I'm interested in the 50mm f/1.2L as an available-light prime, and after reading lots of posts on this site, shot a bunch of raw images at varying apertures at my local photo store last week with my 5D.
The images look good - reasonable sharpness wide-open with improvement as it's stopped down but with better contrast and sharpness at wide apertures than my previous Canon 50/1.4's. Although the 1.2L produced a fair bit of CA, this was easily correctable in Lightroom.
I went to the store to buy the 1.2L today and ran a quick focus test on the lens before committing to purchase - good thing I did, as when focussed at subject distances of 2-4 feet the store's demo lens consistently back-focused. The strange thing is that this back-focus didn't appear like the focus-shift that's documented elsewhere in these forums, but had the same focus point at each aperture but with this point of focus being equally shifted backwards from the intended point of focus regardless of the aperture used to shoot. In other words, the point of critical focus was virtually identical for all apertures, but each was also identically shifted rearwards from the subject's focus plane.
Testing another copy of this lens led to exactly the same results - very consistently accurate mis-focus.
My 5D focuses dead-on with my other lenses, as it did with my 100 macro USM to show the salesperson that my 5D was not the issue - the 100 macro provided perfectly sharp subject focus in every shot at a variety of apertures and subject distances.
Other than this AF "issue" I love the imaging and build qualities of this lens but don't want to risk its purchase unless it can produce accurate focus. I don't think it appropriate to spend these sorts of funds for a semi-AF lens.
Although I love the 50mm focal length on a FF body, I'm not a fan of either the 50/1.8 or the 50/1.4 (having owned a 1.8 and two copies of the 1.4). I'm reluctant to purchase a lens that needs calibration straight out of the box when there's no assurance that calibration will correct the problem, and when I likely wouldn't receive the lens back from servicing before the end of the two week return period.
Has anyone had their 50mm f/1.2L calibrated after initial back-focusing problems and obtained AF accuracy after re-calibration? Anyone had their lens re-calibrated and not been satisfied with the AF accuracy afterwards??
I've had the focus shift also. No amount of calibration can fix this. It will be calibrated for f1.2 and as soon as you stop down it will still back focus. I would love this lens if it didn't do it, but it's a design flaw (or feature depending on who you ask). Unless Canon changes the design or makes a fix, it's not going to change.
This issue comes up all the time, and even so, many people like the
lens enough, optically, to put up with this issue. Use of peripheral
focus points seems to help. joemama over at dpreview agonized over it
for a long time then finally purchased it and loves it.
One shouldn't have to put up with this on a lens of this price range.
Frankly, it's an embarassment for Canon, as is the lingering 1DmkIII focus
issue.
What this teaches me is this cardinal rule: Do not be an early adopter.
If many others rave about a body or lens I'll consider it. But I won't preorder
any equipment.
My prediction is that this will get fixed through some sort of redesign and
those who bought the current 50/1.2L will be stuck with focus problems,
but will find a way to work around them and just live with it.
Your description fits quite well with what I experienced. I had backfocus also at f1.2, which should rule out the focus-shift. I tested three copies of the lens on two different XTi bodies and all had the problem. I picked one of lenses and sent it in together with my XTi.
The problem is that I haven’t got it back yet.
But it has been seven weeks now! so I expect it any day. Tell me if you are interested in the outcome even if you don’t get it now and I will give you a message as soon as possible.
I got my return period prolonged until this issue is resolved. I think that is the least one should expect.
This issue and others like it are probably the reason Canon is introducing the focus calibration system we see in the 1D3.
I am interested if the backfocus occurs when you focus manually. Has anybody tried it? Also manually focussing while stopped down might be an option. On my 30 year old Contax RTS there is a lock on the stop down button so that you can focus while stopped down. Even with Canon's electronic system it would be easy to include a custom-function that set the stop-down button to be sticky: you could press to stop down, and press a second time to open up again. Needless to say, Canon's usability experts have seen to it that Canon bodies still lack that excellent feature so it's practically impossible to focus while the lens remains stopped down.
The focus problem is the same whether I use AF or MF on my 5D.
I was wondering if this issue is camera specific. May be it's only on the 5D and 1D Mark III. I have read some happy results from the 40D camp. Does any one know for sure?
menatwork wrote:
The focus problem is the same whether I use AF or MF on my 5D.
I was wondering if this issue is camera specific. May be it's only on the 5D and 1D Mark III. I have read some happy results from the 40D camp. Does any one know for sure?
I don't think it's camera specific. But a 1.6x camera will help to hind the issue. The back focus is something like 1 cm on a 5D. On a 40D, that would be 2/3 cm due to DOF differences. So it's easier to miss the mis-focus.
The focus problem is the same whether I use AF or MF on my 5D.
Did you try manual focussing while stopped down? As I said, you will need to be a contortionist to get around Canon's strange attitude to ease-of-use, but presumably it can be done. Or does the 50L have an annoying electronic manual focus ring which gets disabled during stop-down?
The back focus is something like 1 cm on a 5D. On a 40D, that would be 2/3 cm due to DOF differences.
As I understand it, there are no DoF differences, and there should be no lens geometry differences either. Using a smaller sensor doesn't alter the lens or its distance from the film plane. If it's 1cm on a 5D it should be 1cm on a 40D. Correct me if I'm wrong.
brainiac wrote:
As I understand it, there are no DoF differences, and there should be no lens geometry differences either. Using a smaller sensor doesn't alter the lens or its distance from the film plane. If it's 1cm on a 5D it should be 1cm on a 40D. Correct me if I'm wrong.
Yes and no. If you shoot the same headshot with a 5D and a 40D, you stand 1.6 times farther away with the 40D which is why the 40D appears to have more DoF at the same aperture. If you shoot from the same distance, then the DoF is the same but the image is cropped in 40D. For most users the practical experience is that a crop body has more DoF.
Edited by LiquidAir on Sep 21, 2007 at 09:44 AM GMT
Edited by LiquidAir on Sep 21, 2007 at 09:45 AM GMT
Brainiac Wrote:
Did you try manual focussing while stopped down? As I said, you will need to be a contortionist to get around Canon's strange attitude to ease-of-use, but presumably it can be done. Or does the 50L have an annoying electronic manual focus ring which gets disabled during stop-down?
Yes, I tried MF and stopped down but the result was still not what I'm expecting for a stop down f1.2 L lens (pictures are still blurry).
brainiac wrote:
As I understand it, there are no DoF differences, and there should be no lens geometry differences either. Using a smaller sensor doesn't alter the lens or its distance from the film plane. If it's 1cm on a 5D it should be 1cm on a 40D. Correct me if I'm wrong.
Actually I think we are both wrong. I believe there is a DOF difference due to the 1.6x multiplier, however I got it backwards. The 40D would be seeing 1.6cm difference. So why someone says they don't have it is beyond me. It is present and I've seen it on both FF and 1.3x crop bodies.
I have to chime in with all the negative. I have had this lens for about 6 months. Focus is dead on 99% of the time on a 5D and Mark II N. It will mis focus once in blue moon, but so did my 1.4 . I shoot it mainly from 1.2 -2.5..never higher and have full confidence in the lens. Just my 2 cents
bcaslis wrote:
I don't think it's camera specific. But a 1.6x camera will help to hind the issue. The back focus is something like 1 cm on a 5D. On a 40D, that would be 2/3 cm due to DOF differences. So it's easier to miss the mis-focus.
The shift is about 2 cm on my XTi. Here is a crop at 3 ft and f/2.8. The vertical black line is the center. Five tick marks to a cm.
jjf88 wrote:
I have to chime in with all the negative. I have had this lens for about 6 months. Focus is dead on 99% of the time on a 5D and Mark II N. It will mis focus once in blue moon, but so did my 1.4 . I shoot it mainly from 1.2 -2.5..never higher and have full confidence in the lens. Just my 2 cents
If you shoot at f/2.5 and around 3 feet you would see the problem if you have it. The shift can be subtle with the focus point just catching the front edge of the DOF or even better if the lens is slightly front focused. Beyond 3 feet and the increasing DOF starts to compensate for the shift and it is not very noticable by six feet.
>>...Using a smaller sensor doesn't alter the lens or its distance from the film plane.
>Yes and no. If you shoot the same headshot with a 5D and a 40D, you stand 1.6 times farther away with the 40D which is why the 40D appears to have more DoF at the same aperture...
Oh yes - I agree with that. I was just envisioning doing a focus test from the same position of the same subject. Once you start moving forwards or backwards it's a different story, as you say.
