No one really knows if there is any correction for this built into that len's firmware, or if this lens even exhibits it. The only time I heard it mentioned as being an issue was from Mark Tucker who was experimenting with ancient Petzval view camera lenses a while ago.
mine seems to be a good copy. i just opened a few shots i took last night around the house of some random small objects and haven't found any sharpness or AF issues.
i'll be doing focus scale and box shot tests later today just for kicks but am confident that this lens is a keeper. the bigger question will be in how much i use the 50mm focal length and if return on client work will pay for the investment.
for the record, my s/n is 661xxx and date code is UU1000. given the date code, it must have been one of the first ones made.
tkjaer wrote:
It also could be that Canon is shipping to the rest of the world now.
I know I still haven't received the one I ordered in Germany early December.
That is very possible, since Canon is just now getting around to introducing the lens in Japan. It is supposed to go on sale there around the 26th of January or thereabouts from what I heard on another forum.
So, initial focus chart results with the 50/1.4 - it actually front focuses a bit at f/1.4 (focusing variance is much greater than the 50/1.2, which is more solid) and by f/2.8 it is right on target where the 50/1.2L was off by a couple centimeters. This would easily explain William Castleman's sharpness results if his lenses were display the same problems. I'm not sure if William couples a focus accuracy test with his sharpness tests, I may have to drop him a line.
OK, I went out to shoot in some semi-frozen stuff - the first of the season here - and found an opportunity for a medium distance focus test on the 50L with 1Ds. This was all done on tripod, with mirror lock up. The AF point was over the leaf in the left of the 100% crops. Yes, there are several shots at 1.2. They were done to check AF consistency. I bumped the contrast in raw conversion because the light was very dull overcast. If you notice grain, it's somewhat enhanced by the high contrast. ISO 100. All were processed consistently. I didn't go past f/2 in this series. Since I've had very sharp and well focused results from every shot at f/2.8 and on, I didn't feel a need to look for focus shift or other problems in that range.
Here is the whole series. The selected focus point was over the leaf on the left. http://www.jlpx.com/FM_forum/fcompmontage.jpg
Clearly the acutance increases gradually from wide open to f/2, and I'm satisfied that the plane of focus is where it should be for each shot. Maybe I'm not fussy enough...
I'll say this - rain was falling for the two hours I was out, and I was happy to have a weather sealed body and lens!
I also feel that I got a few shots that are much more worthwhile than this one (that's not saying much!), but after all we're discussing technical problems and shortcomings, not artistic merit. At least this was more fun to shoot than a ruler.
Jess, your plane of focus does indeed look like it is where it needs to be. I would have gladly accepted my lens if I had been able to produce results like that.
i just ran box shot tests and a focus scale test and my copy failed miserably. gary's copy blew mine away in sharpness and mine seems to backfocus quite a bit. and, i've never had an issue like this with L lenses so this one's a new experience -- and i've owned/used just about all of them. it's getting exchanged in-person tomorrow for another copy if they have one in stock.
i'm bummed because even though i don't need f/1.2, i'd like to be able to use it whenever i want. there's no way i'm paying this much for a lens that doesn't perform within its intended specs.
at 10+ feet this lens is razor sharp all the way to the extreme corners. if i needed it solely for landscapes at f/8~16 then i'd be happy as a clam. but, if i needed to shoot an up-close portrait of my girlfriend then she'd probably be quite pissed with the results.
Alright, the 50L is outta here. The backfocusing issue is outta hand. Seeing it both in practical shots and focus tests. Here's how bad it is (focused on her right eye, of course):
50/1.4 @ f/2.8
50/1.2L @ f/2.8
Just a bit of background... the above shots are the "best of three" for each lens. That's right, the 50/1.2L shot shown here is the best of three shots taken. ALL of the 50/1.4 shots were acceptably in focus. My process is to defocus the lens, shoot, defocus, shoot, defocus, shoot.
The 50/1.2L's performance is nothing short of dismal, and this is after having Canon work on it. It seems pretty clear that there's good copies out there, but this clearly isn't one of them. I'm seeing the same thing across the board - the lens focuses very well around wide open, but gets progressively worse as I stop down. The problem is most noticeable below about 6ft, but the the 50/1.4 is still noticeably more accurate around 10ft, which is closer to where I need to use the lens.
Sam/Gary,
Don't the two of you have the same problem with this lens?
--Lens appears to focus correctly wide open, but once stopped down it starts backfocusing?
sam, i'm right there with you. my 50 1.8 mkII just blew the 50L away in wide open tests. not only is it sharper but it also autofocuses accurately. a $75 lens should not be better than a $1600 lens. that's frustrating.
CMOS wrote:
Don't the two of you have the same problem with this lens?
--Lens appears to focus correctly wide open, but once stopped down it starts backfocusing?
Yes.
Focus distance on the eye shot was probably about 3-4ft, didn't measure it.
Sam Bennett wrote:
Alright, the 50L is outta here. The backfocusing issue is outta hand. Seeing it both in practical shots and focus tests. Here's how bad it is (focused on her right eye, of course):
Just a bit of background... the above shots are the "best of three" for each lens. That's right, the 50/1.2L shot shown here is the best of three shots taken. ALL of the 50/1.4 shots were acceptably in focus. My process is to defocus the lens, shoot, defocus, shoot, defocus, shoot.
The 50/1.2L's performance is nothing short of dismal, and this is after having Canon work on it. It seems pretty clear that there's good copies out there, but this clearly isn't one of them. I'm seeing the same thing across the board - the lens focuses very well around wide open, but gets progressively worse as I stop down. The problem is most noticeable below about 6ft, but the the 50/1.4 is still noticeably more accurate around 10ft, which is closer to where I need to use the lens....Show more →
Sam I had excatly the same problem but it only happens every so often, Sometimes it is alright. I did exactly the same test and had exactly the same result as you. I mean exactly.
So I rang my dealer this morning and told him I don't want a replacement lens. He can keep it and he agreed to it. I will leave the credit there and I will maybe purchase the 1dmk3 or 1dsmk3 when it is release.
Wide open my copy has been excellent and very consistent. Even stopped down I have gotten excellent shots. However after conversing with Gary offline for the last several weeks as well as reading Sam's findings I was intruiged to see how mine would fare in a similar close portrait test.
At 2.2 and 2.8 at close distances indoors with tungsten lighting (and using flash assist), my copy too backfocuses approximately 1", and it does so consistently after locking focus. I also noticed I got significantly less in focus images when shooting in AI servo and knowing my subject was in focus. Out of 16 shots done in one shot mode last night, only two focused correctly and were sharp. One was at 2.2, the other at 2.8.
added:
Oddly enough just last week, I ran focus tests on the lens with stationary objects from 4 feet to 10 feet and it passed flawlessly
I am now faced with the dilemma of asking my retailer to exchange it or gamble with Canon's service. Seeing Gary send his in several times and Sam's visit to Jamesburg doesn't have me convinced they'd be able to do anything with it. (And unfortunately, I cannot send my body in with it as I need it for regular shoots, and truth be told no one should have to send a body in for their lens to work properly).
Manually focused the lens is outstanding at all apertures. I truly believe there is something wrong with electronics and how the lens is sensing the focus distance and that is something I've been discussing with Gary the last week. It cannot be the optics since manually focused it functions as expected. So there has to be a bad batch of electronics with this lens.
It wouldn't surpise me if all copies of this lens had the same problem. But what is certain, is am sure none of us paid close to $1600 for a manually focused lens.
At least this was more fun to shoot than a ruler.
