-ordered.


http://vimeo.com/64665246

[mmodica]

Do you think there is any chance of Canon adding a similar function to EOS utility if they lose enough sales, and face, to a third-party vendor with a superior adjustment system?

Edit: I suppose I should mention that I've already purchased two Sigma lenses and see more of their lenses in my future.

If they do, I would hope it would have different adjustment values for different focal lengths, different focusing distances, and different apertures - since I know the focus shift present due to spherical abberation will differ for all three of those :D

[Busted Knuckles]

So working the AFMA on my lenses w/ a tripod this time...........

Tethered shooting, decent targets that allow me to measure the DOF from the precise focus point.

I thought I had read somewhere that the DOF should be 1/3 in front and 2/3 behind - or was that some sleep deprived hallucination.

I think the logic of this was that most times we are looking at the front of the object, say focusing on the eye, front 1/3 brings the nose in to focus and back 2/3 brings most/all of the head/neck into focus.

When shooting sports, the guy w/ the ball and the folks chasing him.... he hopes they are behind him

When shooting wild life - it is the eye again and most of the animal is behind the eye most of the time (adj for going away shots note to self)

Thoughts or is it to be 50/50?

Thanks

[HDNitehawk]

I had heard the 1/3 - 2/3 rule before.
I go for 50/50 if I am doing it manually.
I think FoCal is probably 50/50 because it looks for the sharpest focus.
I prefer mine to favor back focus for the eye reason you give.

Honestly though, most camera models are not so accurate that you will pin point the 50/50. They are not that precise.
I have tracked this in the past when using lens align. Out of 10 shots you might get 4 that favor front focus, 4 that favor back focus and 3 dead on.
So if you're settings are front focused or back focused you will still get a fair amount of shots that hit the critical focus of the eye even if you are wrong.
You might just have a few less that the DOF captures a birds tail or the back of a models hair. If you set your camera to back focus you might occasionally fall in a zone that the critical focus on the eye misses slightly but is still in focus.

My 1D IV is the most precise body I have. The 5D II is a bit less. The 7D II, so far it isn't as nearly as precise as the other two which makes it a bit harder to take advantage of the increased resolution.

[neuroanatomist]

RE the 1/3 in front 2/3 in back, that's the way DoF is distributed around the focal plane *for wide angle lenses* (ultrawides can be 30:70 or more). As the FL increases, the distribution trends to 50:50...by 100mm, it's 49:51.

[Busted Knuckles]

Thanks,

As best I can tell I have it consistent on 1/3-2/3 on everything up to the 70-200, the tamzooka I have reset to 50/50 - before it was simply way back focusing

Going to pic some birds today w/the daughter and see how things go.