Any recent experience on 50 1.2L focus-shift issue?

[waving_odd]

I've been checking this lens since its announcement 3.5 years ago.


Heard about many focus-shift issues (while some say back-focusing issue that I think it's misleading) that are caused by the lack of floating elements to correct spherical aberrations -- hence by design.


This makes me hesitate to get it because I plan to take advantage of its short minimal focusing distance (so I can get nicer close-up on subjects) with the possibility (depending on creativity) to stop down a bit to f/2 or f/2.8 (although I primarily want it wide open at 1.2) -- all these will put it into the danger zone of focus-shift.


Some people said Canon will improve the more recent copies by having better quality control during manufacturing. Some said it's irrelevant because the issue is by design. [:S][*-)]


Anyone has some recent first-hand experience? Thanks in advance!

[genotypewriter]

I don

[Keith B]

I read many an article/review of this lens for years and held off on purchasing. But then someone wrote a more recent review of the lens and stated his experience with the current builds were different than the earlier models he had tested. That was good enough for me so I decided to give it a try.


My experience is; this lens is amazing. I do not experience focus shift. The consensus from what I read is that because the lens is wide open to focus then squints down to the actual f/stop when the shutter fires the focus will shift. From what I've read, the most notable shift is between f/2.0 and f/5.6 where the DOF gets tight enough you won't notice the shift.


I do not experience this at all. I have the same experience with any other fast lens. It is what you'd expect at 1.2 (reasonably sharp but not as sharp as the 85 1.2II) and gets sharper as the aperture gets smaller. I did the test running through all the full f/stops (except 1.2 was used instead 1.4). At f/8 it is amazingly sharp. This is the first time in the year I owned the lens I felt the need to test it. I was just always super happy with the images I got from the lens.


A little disclaimer; I have lots of experience with fast lenses and I have a very steady hand/technique. I usually attribute this to my successes with "finicky" fast lenses where others may have issues. Steadiness is huge factor when shooting at 1.2. Not just hand shake but also body sway control.I have had so much success with this lens that I know if I missed the focus, I'm sure it was user error.


As you mentioned, with the short MFD and shooting at 1.2 you can have some very amazing effects. It is a very fun lens. It is on my camera(s) 80% of the time.





PS





My copy is UY date code. I have read articles mentioning how Canon doesn't make all the pieces and parts in the lenses. Some are provided by other manufacturers and sometime they may change these outside manufacturers for one reason or another and this may not be considered QC but it could explain differences between production runs. Another factor may come into play; Canon does not have a production line dedicated to any one lens. They produce X amount of one lens, break down the line and set it up to produce another lens for X amount run. This would allow for evaluations and changes to be made since the production gets halted routinely. Not saying any of this happened (impossible to know since Canon would never own up to it) but it seems plausible.

[Dr Croubie]

just a quick check using an online Depth of Field calculator here: www.dofmaster.com/dofjs.html


Depth of field for 50mm f1.2 30cm away is 1.4mm for a 7/60/550/600d, 2.1mm for a 5d2/1ds3 (because of the larger pixel size).


I

[Fast Glass]

Actually those DOF numbers don't apply to real life, because if you were to compare different formats using the same focusing distance you have to use different focal lengths to get the same picture. Using DOF Master; 1.6 crop sensor, 50mm at f/1.2 has 6.51" of DOF at 10'. Full frame, 80mm at f/1.2 has 3.97" DOF at 10'. If you were to use the same lens and get closer with the FF you would get the exact same DOF as with using different focal lengths.


The only time you use that formula if you are focus calibrating your camera/lens combo, then 1.6 has thiner DOF.


John.

[neuroanatomist]

Depth of field for 50mm f1.2 30cm away is 1.4mm for a 7/60/550/600d, 2.1mm for a 5d2/1ds3 (because of the larger pixel size).

In addition to FastGlass' comments, just to be clear, pixel size has no effect on DoF - sensor size is the relevant factor.


John (FastGlass), I thinkDr Croubie's numbers have some merit in the sense that the focus shift issue is most prominent at or near the MFD, and the MFD is a lens property that is independent of sensor format. However, the 30 cm distance chosen as an example was a poor choice, since the MFD of the 50mm f/1.2L is actually 45 cm.

[Fast Glass]

John (FastGlass), I thinkDr Croubie's numbers have some merit in the sense that the focus shift issue is most prominent at or near the MFD, and the MFD is a lens property that is independent of sensor format.

True, but to bad DOF Master doesn't acount for optical design such as floating elements.


John.

[Daniel Browning]

In addition to FastGlass' comments, just to be clear, pixel size has no effect on DoF - sensor size is the relevant factor.

While it's true that pixel size is generally less important than sensor size when it comes to affecting DoF, I wouldn't go as far to say that it has no effect. Here are some other factors that affect DoF:

• Aliasing
• Processing (e.g. demosiac, sharpening)
• Display size
• Viewing distance
• Display resolution
• Viewer acuity
• Viewing environment

All of those factors and more can affect what can be discernable by the viewer as acceptably sharp.


For example, let's say I take a self-portrait and find that it looks perfectly sharp on the rear LCD of my camera such that about 3 feet (object space) looks in perfect focus. It could not be any sharper. Then, I get home and put it on the computer and decide to use it for my 30-pixel TDP avatar image. But at this reduced size, what used to be a blurry background is now much sharper, so the DOF is now more like 10 feet. The maximum sharpness for that display is achieved over a wider distance in object space.


Then I decide to make a 4x6 print. Now that I can see a lot more resolution, it becomes clear that some of the image is far sharper than other parts. Now it is possible for some of the image to be sharper, but it isn't. The acceptably sharp standard has changed again, but in the other direction. Now it's clear that the image is slightly front-focused, and the DOF only looks like 6 inches.


Next I make a 30x20 blow-up. Now it's painfully obvious how badly out of focus the image is. Add a more detailed demosaic and additional sharpening and it becomes even more clear just how thin the very sharpest part of the image is.

[neuroanatomist]

Thanks, Daniel. My bad. [] Perhaps I should have stated that as, "pixel size has no effect on the DoF value reported by standard calculator tools such as DoFMaster." Certainly, since true DoF is defined by that which is perceived as acceptably sharp, as opposed to that which is not, many other factors come into play. Here's one more for your list:

• viewer's state of inebriation

[:P]

[Dr Croubie]

yeah, ok, sensor vs pixel size, wasn't thinking straight. but anyway, i was just going on the aps-c CoC being 0.019 and 35mm CoC being 0.03mm. whatever causes it.


and yeah, i also noticed that the min focus distance was 45cm after i posted too.


as for the same-framing needing wider-lens on aps-c, and corresponding f/number differences too, yeah i know that too. but he said he's using 50mm f1.2 @ MFD, so that's what will be used, regardless of framing...





but anyway, wasn't this about the 50 f1.2 itself? the 85 f1.2 will have shallower DOF on the same camera at the same distance, and it has a floating system too. The MM figures are worse though (0.15x vs 0.11x without tubes)...