Back Focus Problem with Canon 50mm f1.2L?

Flyinggina

I recently bought the 50 mm f1.2L lens for low light situations. I knew when I bought it that the very narrow DOF when the lens is wide open or close to wide open would be a challenge, so I assumed that the fact that the sharp spot on the photos seemed never to be the spot I had focussed on was simply bad technique on my part. I got some lovely photos, but it was as though I was using one of those soft portrait lenses. When my husband remarked that the sharpest area in my photos (and also some he took with the lens) were consistantly behind the plane on which he and I had focussed, we decided it was time to test.

We used Tim Sullivan's focus chart, took the results to Calumet, and they replaced the lens with a new one and returned the first to Canon. The second one turned out to be as "bad" as the first, especially when compared with my 50mm 2.5, which we tested as well. (Of course, it has more DOF wide open than the 1.2, but it is instructive to see the difference in sharpness.)

So it was back to Calumet today where I exchanged the 1.2 for a third new lens and bought a 50mm f1.4 to try. Brought them home and tested them immediaetly, again using Tim Sullivan's Focus Chart.

The set up was the same for all photographs. Camera on a tripod. Tripod, level and chart taped to table. Camera at 45 degree angle to paper (measured using a triangle thingie -- OK, I'm not a technie - my techie husband did it). ISO 100. Remote shutter release. Single center focus point focussed on the "focus here" in the middle of the chart.

Here are the results:

50mm f1.2L lens at f1.2


50mm f1.2L lens at f1.4


50mm f1.4 lens at f1.4


50mm f2.5 lens at f2.5 (taken last night)



It sure looks as though my latest 50 f1.2 lens has a significant back focus problem. Am I reading this info correctly, or have I missed something.

Your thoughts?

Virginia

ziggy53

While both the EF 50mm, f1.4 and the EF 50mm, f1.2L are rated at minimum focus of 45cm (1.5 ft), I suggest a minimum of around 90 cm (3 ft) for this test. It looks like like you may be shooting closer than that.

Flyinggina

Interesting point, Ziggy. The photographs were taken at between 2 feet and 2 feet 3 inches depending on the lens. We'll try again tomorrow on the 1.2 and the 1.4 at 3 feet and see what we come up with. Thanks very much for the tip.


Virginia

LiquidAir

Flyinggina wrote:

Interesting point, Ziggy. The photographs were taken at between 2 feet and 2 feet 3 inches depending on the lens. We'll try again tomorrow on the 1.2 and the 1.4 at 3 feet and see what we come up with. Thanks very much for the tip.


Virginia

I usually do my focus tests at 2m (6 feet) because for me that is a fairly typical working distance so it is more representative of what I am likely to see in practice. The other thing I do is multiple tests alternately moving the focus closer and farther than the test point between each one.

kdog

Couple of things. Changing the distance to the subject may affect the error, but it's still an error. You should be able to get a reliable focus within the specifications of the lens.

I'm not a fan of that calibration test. I think you're better off focusing on a target that's parallel with your sensor plane, not at a 45 degree angle to it. Here's a link to a picture of my test setup, but you don't need to get that elaborate. http://www.jacara.com/photo/focus_test_jig_small.jpg

Odds are that if you have a problem, the problem is with your camera body, not the lenses that you've tested. I'm guessing you haven't ever attempted such a narrow depth of field with it before, which magnifies any back or front focus error. You would want to send a lens plus your body in for calibration.

Finally, I'll probably take some flak for this, but I don't think the focus system on the 20D is worth a damn at under F2. It probably still won't be perfect even after calibration. At least that's been my experience.

Have fun, and good luck.
-joel

Flyinggina

We redid the tests as Ziggy suggested at 3 plus feet. The problem was even worse for the f1.2 but the f1.4 continues to look pretty good.

Thanks for your suggestions, Liquidair. I often photograph at closer in than 6 feet (and am happy to have the shallow DOF) but will test at a greater distance tomorrow, just to see what happens. I'm not sure I understand your second point. We do start with the camera out of focus each time so that the AF has to do its job. We don't systematically go from closer to farther away focus, though. Will try that too.

Joel, your comments are intriguing. You are right that I have never used a lens with an aperture under f2.5 on the 20D before. I'll try to test it on a 5D before returning the lens. That should tell me whether the camera's AF is at fault.

Since the tests on the f1.4 suggest that the AF on that lens is working well with the camera, I'm going to try it out in the field to see if it does what I need. It may be that it is the best choice for what I want and it has the benefit of being considerably less expensive that the f1.2. And lighter too. I'll check out your test setup too.

I'm very intrigued by the idea that the 20D can't handle an f1.2 lens. I assume this is true of the 30D as well?????

Thank you all very, very much for your input. It is incredibly helpful and I'm learning a lot.

Virginia

ziggy53

Flyinggina wrote:

We redid the tests as Ziggy suggested at 3 plus feet. The problem was even worse for the f1.2 but the f1.4 continues to look pretty good.

Thanks for your suggestions, Liquidair. I often photograph at closer in than 6 feet (and am happy to have the shallow DOF) but will test at a greater distance tomorrow, just to see what happens. I'm not sure I understand your second point. We do start with the camera out of focus each time so that the AF has to do its job. We don't systematically go from closer to farther away focus, though. Will try that too.

Joel, your comments are intriguing. You are right that I have never used a lens with an aperture under f2.5 on the 20D before. I'll try to test it on a 5D before returning the lens. That should tell me whether the camera's AF is at fault.

Since the tests on the f1.4 suggest that the AF on that lens is working well with the camera, I'm going to try it out in the field to see if it does what I need. It may be that it is the best choice for what I want and it has the benefit of being considerably less expensive that the f1.2. And lighter too. I'll check out your test setup too.

I'm very intrigued by the idea that the 20D can't handle an f1.2 lens. I assume this is true of the 30D as well?????

Thank you all very, very much for your input. It is incredibly helpful and I'm learning a lot.

Virginia

Virginia,

You are approaching this with a great attitude, exploring every avenue and alley. I applaud you for your tenacity.

Too many folks discard things based on a single test, and that is not fair. It really does take the sort of "due diligence" that you are demonstrating to rule out one problem or another.

Do also test the f1.2L in a more organic and realistic environment before declaring it to be probematic. Test it in a similar situation to how you would actually use it were you to keep it.

I really do love my Canon EF 50mm, f1.4, wishing only that it had the ability to relay distance information to the camera for E-TTL II purposes. (Even though I usually use the flash in some sort of bounce mode, which turns off E-TTL II.)

LiquidAir

Flyinggina wrote:

We redid the tests as Ziggy suggested at 3 plus feet. The problem was even worse for the f1.2 but the f1.4 continues to look pretty good.

Thanks for your suggestions, Liquidair. I often photograph at closer in than 6 feet (and am happy to have the shallow DOF) but will test at a greater distance tomorrow, just to see what happens. I'm not sure I understand your second point. We do start with the camera out of focus each time so that the AF has to do its job. We don't systematically go from closer to farther away focus, though. Will try that too.

Joel, your comments are intriguing. You are right that I have never used a lens with an aperture under f2.5 on the 20D before. I'll try to test it on a 5D before returning the lens. That should tell me whether the camera's AF is at fault.

Since the tests on the f1.4 suggest that the AF on that lens is working well with the camera, I'm going to try it out in the field to see if it does what I need. It may be that it is the best choice for what I want and it has the benefit of being considerably less expensive that the f1.2. And lighter too. I'll check out your test setup too.

I'm very intrigued by the idea that the 20D can't handle an f1.2 lens. I assume this is true of the 30D as well?????

Thank you all very, very much for your input. It is incredibly helpful and I'm learning a lot.

Virginia

I alternate directions to see if there is any loosness in the coupling between the moving elements and motor or position coder. Sometimes a lens will be more accurate in one direction than it is in the other. The Canon AF system is open loop in the sense that the camera does not recheck focus after the lens has moved so accurate focus depends heavily on how accurately the lens position is measured.

claudermilk

Flyinggina wrote:

I'm very intrigued by the idea that the 20D can't handle an f1.2 lens. I assume this is true of the 30D as well?????

OK, starting the flak: :gun2 I disagree with that. I've used a couple of differen 85/1.2Ls on my 20D several times--some of them in quite challenging lighing fully utilizing the lighbucket f1.2 capabilities--and the AF worked just fine. Oh and since the comment was "under f2" I'll add in that my 50/1.8 works fine. There is a point where the AF starts losing it with little light & no high-contrast edges to latch onto--but this should be true even of a 1D AF system--just that the threshold is different.

IIRC there was a big brouhaha over the 50/1.2 and focus errors. Since I never intend to purchase the lens I didn't pay much attention. I know there was a long thread started by Andy when he got his and a couple of big ones over at FM as well. Maybe do some archive digging to see what the fuss was about.

kdog

claudermilk wrote:

OK, starting the flak: :gun2 I disagree with that. I've used a couple of differen 85/1.2Ls on my 20D several times--some of them in quite challenging lighing fully utilizing the lighbucket f1.2 capabilities--and the AF worked just fine. Oh and since the comment was "under f2" I'll add in that my 50/1.8 works fine. There is a point where the AF starts losing it with little light & no high-contrast edges to latch onto--but this should be true even of a 1D AF system--just that the threshold is different.

Thanks, I needed that. :giggle

I'm glad to hear you're getting good results at those aperatures. I think you must have a very good copy of the 20D. I spent a lot of time on dpreview in the past and many folks with the 20D report focus problems with the fast 50's. Now that I think about it, I haven't heard the same complaints on the fast 85's. I haven't tried an 85, nor have I tried a fast L lens. Perhaps they behave better on this camera.

Cheers,
-joel

claudermilk

Any time.

I'm starting to think I'm somehow blessed when getting camera gear. My 20D is usually run at the ragged edge of it's capabilites at ISO 3200 and the lenses on it wide open with barely enough light to make it work. It's shutter is well over 80,000 clicks. It's like the old Timex ads--it just keeps on clicking. Same goes for all my lenses--I always find myelf of this side of the good/bad copy debates since everything I've used has been at minimum perfectly functional if not simply magic (300/2.8IS and my 70-200/2.9 fall under that category).

water_boy

I brought mine 2 weeks ago and have found manual focus gives best results. I have learnt not to focus and recompose.
Great lens but need to learn how to use it.

Flyinggina

Just got back from returning the third Canon f 1.2 because of the back focussing issue. The ones I "bought" were the last three in inventory at our local Calumet. Careful testing showed even more severe back focussing in the last of the three. I began to think it must be the 20D and not the lens as kdog suggested might be the case. (If you read the manual, they specifically talk about excellent AF results at 2.8 and above.)

However, we were finally able to ran a test using a 5D body, and got the same results. Also did some quickie shots at Calumet today and there was no way around it. The AF was back focussing. I know it sounds unlikely, but it does appear that the last three f1.2L's in the store's inventory had AF problems.

I'm keeping the f 1.4. It is a little slow focussing in low light, but it gives very nice results, especially at 2.2 when there is sufficient light for easy focus. Plus it is lighter and a lot less expensive.

Water-boy - I would love to be able to use manual focus. Alas, I have a very, very strong lens correction in my glasses which is not (and cannot be made) 100%, so what looks in focus to me through the viewfinder isn't. (Sort of a human back focus problem.) I was aware of the dangers of focussing and recomposing when using a very narrow DOF, so when I was shooting with the lens I made sure that the focus point was exactly where I wanted it. I agree that it is a lovely lens. I'm glad yours is working out for you.

Thanks, everyone, for all of your help and advice. There is still a f 1.2L in my future. When I upgrade to the 5D (6D?), I plan to purchase the 85 mm f1.2L.

Virginia

claudermilk

Flyinggina wrote:

Water-boy - I would love to be able to use manual focus. Alas, I have a very, very strong lens correction in my glasses which is not (and cannot be made) 100%, so what looks in focus to me through the viewfinder isn't. (Sort of a human back focus problem.)

So use the viewfinder diopter to make the final adjustment.

kdog

Flyinggina wrote:

J... I know it sounds unlikely, but it does appear that the last three f1.2L's in the store's inventory had AF problems.

Were the boxes sealed? I'm wondering if they were returns. Maybe they started with a bunch of these lenses, and these three out of the bunch got returned for focus problems. So they keep reselling them in the hopes of eventually finding buyers for them. Cynical, I know...

-joel