AFMA Software and Suggestions

[pierceography]

Hello All,
I know this has been discussed in previous threads, however I'm starting a new one specifically related to someone who's never worked with AFMA and looking to get a little guidance.

My setup currently includes a 7D, 5Dm3, Sigma 12-24mm m2, Canon 24-70mm f/2.8L (I), Canon 50mm f/1.4, Canon 70-200mm f/2.8L IS II USM, Canon 100mm f/2.8L.  Particularly, I've noticed the 50mm and 100mm could use some adjustment.

So I have a few questions regarding AFMA:

1) I've read a bit about FoCal, but have no personal experience with it.  My budget for software (or anything else, really) is pretty tight right now.  The Pro version runs around $114 USD.  It sounds like most people in here use Pro.  So would Plus or even Standard work fine for just some basic AFMA?  I don't necessarily need the extraneous bells and whistles that come with Pro.  However, I also have a mac... and from what I've read, you need a Pro license to use the mac beta.

2) Are there any other methods/software recommended for AFMA?  I read neuro's guide, and while informative, I certainly don't have the gear to accomplish it.

3) And finally, is it a lengthy process?  I ask, because I'm heading to the Caribbean for a week this coming Friday, and would like to "fine tune" my camera(s) before heading down there, as it's something of a "once in a lifetime" type of trip for me.  I wanna make sure I can take full advantage of the scenes I'll likely encounter.  Probably should have investigated this more than a few days before I leave.  ;-)

Thanks in advance for the help!

[neuroanatomist]

Honestly, I think FoCal is the way to go.  Yes, you need the Pro version to run on a Mac (although if you have a Windows virtual machine - Parallels, VMware, etc., you can run a lower version that way).  The full auto takes about 10 minutes per run.

The way I see it, you've got somewhere north of $10K of gear, and are heading out on a "once in a lifetime" trip.  Spending ~1% of your gear value to get the best from your lenses seems like a pretty reasonable investment...

[pierceography]

I certainly agree about the cost ratio of current gear versus FoCal Pro license.  Only reason I bring up a budget is we have a kid on the way, and the wife has put our finances on lock down.  Time to start squirreling money away for my photography habit/obsession.  ;-)

I'll probably just wind up getting the Pro license.  The VM suggestion is a good one, but my mac isn't cutting edge, and pulling resources away from the native OS is annoying via VirtualBox.  Booting into another OS with the amount of terminals I actively maintain is always a lengthy process.

Thanks for the feedback, neuro!

[Dylan777]

Honestly, I think FoCal is the way to go.  Yes, you need the Pro version to run on a Mac (although if you have a Windows virtual machine - Parallels, VMware, etc., you can run a lower version that way).  The full auto takes about 10 minutes per run.

The way I see it, you've got somewhere north of $10K of gear, and are heading out on a "once in a lifetime" trip.  Spending ~1% of your gear value to get the best from your lenses seems like a pretty reasonable investment...

+1....on FoCal Pro. Even though this is my 1st software and 1st time using FoCal as tool for AFMA, I found it's easy to use.

[mirekti]

I belive I'll get FoCal as well.
Reading other articles about FoCal just brought some confusion to my head.

Do you think Plus version is good enough? I don't see benefits of Pro version that would make me buy it, but I might be wrong. Is there anythin that I would miss/regret for not buying the pro version?

The other confusing matter is the process itself. For example I'd likt to calibrate 70-200. As I understood one can do the calibration on different focal lenghts and many say they do it on the telephoto end.

What I noticed 5d III had AFMA with two parametrs wide and telephoto.
What would be a workflow for 70-200? Should I just run the test on 70 and put the result in W and than on 200 an put the result in T and that is it?

Than, there's this apperture test as well. What apperture should I use for the test, wide open?

[neuroanatomist]

Minimally, test at the wide end for W and the long end for T.  I test in between as well, and may alter the W/T choices slightly depending on that. I also test at two distances.

The aperture test is designed to tell you the aperture at which a lens is sharpest, but of course you'll want to shoot at the aperture you need.

[mirekti]

Thanks, but what would mean 70 +1, 135 +3, 200 +7

I can only put +1 and +7 but what should I do with 135? This is way to confusing. Theoretically, I can do a test for each and every focal lenght, print it out and change it before the shoot. 
OMG, this is way too complicated. I guess for prime lenses it is much easier choice.
Do you know how Canon calibrates lenses?

What are the two distances you chose. What I read recommended is 5x the focal lenght.

[Mt Spokane Photography]

Thanks, but what would mean 70 +1, 135 +3, 200 +7

I can only put +1 and +7 but what should I do with 135? This is way to confusing. Theoretically, I can do a test for each and every focal lenght, print it out and change it before the shoot. 
OMG, this is way too complicated. I guess for prime lenses it is much easier choice.
Do you know how Canon calibrates lenses?

What are the two distances you chose. What I read recommended is 5x the focal lenght.

If you have a zoom lens, AF accuracy is not the same at different focal lengths, so test the AFMA at the focal length you usually use and set it accordingly.
If you use a lens like the 24-105mmL at all focal lengths, at the minimum, set it to about 60mm to do your AFMA.  If you test it at 24, 60, and 105mm, you can find the best AFMA values for those focal lengths and see if its only a slight difference, or a major one.
For example, notice the two calibration curves for my 24-105mmL.  at 24mm, its a +1 adjustment while at 105mm, its a -2 adjustment.  Notice that the curve at 105mm is much flatter, so a small change in AFMA makes little difference.  With this lens I'd just use a setting of zero and it would be fine at all focal lengths.  But, since the 5D MK III has two AFMA settings, one for wide and the other for telephoto, I put in +1 for wide and -2 for telephoto, and it will interpolate at 60mm by using AFMA=0 or -1
With this lens, there was no real benefit from AFMA.
24-105mm L at 24mm

 
24-105mm L at 105mm

[Canon 14-24]

I have heard a lot about this FoCal software for AFMA.

My question is it really worth it?

Doesn't AF vary by distance to the subject too? From minimum focus of a focal length to a more distant shot at the same focal length?

So this software just adjusts the AF for a specific distance of a focal length (assuming you use that same distance quite a bit), if you change that distance from the subject...you aren't going to get that same perfect focus at the same focal length? Am I missing something?

I understand the software can do many other tests, but for the main purpose for AF calibration is FoCal worth it?

[Mt Spokane Photography]

I have heard a lot about this FoCal software for AFMA.

My question is it really worth it?

Doesn't AF vary by distance to the subject too? From minimum focus of a focal length to a more distant shot at the same focal length?

So this software just adjusts the AF for a specific distance of a focal length (assuming you use that same distance quite a bit), if you change that distance from the subject...you aren't going to get that same perfect focus at the same focal length? Am I missing something?

I understand the software can do many other tests, but for the main purpose for AF calibration is FoCal worth it?

It only measures one focal length at whatever distance you set the camera back from the target.  How could it do anything different?
 
However, you can vary the focal length and distance to the target and compare the best settings, picking the one that makes sense to you.
Knowledge is power, and by taking AFMA measurements at different focal lengths and distances, you can be the one who determines the setting. 
Of course, if You have a 5D MK III, or a 1D X, there are two settings for zooms - Wide and Tele, and the camera interpolates the AFMA for focal lengths inbetween.
As to distance, most bodies do tend to front focus at minimum autofocus distances, so you do need to test and see if its a real issue.

[Mt Spokane Photography]

I understand the software can do many other tests, but for the main purpose for AF calibration is FoCal worth it?

 
Thats a value judgement that only you can make.  For me, it was worth it.
 
Here, for example is the sharpness versus AFMA curve for my 50mm f/1.8 MK I lens.  The correct AFMA is -10, but truthfully, it looks pretty darn good at zero.  You do not notice the difference except at 1:1. Its the difference between about 600 and 750 on this scale.

[neuroanatomist]

Doesn't AF vary by distance to the subject too? From minimum focus of a focal length to a more distant shot at the same focal length?

So this software just adjusts the AF for a specific distance of a focal length (assuming you use that same distance quite a bit), if you change that distance from the subject...you aren't going to get that same perfect focus at the same focal length?

I understand the software can do many other tests, but for the main purpose for AF calibration is FoCal worth it?

Yes, yes, and yes. 

FWIW, I test at both 25x and 50x the focal length, and generally find very little difference (1-2 AFMA units). 

For me, it's absolutely worth it. I can get the same results with a manual tool, but FoCal is much easier to use.  I've got 15 AF lenses and lens+TC combos that AF (well, technically 17, but I don't use TCs with my 135L), and anything that makes the process easier is worth it.  It takes me about 10 minutes per test (one lens, one focal length, one distance) with FoCal, that's down from ~20 with manual testing (LensAlign Pro).  Since I test all lenses at two distances, and zooms at 3-5 focal lengths at both distances, 10 minutes less per test is literally saving me several hours.

[mirekti]

Since I test all lenses at two distances, and zooms at 3-5 focal lengths at both distances, 10 minutes less per test is literally saving me several hours.

Could you describe the workflow for 70-200 lens for example.
Would this be a good way?

1. Set the target at 10m (200mm x 50), maybe a feet or two more.
2. Perfrom test at 70mm than at 200mm.
I can see there are two results per messuring (50mm f/1.8 MK I) e.g. you have for 0 something below 560 and something above 640. Where those results from two messurments at different distances?
3. Simply put best wide end and telephoto end to 5d III

What do you do with the results from the focal lenghts in between?

[neuroanatomist]

Could you describe the workflow for 70-200 lens for example.

What do you do with the results from the focal lenghts in between?

Sure. I test the 70-200mm:

• set to 70mm at 1.75 m and 3.5 m from the target
• set to 135mm at 3.375 m and 6.75 m from the target
• set to 200 mm at 5 m and 10 m from the target

With a 1.4x TC or 2x TCmounted (the combo is treated as a separate lens), I'd set the same zoom settings, but test at 1.4x or 2x the above distances.

Selecting an AFMA value is a judgement and a compromise (across distance for primes, and distance + focal length for zooms), since you get one value for bodies other than the 1D X and 5DIII, and two values for those two bodies - Wide and Tele and they are applied with a linear regression between W and T.  IMO, more data points drive a better compromise.

For wider zoom ranges, I test more intermediate focal lengths (e.g., 24-105L at 24mm, 50mm, 85mm, and 105mm). For Macro lenses, I also test at close distance (30-50 cm).

In making a decision, I consider the focal lengths and distances I use most with that lens, and the DoF (thinner at the long end of zooms) in selecting the value(s) to apply.

[neuroanatomist]

Thanks, but what would mean 70 +1, 135 +3, 200 +7

I can only put +1 and +7 but what should I do with 135?

In that example, likely nothing - I'd probably use +1 and +7, knowing that at 135mm the camera would apply +4 which is close enough.

But for example, let's look at my 100-400 on my 1D X:

               50x   25x
100mm   0   -3
200mm   0    0   
300mm   0   -1   
400mm   1    2   

If I had just done the Canon recommended 50x and tested the ends, the values would have been W=0, T=+1.  If I had done just the ends at 25x, I'd have had W=-3, T=+2.  I went with W=-1, T=+1 for the adjustments.  Honestly, that's not too different than just testing the ends at 50x, but personally, I have more confidence in the decisions due to the extra data points.  In fact, being completely honest, with an f/4.5-5.6 lens like that, the one unit adjustments likely make almost no difference in real world shooting.  But that would not be the case with an f/2 or f/1.2 lens.

You also learn something about your lens.  All those values were pretty close, but what if your 100-400 is very different at 300mm (say, 50x +6 and 25x +7)?  If I saw that, I'd return the lens if new, or send to Canon if not new.  Canon actually changes the electronics package inside the lens to adjust the focus.

What are the two distances you chose. What I read recommended is 5x the focal lenght.

I use 25x the focal length and 50x the focal length.  LensAlign recommends 25x, Canon (Chuck Westfall) recommends 50x.  For some lenses (e.g. 85L, for me) I almost always use them at closer distances, so the 25x test gets more weight in my decision.