August 29, 2014, 04:32:29 AM

Author Topic: AFMA Callibration - How much is too much?  (Read 7375 times)

neuroanatomist

  • CR GEEK
  • ********
  • Posts: 13952
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #15 on: August 07, 2012, 02:53:58 PM »
FoCal does not yet support the 1D X (and only supports the 5DIII in a semi automated mode where you have to set the AFMA values on the camera yourself, when prompted by the software).  However, there is a fully manual mode (I think perhaps only on the Pro version), where are you capture the images yourself at different adjustment values, then load those images into the software, and it performs the analysis.  I've used that method with 10 different lenses on my 1D X.
EOS 1D X, EOS M, and lots of lenses
______________________________
Flickr | TDP Profile/Gear List

canon rumors FORUM

Re: AFMA Callibration - How much is too much?
« Reply #15 on: August 07, 2012, 02:53:58 PM »

JaxPhotoBuff

  • PowerShot G1 X II
  • ***
  • Posts: 68
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #16 on: August 08, 2012, 10:51:01 AM »
Like Dr. Deano, I would like to know:

1. the cost involved with having Canon calibrate lenses to a specific body; (is there a per lens price?)
2. the time it usually takes;
3. the process required to ship the equipment to Canon (where do I find the information about where to ship it and how?).

I am not a CPS member, so I have not gone through the process before. Any info would be appreciated.

I may not do this, if microadjustment ends up being satisfactory, but I would like to have the information just in case (and/or to know that it's as good as Canon can make it).

Thanks in advance.

canon816

  • EOS M2
  • ****
  • Posts: 211
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #17 on: August 08, 2012, 12:26:07 PM »
Like Dr. Deano, I would like to know:

1. the cost involved with having Canon calibrate lenses to a specific body; (is there a per lens price?)
2. the time it usually takes;
3. the process required to ship the equipment to Canon (where do I find the information about where to ship it and how?).

I am not a CPS member, so I have not gone through the process before. Any info would be appreciated.

I may not do this, if microadjustment ends up being satisfactory, but I would like to have the information just in case (and/or to know that it's as good as Canon can make it).

Thanks in advance.


The biggest problem with doing this is that if you use multiple bodies your lens could only be calibrated by canon to one of those bodies.  If you never intend to use more then one body then this is certainly an option.  However... not knowing the cost I would be willing to bet that sending in a whole kit worth of lenses would likely cost more then the Pro Version of Focal or some other method of calibration.

Also, I would hate to be without my gear while it is shipped to canon and you never know what kind of trauma it could experience in shipping or while there.  I have heard of many people who send in bodies to canon that end up coming home full of dust and debris (it actually happened to my father in law)  He sent in a body for a technical fix and ended up having to send it back two more times because it was full of "sawdust" like material. 

Just my .02

Jamesy

  • 5D Mark III
  • ******
  • Posts: 710
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #18 on: August 08, 2012, 12:36:00 PM »
I just bought FoCal Pro on the weekend and calibrated my lenses on my 5D3 - here are my results:

50 1.8 II: -6
70-200L F4 IS: W=-1, T=1
24-105L IS: W=0, T=-2
135L: -6
85 1.8: -4

It is funny, my 24-105 is not performing the way I would expect since I got it with my 5D3 in the kit. It does not seem as sharp as the 70-200. I will shoot an event this weekend and see how it does after the AFMA as prior to FoCal Pro, I never had used AFMA before.

I am sending in the 5D3 for the light leak issue, just so it is done and I also opened a case for Canon to look at the 24-105 while it is still under warranty. Perhaps I should borrow a 24-105 from one of my buddies to compare copies.

CanonCollector

  • Guest
Re: AFMA Callibration - How much is too much?
« Reply #19 on: August 08, 2012, 01:33:11 PM »
There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

Jamesy

  • 5D Mark III
  • ******
  • Posts: 710
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #20 on: August 08, 2012, 01:37:22 PM »
There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

Too early to tell for me, preliminary test say yes but I will have more to report after shooting an event this weekend.

canon816

  • EOS M2
  • ****
  • Posts: 211
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #21 on: August 08, 2012, 01:53:38 PM »
There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

Absolutely.  The increase in IQ by calibrating AFMA is staggering.  It is both quantifiable and easily noticed when shooting is the field.  If your AFMA is off by even a little bit your image IQ will drop dramatically.  I have attached a chart that plots IQ for one of my lenses through the AFMA range that was generated by foCal.  Before I calibrated this lens I was always complaining that my images looked soft at 100%.  As you can see the optimal AFMA point for this lens is +8.  I was shooting at the default of 0 and my IQ was suffering.  Now I consistently bring home tack sharp images at all shooting distances whether short or long.  I am more careful now on how many images I take in a shoot because it is daunting to have to go through hundreds of images that are all sharply in focus.  I used to be happy with the 10-15 "lucky" hots I brought home... now my keeper rate is much much higher.

This chart represents the IQ of the following test setup:
Canon 1DIV
600mm F4 NON IS
Distance to target 43 feet
12EV lighting
Fully Automatic Test

canon rumors FORUM

Re: AFMA Callibration - How much is too much?
« Reply #21 on: August 08, 2012, 01:53:38 PM »

Jamesy

  • 5D Mark III
  • ******
  • Posts: 710
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #22 on: August 08, 2012, 02:13:13 PM »
What is the name of the FoCal test you ran to generate that chart? I did a sharpness (aperture test) test on my lenses but it did not look like that.

bkorcel

  • Guest
Re: AFMA Callibration - How much is too much?
« Reply #23 on: August 08, 2012, 02:17:41 PM »
If you consistently shoot at F8 or so the AFMA adjustment may not be noticeable as most lenses will AF within the DOF range at smaller apertures.  AFMA adjustments shine when shooting wide open or close to wide open or when the lens or camera are terribly out of spec with each other.

Some lenses are soft anyway wide open so results may vary on consumer grade glass.  Focal allows high quality lenses to work as designed.  Even so, the affects of uncalibrated focus and diffraction will compound so getting the focus right in the first place may help even a consumer lens appear sharper.

Where I'm not 100% totally sold on it is that the software bases it's results on repeatability at different AFMA settings.  Honestly if your lens focus varies at one AFMA setting, it should vary just as much statistically at any other AFMA setting.  So say if AFMA 4 gave you 100% repeatability but AFMA 2 maybe 25% but 2 resulted in some shots with a higher IQ, AFMA 4 would be favored.

I am struggling with an instance with my 300mm where it recommends AFMA 4 but visually the image resolves the test target at AFMA 2 much better.  The nice thing is that the software charts the results as shown by another user here.  You can decide which AFMA value to use based on the results.

I'm leaning toward using the AFMA that produced the highest IQ on the test chart.  If my combo seems to have a higher repeatability with a soft image it's not really giving me anything.

There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

canon816

  • EOS M2
  • ****
  • Posts: 211
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #24 on: August 08, 2012, 02:19:41 PM »
What is the name of the FoCal test you ran to generate that chart? I did a sharpness (aperture test) test on my lenses but it did not look like that.

"Fully Automatic AF Microadjust"

When the test is complete make sure you save it.  This will create a PDF which you can view all of the data from the test as well as this visual chart of the data.

I use FoCal Pro.... not sure if that matters but not all versions show the detailed analysis.

Jamesy

  • 5D Mark III
  • ******
  • Posts: 710
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #25 on: August 08, 2012, 02:24:12 PM »
What is the name of the FoCal test you ran to generate that chart? I did a sharpness (aperture test) test on my lenses but it did not look like that.

"Fully Automatic AF Microadjust"

When the test is complete make sure you save it.  This will create a PDF which you can view all of the data from the test as well as this visual chart of the data.

I use FoCal Pro.... not sure if that matters but not all versions show the detailed analysis.

Thanks for that. I use FoCal Pro. It is on the second page of the detailed report

canon816

  • EOS M2
  • ****
  • Posts: 211
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #26 on: August 08, 2012, 02:27:52 PM »
If you consistently shoot at F8 or so the AFMA adjustment may not be noticeable as most lenses will AF within the DOF range at smaller apertures.  AFMA adjustments shine when shooting wide open or close to wide open or when the lens or camera are terribly out of spec with each other.

Some lenses are soft anyway wide open so results may vary on consumer grade glass.  Focal allows high quality lenses to work as designed.  Even so, the affects of uncalibrated focus and diffraction will compound so getting the focus right in the first place may help even a consumer lens appear sharper.

Where I'm not 100% totally sold on it is that the software bases it's results on repeatability at different AFMA settings.  Honestly if your lens focus varies at one AFMA setting, it should vary just as much statistically at any other AFMA setting.  So say if AFMA 4 gave you 100% repeatability but AFMA 2 maybe 25% but 2 resulted in some shots with a higher IQ, AFMA 4 would be favored.

I am struggling with an instance with my 300mm where it recommends AFMA 4 but visually the image resolves the test target at AFMA 2 much better.  The nice thing is that the software charts the results as shown by another user here.  You can decide which AFMA value to use based on the results.

I'm leaning toward using the AFMA that produced the highest IQ on the test chart.  If my combo seems to have a higher repeatability with a soft image it's not really giving me anything.

There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

You are absolutely right.  Shooting at F8 "hides" the fact that a lens might not have optimal AFMA settings.  Also you are correct that the repeatability of this test is more reliable with higher end lenses.  While wide open may not be the "sweet spot" for real life shooting you need to test wide open so that the images produce the narrowest DOF for the software to analyze appropriately.

It is worth mentioning that FoCal is not only a tool for calibration, but can expose other issues as well.  If you have a lens that will not produce a repeatable AFMA point or wide variability along the plotted chart line it can indicate that there is a problem with the lens (or body).  No amount of calibration will correct defaults that produce results like this.  A problem like this may need service from canon... but you might never know without running tests...

bkorcel

  • Guest
Re: AFMA Callibration - How much is too much?
« Reply #27 on: August 08, 2012, 02:47:33 PM »
Yep it does uncover repeatability issues but what I was saying is that it might not be a good idea to base your result on an AFMA that appears to be repeatable.  It may not be the AFMA that has the best IQ.  That is what I'm finding. 

If you consistently shoot at F8 or so the AFMA adjustment may not be noticeable as most lenses will AF within the DOF range at smaller apertures.  AFMA adjustments shine when shooting wide open or close to wide open or when the lens or camera are terribly out of spec with each other.

Some lenses are soft anyway wide open so results may vary on consumer grade glass.  Focal allows high quality lenses to work as designed.  Even so, the affects of uncalibrated focus and diffraction will compound so getting the focus right in the first place may help even a consumer lens appear sharper.

Where I'm not 100% totally sold on it is that the software bases it's results on repeatability at different AFMA settings.  Honestly if your lens focus varies at one AFMA setting, it should vary just as much statistically at any other AFMA setting.  So say if AFMA 4 gave you 100% repeatability but AFMA 2 maybe 25% but 2 resulted in some shots with a higher IQ, AFMA 4 would be favored.

I am struggling with an instance with my 300mm where it recommends AFMA 4 but visually the image resolves the test target at AFMA 2 much better.  The nice thing is that the software charts the results as shown by another user here.  You can decide which AFMA value to use based on the results.

I'm leaning toward using the AFMA that produced the highest IQ on the test chart.  If my combo seems to have a higher repeatability with a soft image it's not really giving me anything.

There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

You are absolutely right.  Shooting at F8 "hides" the fact that a lens might not have optimal AFMA settings.  Also you are correct that the repeatability of this test is more reliable with higher end lenses.  While wide open may not be the "sweet spot" for real life shooting you need to test wide open so that the images produce the narrowest DOF for the software to analyze appropriately.

It is worth mentioning that FoCal is not only a tool for calibration, but can expose other issues as well.  If you have a lens that will not produce a repeatable AFMA point or wide variability along the plotted chart line it can indicate that there is a problem with the lens (or body).  No amount of calibration will correct defaults that produce results like this.  A problem like this may need service from canon... but you might never know without running tests...

canon rumors FORUM

Re: AFMA Callibration - How much is too much?
« Reply #27 on: August 08, 2012, 02:47:33 PM »

neuroanatomist

  • CR GEEK
  • ********
  • Posts: 13952
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #28 on: August 08, 2012, 03:07:21 PM »
There is some great information here. What I am not reading is that these adjustments have made a noticeable difference in IQ. Is that just implied?

One unit of AMFA is 1/8 of the depth of focus for the lens at max aperture, so an adjustment of units means your combination is off by a full depth of focus.  Whether or not that makes a difference depends on how you shoot.  If you only shoot stopped down to f/5.6 or narrower and your subjects are always relatively distant, AFMA likely will not matter.  If you shoot at wide apertures, with close subjects, basically any time the DoF is shallow, it makes a big difference.  Obviously, the amount matters, too - a 2 unit adjustment on an f/4 lens isn't going to make much difference, but a 4 unit adjustment on an f/2 or faster lens will be noticed.

Absolutely.  The increase in IQ by calibrating AFMA is staggering.  It is both quantifiable and easily noticed when shooting is the field.  If your AFMA is off by even a little bit your image IQ will drop dramatically.  I have attached a chart that plots IQ for one of my lenses through the AFMA range that was generated by foCal.  Before I calibrated this lens I was always complaining that my images looked soft at 100%.  As you can see the optimal AFMA point for this lens is +8.  I was shooting at the default of 0 and my IQ was suffering.

Right - a big difference because you were off by a full depth of focus.  But it really is lens/camera dependent.  If you look at your plot, you can see that if you set a value between 7 (or perhaps 5) and 13, there would not be a meaningful, real-world difference.  If the lens with your camera had been in the -3 to +3 range, you'd likely not have needed the adjustment at all.

But...when it's needed, it makes a huge difference.
EOS 1D X, EOS M, and lots of lenses
______________________________
Flickr | TDP Profile/Gear List

JEAraman

  • Rebel T5i
  • ****
  • Posts: 108
    • View Profile
Re: AFMA Callibration - How much is too much?
« Reply #29 on: August 08, 2012, 03:35:13 PM »
I would be interested to see what others have had to AFMA input into their 1DX's.

Here are mine.

16-35L II: W=4, T=7
24-105L IS: W=4, T=2
28-300L IS: W=4, T=4
70-200L IS II: W=2, T=2 (haven't done it with 1.4x/2x TCs yet)
100-400L IS: W=-1, T=1
35L: 4
40 pancake: -1
85L: 0
100L Macro IS: 3
135L: 0

Did you do them manually or did you use Focal?

thanks

canon rumors FORUM

Re: AFMA Callibration - How much is too much?
« Reply #29 on: August 08, 2012, 03:35:13 PM »