Sharpness of EF 50mm 1:1.2L USM?

[chris_shawn]

Hello,

 

I own the EF 50mm 1:1.2L USM and have the feeling that it just doesn't focus well enough. Maybe it is just me,

but I find it difficult to get the focus 100% right with this lens.

 

Therefore I would like to ask if there is a trick of testing if the lens is working correctly. Are there

testsheets somewhere that I can try out? Or do I have to bring the lens to a Canon dealer and have it tested?

 

Are there any known problems with that lens?

 

I know my questions are difficult to answer but maybe someone can help.

 

 

Thanks a lot

Chris

[mark u]

What body are you using, with what focus screen? It's probably meaningless to attempt an answer without knowing this.

[chris_shawn]

I use a Canon EOS 1Ds MKII with the regular focus screen (that is included when you buy it).

[rainer_t]

Chris, as a formal test, you can do the focus test as described at the end of ...

 

http://focustestchart.com/chart.html

 

Just a few thoughts ... if you use the 50/1.2 wide open, or near there at portrait distance, depth of field is

very shallow. The slightest movement of the camera or the subject, will change results dramatically. Also,

selection of the right focuspoint is important.

[chris_shawn]

Thing is, I used it as closed (f16) as possible and focussed manually. Now, I just want to find out if it was my failure or a technical problem.

[rainer_t]

From ...

 

http://www.photozone.de/canon-eos/158-canon-ef-50mm-f12-usm-l-test-report--review?start=1

 

... I would think that you get maximum sharpness around f/5.6 ... which doesn't mean f/16 is bad .

[chris_shawn]

Here is my result from the test. It doesn't look good but I am afraid that maybe my test circumstances weren't 100% right. E.g. if my camera is at 46 degree or 44 degree the result will be different than it should...<div></div>

[mark u]

The lens has a reputation... try looking at the more sensible parts of this thread and some of the links from it:

 

http://forums.dpreview.com/forums/read.asp?forum=1029&message=27757148

 

Basically, at close subject distances, focus varies with aperture. Even wide open, the lens may need careful calibration. However, "user error" is extremely common when using this lens - the slightest recompose or movement after focus will show up.

 

In the mean time I strongly suggest you consider alternative focus screens, as these will help:

 

http://www.prophotohome.com/forum/canon-1-series-digital-slr-eos-5d/72864-canon-1dii-focusing-screen-manual-focus.html

 

It can also be useful to adjust your technique: I borrow one from macro work. Focus lock, lean back slightly and then take a series of shots as you lean forward again to provide a focus bracketed series. The movement required is very slight, and needs some practice - however, it makes up for the fact that no viewfinder screen I know of is actually optimised for lenses as fast as f/1.2, and likewise the AF system is probably operating right at the margins of acceptability. Those 1/3rd DoF f/2.8 points aren't significantly more accurate with a lens faster than f/2.8 - at f/1.2 they are about equivalent to "within DoF", while the normal precision points may not be accurate enough. If you want to understand why this is so, then read this:

 

http://doug.kerr.home.att.net/pumpkin/Split_Prism.pdf

[rainer_t]

The testresults arent overly sensitive to the 45 degrees. (40 or 50 degrees will do as well ... the amount of malfocussing wil change a bit, but not the tendecy itself).

 

From the picture, I would guess you have about 10mm backfocussing. Is this repeatable ... if you do 10 test, do you get this result in 8 of 10 cases, or is the distribution of the results different. Also, what do you get if you focus manually?

[chris_shawn]

@Rainer T

 

I guess I have a problem with the lens. I shot with auto-focus and the red squares appeared totally on the thick black line. In the viewfinder it looked perfectly in focus.

 

 

@ Mark U

 

Focus bracketing sounds like a good choice to have a perfect photo. But it only works for still live. I only shoot living models (people) and they usually move in front of my camera.

I will read the links you provided now. Thank you!

[mark u]

The focus test has shown you how narrow the DoF is at f/1.2: if your models move you will only achieve correct focus by accident at this aperture. You will need to persuade them to hold their poses while you shoot, or arrange that their movement provides the focus bracketing with the camera on a tripod.

[chris_shawn]

When being a perfectionist this is all really a pain. Thanks indeed for your professional advise. Reading in the dpreview forum made me happy after all. I am not the only one with that problem.

 

Maybe I will try to get my lens calibrated. Maybe they can work it out.

[mark u]

You can get an idea of the effective aperture at which the standard viewfinder screen operates by progressively stopping down while using DoF preview. When the image darkens, the lens aperture is narrower than the limit imposed by the focus screen. I doubt you will notice any change until f/2.8 at least. This is why the standard screen is quite useless for judging focus with a fast lens.

 

Another factor with this lens is that the soft background bokeh is achieved at the expense of some lateral CA (visible in your focus test shot - at least the foreground comes out warmer than the background), and spherical aberration:

 

http://www.vanwalree.com/optics/spherical.html

 

this also accounts for the aperture related focus behaviour and some of the other focus difficulties.

[chris_shawn]

Do you recommend a viewfinder magnifier? Would this make it easier for me to focus (when focussing manually of

course; which I do most of the time as I try to focus on the eyes always and selecting the focus point to the eye

is just way much work on a moving subject than focussing manually)? Are there good ones available?

[mark u]

Viewfinder magnifiers are best with still life or macro work. A super precision screen is better otherwise. I think I'd try the Ec-S screen, notwithstanding the metering problems on the 1Ds Mk II. After all, with studio work you are probably working with manual exposure settings and manual strobe settings too. And you may be able to determine some compensation offsets (these may vary according to the aperture you shoot at).

[derek_stanton2]

Here's my experience with the 50L

 

I bought one online, from a local dealer. It was shipped to me via UPS. I received it, and made some simple test shots to see

how it compared to the 50/1.4 i had owned for years. The tests were shot at 1.2 - 2.8, at close range, specifically looking for

focus accuracy, and sharpness. I also tried some out of focus stuff, to see how the bokeh differed from the 1.4 lens.

 

That particular 1.2L lens did not focus accurately at wide apertures. It front- or backfocused (can't remember which) by a

few centimeters.

 

I returned the lens to the store, in person, in order to try other samples. At the counter, i tried similar focus/sharpness tests,

with two other samples. The second lens also had the same mis-focus issues. The third sample seemed to focus

accurately. But, even when it was in focus, it was still not as sharp as my old 50/1.4.

 

The bokeh on the 50L, however, is even nicer than that of the 1.4 lens, and i've always liked the 1.4 lens.

 

I desperately wanted to keep the 50L, as i had been posting online my desire for one well before it was even announced.

But, because it did not outperform my existing lens, and at the expense of both money, and size/weight, it made no sense

for me to keep the L.

 

If you read dpreview, there are many conflicting opinions on this lens. There are a great many who, like me, suffered too

many problems and feel the lens is not worth the headaches and required workarounds. Then, there are a few people who

are delighted with the lens. But, i rarely/never notice any positive opinions when that person ALSO has a good sample of the

50/1.4 to compare. The people who like the 50L seem to be judging it on its own merits and not making any value

judgements.

 

As well, there are threads about people having the lens calibrated to their body by Canon. This was something i also

explored. I asked Canon about this, and was told that it could be done "to factory spec." But, when i asked how i could

receive two NEW lenses that were not calibrated to factory spec, i was told that perhaps issues occur during shipping.

Perhaps due to vibrations or somesuch. That's almost a reasonable explanation. But, then, i asked, If i send the lens to you

and you calibrate it accurately, why wouldn't the same vibrational issues occur in transit when it's returned to me? And, what

about further transport? As i intended to use the lens during frequent, long flights, i imagined that would, again, be a problem.

 

In the end, i've abandoned hope, unless Canon specifically addresses the issue. I've read that people should NOT expect

the lens to focus accurately at close range, and that's the way it was designed, and that that problem was a necessary

design compromise — things of that nature. Which are, frankly, ridiculous rationalizations.

 

You may want to try a different focusing screen. I now have the EE-S installed in my 5D, and it seems to make a difference

when using other brands of lens via adapter. Still, though, i think you might be better off finding a really good sample of the

50/1.4, or trying the new Sigma 50/1.4, unless you're convinced you absolutely need f1.2. But, really, who does? How many

books or exhibits have you seen with images shot at 1.2?

[mendel_leisk]

Are you still in the warranty period? Email Canon: you can likely get the back focus remedied at a Service Center, for the cost of shipping the lens to them. If it's out of warranty, it still might be worth your while: they tend to have a fixed price for repairs/adjustments, by categories, and I think they will pay the return shipping.

 

It's a little frustrating being part of their Quality Assurance program though. They should catch this.

[mark u]

I've bookmarked this thread:

 

http://www.fredmiranda.com/forum/topic/652568/0

 

which shows the wide open bokeh performance of a good selection of fast 50-60mm primes. I really quite like the Canon EF 50 f/1.2L, but I think I'd be happy with the Sigma and Zeiss f/1.4 as a pair with different qualities, although there are a number of other outstanding contenders in an area which is at least partly down to artistic taste - I could probably make a case for adding at least one more. The Sigma also gets my vote for its AI Servo performance.

[mark_petrie]

While certainly not as sharp as the 85L or 135L, the 50L is no slouch. You'll find many f/1.2 lens that are a bit off

when doing that test. You may want to try an ec-b screen and MF, as that is ideal. Alternatively, send the body and

lens into canon for calibration.

[philosomatographer]

The EF 50mm f/1.2L is a ridiculously sharp fast lens (well, my copy certainly is) - at f/1.2, if you're not too

far out to the edges of the frame, it can out-resolve my 1D MKIIN sensor to the point of producing moire.

 

However, focusing accuracy is very difficult, and there it certainly is not always 100% correct. I treat mine as

a manual-focus lens, and the Ec-S focusing screen has made a *world* of difference, literally opened my eyes to

the in-focus area of the image.

 

Why do I treat this expensive lens as a manual-focus lens? Because in at least 25% of shots I take at f/1.2, AF

does not focus precisely where I want to. Whether this is a fault of the lens (likely), the camera (unlikely, my

1-series focuses spot-on with all my other lenses), user-error because of the difficult nature of placing a 50mm

f/1.2 focal plane (manual or auto) I don't know. When it hits the target, this lens is better than any I have

ever used - this includes lenses like the 100mm f/2.8 Macro, one of the sharpest canon makes.

 

Attached is an auto-focused image taken in harsh daylight at f/1.2, I focused on the Mercedes' badge. It's spot

on, and tack sharp. It gets soo much better when you stop it down. So if you are seeing softness, there is

something severely wrong with yours. Just note, however, that at f/16 on a 1Ds MkII any lens is going to start

getting seriously soft because of diffraction. I can already see the serious impact on my lower-resolution camera

at that aperture.<div></div>