December 22, 2014, 12:21:26 AM

Author Topic: Help me test large aperture Prime AF accuracy  (Read 982 times)

@!ex

  • Canon 70D
  • ****
  • Posts: 295
    • View Profile
    • My portfolio
Help me test large aperture Prime AF accuracy
« on: March 27, 2012, 01:40:57 PM »
So I have a the new 5dmk3 and a 50mm 1.2L.  I have been trying to get the micro adjustment on the lens just right, but but I figured out why it had been a bit confusing up to this point.  If I use the center point (or any AF point within the metering ring) I get spot on tack sharp AF on like 90% of the shots with 0 MFA.  BUT, if I use one of the outer AF spots I get pretty bad back focusing (like +15).  I thought maybe it was the body I had, so I just went back to the camera store and exchanged it for another body hoping that I would at least be able to figure out if that was the variable.  Turns out this body is exactly the same.  This makes me think it is either difficult for the AF to get a good lock due to the aberrations of the glass the closer you get to the edge, or something wrong with the lens itself (it is brand new and date code is uz0909).  I'm also just wondering if this is the price you pay with AF on really fast glass.  That is where you come in.  If you have fast glass (1.2) try checking the AF accuracy on center point and a more lateral point and lets see if we can come to a consensus.  I'm especially interested in 50mm 1.2 as thats what I'm working with.  I tested my 70-200 2.8 and it's pretty spot on with no MFA, on both center and lateral AF points.  I'll post a few test shots of my 50 when I get home later.

canon rumors FORUM

Help me test large aperture Prime AF accuracy
« on: March 27, 2012, 01:40:57 PM »

xthebillx

  • Guest
Re: Help me test large aperture Prime AF accuracy
« Reply #1 on: March 27, 2012, 02:18:42 PM »
So I have a the new 5dmk3 and a 50mm 1.2L.  I have been trying to get the micro adjustment on the lens just right, but but I figured out why it had been a bit confusing up to this point.  If I use the center point (or any AF point within the metering ring) I get spot on tack sharp AF on like 90% of the shots with 0 MFA.  BUT, if I use one of the outer AF spots I get pretty bad back focusing (like +15).  I thought maybe it was the body I had, so I just went back to the camera store and exchanged it for another body hoping that I would at least be able to figure out if that was the variable.  Turns out this body is exactly the same.  This makes me think it is either difficult for the AF to get a good lock due to the aberrations of the glass the closer you get to the edge, or something wrong with the lens itself (it is brand new and date code is uz0909).  I'm also just wondering if this is the price you pay with AF on really fast glass.  That is where you come in.  If you have fast glass (1.2) try checking the AF accuracy on center point and a more lateral point and lets see if we can come to a consensus.  I'm especially interested in 50mm 1.2 as thats what I'm working with.  I tested my 70-200 2.8 and it's pretty spot on with no MFA, on both center and lateral AF points.  I'll post a few test shots of my 50 when I get home later.

The 50 1.2 is known for sometime severe focus shift. That particular lens is optimized optically for 1.2, thus exhibiting certain tradeoffs further down the aperture range (most significantly between 2.8 and 4.0). You may be experiencing field curvature/spherical abberation and not an AF issue.
http://diglloyd.com/articles/Focus/FocusShift.html

Try manually liveview-focusing on the chosen AF point, make an exposure, then turn on AF and make another exposure.

The 50 1.2 is a bear to learn and has a steep learning curve, as 1.2 offers almost no focus, so focus-recompose isn't really an option.

canon rumors FORUM

Re: Help me test large aperture Prime AF accuracy
« Reply #1 on: March 27, 2012, 02:18:42 PM »