What distance to be at for AFMA

[-Gamer-]

Does anybody know at what distance or approximately the distance I should be at to AFMA, I'll be using the Canon 400 f/5.6L with the 7D Mark ll?

Overall I'm happy with the photos i'm getting and love the lens/camera combination but there are times when i'll get say, a run of a pelican and 1 or 2 out of 5-7 shots will be tack sharp the other's will be soft. It's not really a big deal for me as I get more keepers then throw aways most of the time but figured I'd give AFMA a try and see how much if any adjustment I may need.

 

[Click]

This might help you

http://www.the-digital-picture.com/Photography-Tips/af-microadjustment-tips.aspx

 

[Marsu42]

Does anybody know at what distance or approximately the distance I should be at to AFMA

You need to afma at the distance you shoot at most often. Unfortunately, Canon lenses don't have distance-dependent afma values (afaik newer Sigmas have), so you have to choose.

If you use the lens at different distances and afma values differ a lot, you need to change them in camera each time - this is the case for my 100L at macro and portrait distances. I'm probably going to program Magic Lantern to swap afma values on a key combination if I come around to hacking ML again.

 

[Loswr]

I test at 25x and 50x the focal length, so that's 10 and 20 meters.

 

[Marsu42]

It's noticeably sharper than without correction.

You probably meant to say "The focus hits more often than w/o correction" or "The zone of the shots I intended to be sharp are noticeably sharper than without correction" :->

Sorry, but I have to point this out, but afma often seems to be confused with a magic bullet that solves every possible blur problem.

I did my 7D2/4005.6 at about 60ft. [..] My value of +4 seems to hold true throughout my working range.

And sorry again, but I'm confused: From what I read about these afma values, a value of 4 for a f5.6 lens shouldn't be able to make such a significant difference. Or am I mistaken?

 

[Marsu42]

At my working distance the 400 5.6 has a very narrow plane of focus. When it's off, you can see it.

Ok, I believe you if you checked in detail - and in this case, I should probably re-check my 70-300L (f4-5.6) because I only did a rough afma check. And most likely, with my handheld wildlife photography at medium iso values (i.e. accepting residual motion, IS, and hss blur) it doesn't make much of a difference.

I believe that I said that I did not expect it to solve his problems with IF/OOF/IF issues so I don't think I claimed any magical properties.

And I didn't say that you claimed ... it was just an observation after reading so many threads about blurry pictures. Sooner or later, someone claims "It's afma, buy FoCal" even if it's so easy to check if your shot is sharp *somewhere*

 

[Loswr]

I believe that I said that I did not expect it to solve his problems with IF/OOF/IF issues so I don't think I claimed any magical properties.

And I didn't say that you claimed ... it was just an observation after reading so many threads about blurry pictures. Sooner or later, someone claims "It's afma, buy FoCal" even if it's so easy to check if your shot is sharp *somewhere*

The best advice is take a set of live view shots, then a set of phase (viewfinder AF) shots, pick the sharpest of each and if LV is sharper, AFMA can help. If that's not the case, look at other solutions (more careful lens testing, often ending up by sending the lens for service).

BTW, I didn't say you weren't aware of that, just an observation that it's good advice.

 

[Marsu42]

The best advice is take a set of live view shots, then a set of phase (viewfinder AF) shots, pick the sharpest of each and if LV is sharper, AFMA can help.

Nononono, if *I* would write that then you'd claim that lv af is also prone to error, and such a simple check cannot replace a real afma software :-> ... but I'll try with the manual method and ML dot_tune again because I did it at far too close distance last time.

So, if you don't get a nice crisp shiny image of the eye you might as well toss them in the bin.

That's exactly what I'm trying to get away from - less pixel-peeping, more story telling and emotion. It's tough with wildlife, but rewarding if you get shots that have an impact beyond "Oh, look, a [insert animal species here]".

Yes, I agree with your overall point. Sorry if I seemed overly defensive but some of these threads do take any ugly turn at times.

At camera body release time, things seem indeed to get ugly - and I recently decided not to participate in "importance of dr" threads or Canon vs. Nikon anymore. I try to demonstrate with my avatar that I'm not around here to get into flamewars, but to learn and share my knowledge (or what I perceive as knowledge ).

 

[Loswr]

Nononono, if *I* would write that then you'd claim that lv af is also prone to error, and such a simple check cannot replace a real afma software :->

If you took just one shot, yes. Taking several (5 or more with each method) and picking the sharpest ones should eliminate the effects of variability (which affects both AF methods).

 

[dcm]

FoCal Test Distance Chart and Explanation here. Two things to note
[list type=decimal]
[*]its the minimum recommended distance, you can be farther. I tend to check lenses at two distances just to be sure - the minimum and 2X the minimum is my default.
[*]you do a teleconverter in the focal length, but you don't consider the crop factor in the focal length when deciding the distance.
[/list]

So 8m is the minimum recommended for your 400mm lens.

 

[jdramirez]

Near minimum focusing distance is where the dof is the thinnest... so I afma near mfd. I assume if I am shooting a target 10x mfd then my dof will be much more significant, so a focus shift (assume I'm using that term correctly) may occur, but it should still be within my dof.

 

[jdramirez]

I test at 25x and 50x the focal length, so that's 10 and 20 meters.

Does the AFMA change with distance? In other words, if you set the AFMA at distance X, will it still be correct at distance 2*X?
I ask because I do my AFMA with a DIY setup and I like to get closer to the target to take advantage of the thinner DOF. The thinner DOF makes it easier to see how much back or front focus there is.

per a dof calculator:

with full frame, 50mm, f/2.8, and 30 cm from subject...

Depth of field
Near limit 29.7 cm
Far limit 30.3 cm
Total 0.51 cm

In front of subject 0.25 cm (50%)
Behind subject 0.26 cm (50%)

but when you double the distance:

Depth of field
Near limit 58.9 cm
Far limit 61.1 cm
Total 2.24 cm

In front of subject 1.1 cm (49%)
Behind subject 1.14 cm (51%)

So double the distance and you wind up with a 4.39x the dof...

 

[Mt Spokane Photography]

It does depend a lot on the distance you normally use a lens at. However, most lenses tend to front focus as you get near mfd, so unless that's where you use it, be sure to check at multiple distances. My new 35mmL varied AFMA considerably with distance, requiring as much as +17 for distant objects.

I had to actually change it for close objects. I sent it to Canon under warranty, they adjusted it on their reference body, and from then on, it was great.

 

[Loswr]

Does the AFMA change with distance? In other words, if you set the AFMA at distance X, will it still be correct at distance 2*X?
I ask because I do my AFMA with a DIY setup and I like to get closer to the target to take advantage of the thinner DOF. The thinner DOF makes it easier to see how much back or front focus there is.

Yes, it changes with distance, and as Mt. Spokane states most of that changing goes on at and near the MFD. So if you're very close to the target when you test, your results may well be inappropriate for subjects at further distances.

 

[2n10]

It's noticeably sharper than without correction.

I did my 7D2/4005.6 at about 60ft. [..] My value of +4 seems to hold true throughout my working range.

And sorry again, but I'm confused: From what I read about these afma values, a value of 4 for a f5.6 lens shouldn't be able to make such a significant difference. Or am I mistaken?

I believe I read that each point is equal to approximately 1/8 of the depth of field at any given distance.

 

[Loswr]

I believe I read that each point is equal to approximately 1/8 of the depth of field at any given distance.

You may have read that (perhaps even in an article by Rudy Winston on Canon's own DLC site), but it's incorrect. Each unit is 1/8 the depth of focus for the lens wide open. Depth of focus is the sensor-side equivalent of depth of field, but the former is largely independent of subject distance, while subject distance can dramatically alter the latter.

For reasonably close subjects at 400mm f/5.6, I'd expect a 4-unit AFMA to make a visible difference (as it would for a shorter lens with an even closer subject...DoF boils down to just magnification and aperture).

 

[Loswr]

Near minimum focusing distance is where the dof is the thinnest... so I afma near mfd. I assume if I am shooting a target 10x mfd then my dof will be much more significant, so a focus shift (assume I'm using that term correctly) may occur, but it should still be within my dof.

Not a good idea, for the reasons stated above.

 

[jdramirez]

Near minimum focusing distance is where the dof is the thinnest... so I afma near mfd. I assume if I am shooting a target 10x mfd then my dof will be much more significant, so a focus shift (assume I'm using that term correctly) may occur, but it should still be within my dof.

Not a good idea, for the reasons stated above.

I'm shooting my little rugrat... so I don't have much predictive capabilities when it comes to consistent distance from the camera she will be.

 

[dcm]

Does the AFMA change with distance? In other words, if you set the AFMA at distance X, will it still be correct at distance 2*X?
I ask because I do my AFMA with a DIY setup and I like to get closer to the target to take advantage of the thinner DOF. The thinner DOF makes it easier to see how much back or front focus there is.

Yes, it changes with distance, and as Mt. Spokane states most of that changing goes on at and near the MFD. So if you're very close to the target when you test, your results may well be inappropriate for subjects at further distances.

The Focal Test Distance Explanation here has a nice graph and discussion on the second page to show why you may not want to do AFMA near MFD, unless you frequently shoot near that distance. The current Canon AFMA has a single value for all distances and possibly two values for zooms so you might want to do it at both MFD and the normal distance. Then you will have to remember to dial in the appropriate value when shooting.

 

[Marsu42]

So, if you don't get a nice crisp shiny image of the eye you might as well toss them in the bin.

That's exactly what I'm trying to get away from - less pixel-peeping, more story telling and emotion. It's tough with wildlife, but rewarding if you get shots that have an impact beyond "Oh, look, a [insert animal species here]".

Agree, but it's hard to achieve any of these things when your subject has a dead lifeless OOF eye that makes it look like it's been stuffed. It's not pixel peeping.

True. However, my subjects' eyes (mostly horses) are so hard to focus that afma has a minimal impact on "in focus or not". So for scenes I find unique and rewarding, I take a couple of shots anyway for safety and select the best one. Last not least, phase af has a "jitter" so no matter how good afma is, taking more than one shot is a good idea - but of course correct afma is to be preferred.

I imagine afma is esp. important for tracking because the camera keeps adding or subtracting the value all the time, and has more points in time to improve on its decision. But with my 6d and tracking, afma is the least of my worries