retrofocus wrote:
Could be cause by slight changes in the light conditions on your setup - makes a huge difference in FoCal Pro. Your copy is okay, you will sometimes see some deviations in MA, especially at very wide apertures. You can get a better view by just running the MA at a bit of slower speed, e.g. f/1.4 or f/1.8. If you get a MA value which is consistent with one of the f/1.2 values, then take this one.
No light changes. But your suggestion to calibrate at smaller aperture is helpful. I'll try that.
The other thing is that it calibrated my 50/1.4 perfectly and I haven't had any focus issues with that lens at all. So I am wondering if my 50/1.2 is not quite right. I'm comparing them right now to decide which to keep.
Depends. If contrast of the lens on that copy is not great and/or if the lens is a somewhat inconsistent performer, it may help to try at 1.4 or 1.8. It doesn't cost much in terms of time to push a few mouse clicks and let the test run...
Have you looked at results curves from FoCal?? Trying to make consistent results out of inconsistent performance is pretty much what it tries to do... Even a good MA value usually has some data points that are below average.
If the software is making a determination based on contrast of edges, it might help. I'd at least try it at look at the results curve. There may be some clues in there as to what is happening. It takes what? 30 seconds of time to kick of the test? So why not try it...
mco_970 wrote:
Have you looked at results curves from FoCal?? Trying to make consistent results out of inconsistent performance is pretty much what it tries to do... Even a good MA value usually has some data points that are below average.
If the software is making a determination based on contrast of edges, it might help. I'd at least try it at look at the results curve. There may be some clues in there as to what is happening. It takes what? 30 seconds of time to kick of the test? So why not try it...
yeah I well but this may very well be a case where doing it by eye is a better way...
jzucker wrote:
yeah I well but this may very well be a case where doing it by eye is a better way...
Could be, but I'd rather tell FoCal to run the test and look at the results curve and output images. It's faster/easier and I seriously hate looking at test images any more than I have to.
I have used FoCal on a number of different cameras and up to12 different lenses across them. In my opinion, FoCal makes very accurate and consistent measurement. That said, it will only be as well as your equipment and technique will allow.
For example, the 7D and 5D2 were quite consistent, the two 5D3's are significantly more consistent. The 50mm f1.4 could be a bit flaky, but the 70-200 2.8 II was amazing.
As far as technique, mistakes I made: not turning off AF, inconsistent lighting, not pluging in WB, not masking the viewfinder eyepiece, putting tripod on a carpet, not weighing down tripod for extra stability....plus others I am sure I ave forgotten at he moment.
FoCal has added function over the releases to protect us from our mistakes, but it can't catch everything. I recommend carefully reading the user manual and checking it for changes with new releases. As I usually do my AFMA in batches, I usually review the guide each time...just to aid the nagging brain.
John: what do you mean by not turning off AF? Isn't AF the whole point?
I have read the manual and I'm pretty sure my technique was good. I have a 5D3 and the calibration was spot-on for my 50/1.4. I also just calibrated my Tamron 24-70. It required a +3 on the tele end and a +6 on the wide end; I haven't had a chance to shoot with it yet and see how that worked.
The 50/1.2 has been erratic and it's the one that, in real life shooting, seems to miss focus the most. (I'm not talking about the back-focus issue. It misses focus even wide open.)
badlydrawnboy wrote:
Alright, I got it working again but now I'm having a different issue. I calibrated my 50/1.2 one more time because the focus didn't seem quite right. I used the exact same target position and distance-to-target on a heavy tripod. The first time FoCal told me to set MFA at -3. The second time it said no adjustment (leave at zero). The third time it said +2.
Granted, these are small increments. But shouldn't it be the same each time if the distance to target is the same?
Does this mean I might have a wacky copy of the 50/1.2?
It means you have a normal, healthy, typical copy of the 50/1.2L. Mine had the same problems.
The issue is that at f1.2, it's basically so blurry that a bit more or less blurry isn't very useful to the software. The trick was to calibrate it stopped down to f2, which so happens to be my typical aperture. It also addresses the issue of focus shift - calibrating it at f2 means that it won't focus shift at f2.
I am so tempted now. then thing is when I was a beta tester, I felt that the program was not working for me consistently. I would want to test my several lenses, several times and and various focus distances, and since its not yet automated for 5d3, it seems like a huge pain. I am also deciding between the middle one and the pro one. the middle one is just over half the price. decisions, decision. But if canon released its sdk, so that he can make it fully automated, I think its worth buying, esp given the easy of recalibration periodically. I wish there was a way of trying it before buying it.
Argh. I'm having another problem, now with my Tamron 24-70. FoCal gave me +3 on the tele end and +7 on the wide end. That seemed to be pretty accurate - as long as I was shooting at the same distance I calibrated at (about 5 feet).
Then I did some close-ups @2-3 ft. and it was back-focusing significantly.
Is this a lens issue, a calibration issue, or both? How would you address this?
badlydrawnboy wrote:
Alright, I got it working again but now I'm having a different issue. I calibrated my 50/1.2 one more time because the focus didn't seem quite right. I used the exact same target position and distance-to-target on a heavy tripod. The first time FoCal told me to set MFA at -3. The second time it said no adjustment (leave at zero). The third time it said +2.
Granted, these are small increments. But shouldn't it be the same each time if the distance to target is the same?
Does this mean I might have a wacky copy of the 50/1.2?
I did a 'manual' MA of my 50/1.2, that is: I used a flat target and a home made ruler at an angle
After I found the right MA, I'm pleased with this lens. And I can get quite good results at f/1.2.
However, I'm impressed enough with FoCal that I didn't mind upgrading to it.