Question about micro focus adjustments (5D mkIII)

[YellowJersey]

I've been busy micro focus adjusting my lenses (all two of them) and was using this technique to get a baseline to work from http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html  (the first one on the page. The difference on the focusing ring between the live view manual focus ideal and the auto focus ideal was very tiny, it moved about 1mm. Then I tried taking various sets of test shots, experimenting with micro focus adjusting in camera by -1, 0, and +1. I see no appreciable difference between the images. I did see an appreciable difference if I adjusted to -2 or +2; the images looked slightly out of focus.

 My question is: is adjusting by -1 or +1 enough to see an appreciable difference? If they all look the same, should I just stick with 0 (no adjustment)?

[Mt Spokane Photography]

I use software called FoCal to do micro adjust.  It is not ready for the 5D MK III yet, the new Canon SDK is needed first.
 
However, the bell curve in the image below illustrates that a small AFMA error does not have much effect on sharpness.
 
You can see the peak for this lens is pretty broad. The peak indicates the sharpest focus.  That green band on the upper right shows where focus is reasonably sharp, and covers a width of about +/- 2 points  Some have a much broader peak, and a few lenses have a narrow peak.
 

[YellowJersey]

Thanks for the info. I might get FoCal once the 5D mkIII update is available. 

 I guess I'll just keep my 24-105 at telephoto at 0 for now. I'll continue tinkering though with the 24mm end... and then move on to the 17-40.

[neuroanatomist]

With an f/5.6 or f/4 lens, a 1-unit difference is likely not detectable, but with an f/2.8 and faster lens, yes...but I think NOT with the moire pattern.  I find that method to be less precise than a LensAlign (which I used to use) or SpyderLensCal.  Actually, I'm currently using the excellent Reikan FoCal software.

[Mt Spokane Photography]

With an f/5.6 or f/4 lens, a 1-unit difference is likely not detectable, but with an f/2.8 and faster lens, yes...but I think NOT with the moire pattern.  I find that method to be less precise than a LensAlign (which I used to use) or SpyderLensCal.  Actually, I'm currently using the excellent Reikan FoCal software.

I exchanged a few emails with Reikan over the last couple of days, I did not ask what approach they might take to the two AFMA settings for zoom lenses, I guess its obvious though.
 
They just received their 5D MK III and are trying to see if they can trick the current SDK into working with it, so we might see a beta if they can.  I suggested that he take a look at EOS count and see if shutter actuation counts could be incorporated as well.  If they can find a way that works, they will. 
 
Since the 1D X will display shutter actuations, there is hope that the SDK might have a more straightforward way of grabbing shutter counts.

[YellowJersey]

With an f/5.6 or f/4 lens, a 1-unit difference is likely not detectable, but with an f/2.8 and faster lens, yes...but I think NOT with the moire pattern.  I find that method to be less precise than a LensAlign (which I used to use) or SpyderLensCal.  Actually, I'm currently using the excellent Reikan FoCal software.

I exchanged a few emails with Reikan over the last couple of days, I did not ask what approach they might take to the two AFMA settings for zoom lenses, I guess its obvious though.
 
They just received their 5D MK III and are trying to see if they can trick the current SDK into working with it, so we might see a beta if they can.  I suggested that he take a look at EOS count and see if shutter actuation counts could be incorporated as well.  If they can find a way that works, they will. 
 
Since the 1D X will display shutter actuations, there is hope that the SDK might have a more straightforward way of grabbing shutter counts.

Any preference between FoCal and doing it manually using something like the spyder calibration device? I'm actually pretty keen to try FoCal out, although I'll hold off until it has 5D mkIII support.

[fotoray]

With an f/5.6 or f/4 lens, a 1-unit difference is likely not detectable, but with an f/2.8 and faster lens, yes...but I think NOT with the moire pattern.  I find that method to be less precise than a LensAlign (which I used to use) or SpyderLensCal.  Actually, I'm currently using the excellent Reikan FoCal software.

I use the LensAlign MkII with good results.  I find it a tedious process, but you (hopefully) only have to do it once.

[neuroanatomist]

Any preference between FoCal and doing it manually using something like the spyder calibration device? I'm actually pretty keen to try FoCal out, although I'll hold off until it has 5D mkIII support.

Before FoCal came out, I used a LensAlign Pro (and I will be again, because I'll have my 1D X before FoCal supports it).  The values from the manual method agree well with FoCal, but as stated above, using the LensAlign is tedious, especially when done right (many shots at each adjust setting).

[Mt Spokane Photography]

Any preference between FoCal and doing it manually using something like the spyder calibration device? I'm actually pretty keen to try FoCal out, although I'll hold off until it has 5D mkIII support.

Before FoCal came out, I used a LensAlign Pro (and I will be again, because I'll have my 1D X before FoCal supports it).  The values from the manual method agree well with FoCal, but as stated above, using the LensAlign is tedious, especially when done right (many shots at each adjust setting).

+1
 
I did not take enough shots, and my results greatly improved the settings, but were still off.  I'm not looking forward to doing my 14 lenses plus 2 TC's manually for my 5D MK III.
 
The Canon SDK, when issued will presumably cover the 1D X as well as the 5D MK III, so it might happen more quickly for 1D X owners.

[AKCalixto]

Here there is a good tutorial about FoCal.  (I prefer to listen it)

http://blog.martinbaileyphotography.com/2012/02/06/podcast-321-lens-calibration-and-microadjustment-with-focal/

Discount ends March 31st.

http://www.fo-cal.co.uk/mbp45

[AKCalixto]

Reikan Focal Posts Update on Canon EOS 5D Mark III

http://www.the-digital-picture.com/News/News-Post.aspx?News=2192

[Mt Spokane Photography]

Reikan Focal Posts Update on Canon EOS 5D Mark III

http://www.the-digital-picture.com/News/News-Post.aspx?News=2192

 
 
Here is the quote
 
"The AF consistency on the 5D Mark III is the *best* I have seen on any camera I've tested with (which is quite a few now!). And it's better by quite a long way - only a few % deviation across 20 shots compared to, say, the 1Dmk4 which deviates by 10-15%."
 
I'm still giving AI Servo a good workout, but the 5D III's One Shot AF has been very impressive."
 
 
I hope that the AF repeatability is better, it is a real eye opener to see the variation in AF performance on my 7D and 5D MK II.  Just a few lenses were able to produce nice smooth curves with very repeatable results.

[YellowJersey]

Reikan Focal Posts Update on Canon EOS 5D Mark III

http://www.the-digital-picture.com/News/News-Post.aspx?News=2192

 
 
Here is the quote
 
"The AF consistency on the 5D Mark III is the *best* I have seen on any camera I've tested with (which is quite a few now!). And it's better by quite a long way - only a few % deviation across 20 shots compared to, say, the 1Dmk4 which deviates by 10-15%."
 
I'm still giving AI Servo a good workout, but the 5D III's One Shot AF has been very impressive."
 
 
I hope that the AF repeatability is better, it is a real eye opener to see the variation in AF performance on my 7D and 5D MK II.  Just a few lenses were able to produce nice smooth curves with very repeatable results.

Thanks for the update. I wasn't expecting them to be saying anything quite so soon. I'm very keen on using this and seeing what the difference (if any) is.

[Viggo]

How do you cope with VERY different AFMA from center to corner on a lens (in this case the 50 L.

I have a 50 L that I have adusted to be spot on, seriously good, using centerpoint, but when I use an outer point the lens focuses 50-100 yards in front, 100% repeatble.

[Mt Spokane Photography]

How do you cope with VERY different AFMA from center to corner on a lens (in this case the 50 L.

I have a 50 L that I have adusted to be spot on, seriously good, using centerpoint, but when I use an outer point the lens focuses 50-100 yards in front, 100% repeatble.

50-100 yards is a lot of front focus for a portrait lens that might normally be used at 10 ft, so you must be using it for landscape.  I think it might benefit from repair at Canon, but, as you have discovered, its not a landscape lens.
 
I use the center to set AFMA.  Lenses with a large field of curvature are not going to be in focus at both the center and edges unless you stop them down a lot.  Lenses also focus differently at different distances, and for zooms, they focus differently at different focal lengths as well.
 
As long as the photographer knows the limitations of the equipment, he can make adjustments.  You can manually focus to bring a item at the edge in focus and let the center be off, but most would not like the result.