bushwacker wrote:
I bet Canon has that on their factory.... how do u think they calibrate EOS bodies and lenses?....
Now if they would only release the software to the consumer ... hmm ... wouldn't that save them a lot of employee hours in calibrating lenses to bodies? I'd love to see them release that for the XXD line of bodies.
I almost had a "soft copy" of my very sharp 24mm F1.4 (and it's been tweaked by Irvine too) when I mounted it on my new 50D. I MA ed the 50D in the venue (using the much better LCD to detect critical focus) before the performance (much better focus somewhere around +15).
With 15 + lenses and 4 MA bodies I have found NO correlation between the lenses and bodies. But the need (and benefit) of adjustment is real. About 70% of my lens/body combos need significant MA adjustment.
I'd like to see the MA feature include not only different lenses, but different focal lengths, for the zooms. I have found different front/back focussing maladies at different FL's, especially in the 24-105!!
Could a firmware update address that?
I ended up testing only at the near end of the focusable distance for the lens thinking that the narrower DoF closer to the lens would require more precise positioning of the elements to get tack sharp focus - if I focus on something 100 yards away, the DoF is more and should prob allow for slight variations in the focus ring position that will still yield tack sharp focus.
The link for downloading the chart is at the bottom of the page.
My methodology:
Position chart, taped down to a flat board, at 45 degrees (short spirit level assures this). Set up camera on tripod at same level as center of chart so that chart just fills screen. Set lens wide open (I've found that the back focus in my most severe example, the 50 f/1.4, went away at f/11).
Take a series of 9 shots with settings at intervals of 5 (-20, -15, etc.), noting carefully which frame number is which, and leave camera on tripod.
Import into Lightroom and immediately note in the keywords what the setting was for each (as far as I can tell, this info is not stored in EXIF data). Examine images one at a time at a magnification where I can see the entire chart. It quickly becomes apparent where the focus changes from front focus to back focus. If that happens between, say, -10 and -5, I put the card back in camera and shoot 4 more (-9, -8, -7, -6) and examine again. I'm looking for the shot where a back number (say, 40) is just as clear (or fuzzy) as the front 40.
Here were my settings using LiveView per http://blog.willshootphotography.com:/2008/10/canon-50d-af-mi.html
50 f/1.4: -10
100 f/2.8: -2
70-200 f.2.8: -2
24-105 f/4: -20
And here were the settings using the chart from focustestchart.com:
50 f/1.4: +4
100 f/2.8: -6
70-200 f.2.8: 0
24-105 f/4: -15
I can't see a pattern in all this. Went out just now with the 100 macro and shot flowers. Spot on. And I'm happy with performance of the 50. Now on to the 24-105.
Open questions to explore: 1) On the zooms, should I do this at both maximum and minimum zoom? 2) I'm setting focus for wide open aperture; what happens when I stop down? To use example from above, now that I've corrected focus on the 50 at f/1.4, is it still good at f/11?
Will Swetnam wrote:
Here is the test that I used - I found that a combination a target I found on the web, and one that I created myself did the trick.
I ended up testing only at the near end of the focusable distance for the lens thinking that the narrower DoF closer to the lens would require more precise positioning of the elements to get tack sharp focus - if I focus on something 100 yards away, the DoF is more and should prob allow for slight variations in the focus ring position that will still yield tack sharp focus.
Regarding your quote below, I use Canon's ZoomBrowser EX program to read the M/A adjustment amount # and check focus spot. Pretty easy to use...
"Import into Lightroom and immediately note in the keywords what the setting was for each (as far as I can tell, this info is not stored in EXIF data)."
Also, I get much better M/A results since I started using Canon's Ec-B focus screen, you can actually see letters and lines bend until they line up correctly when manual and M/A focusing, especially useful on my 600 and 1.4/2.0 tc's.
I got an aftermarket Ec-B focus screen, probably not a good idea as it's not quite square with the camera. The bigger problem, though, is that not all my glass is real fast, so it's difficult to see things at, say, f/5.6.
After posting I noticed that the MA amount shows up on some screen (Func? Info?). Very handy.
Polansky wrote:
Here you can find a very accurate method, which works great and is also easy to use: Perfect AF Microadjustment method
For the benefit of more recent owners of the 1-series, 5D2, and 50D, I'm showing your link again. This method is pretty easy. Interestingly, I found my fisheye was backfocusing about 8 marks which surprised me since I've never noticed any issues with it (the benefit of stopping down with a very short lens). The 16-35, 24/1.4 and 35/1.4 were both front focusing just a bit (-3 adjustment each). My longer lenses were all right on the money. The 24-70 was right on as well.
I found the microadjustment tool quite useful on my 5D MkII camera. This method worked quite well for me and it is fairly easy to set it up if you have a larger monitor screen:
I adjusted my zoom lenses for the most often used focal length, e. g. at 17 mm of the 17-40 lens and and an average number out of different focal lengths with the 24-70. There was a small improvement after microadjustment, visible in a bit larger RAW file size:
17-40 f4 L lens: -3 microadjustment (17 mm)
24-70 f2.8 L lens: -5
