It is an interesting idea, it may not have anything
to do with the focal length per say but maybe just how the lens communicates with the body. As stated earlier maybe a firmware update would fix on the D2, Ds and Ds2??
i wouldn't be surprised if the high-speed AF sensors in the mkII/N coupled with the 1.25x crop markedly helps the AF accuracy on those bodies compared to the 1Ds/II. i haven't followed all 150+ pages of this thread closely enough to recall if any 20D/30D/5D users had any issues. it seems logical that this could be an isolated 1Ds/II issue.
if CF13 fixes everything then i can definitely live with it until i upgrade to a 1Ds3. limiting the 1Ds to 11 AF points doesn't hinder me in any way. besides, i kinda like the spot meter following the selected point for the times i use spot.
Nothing against the older bodies, I want a Ds2 Just something I have been pondering the last few days.I usually keep mine on CF 13-1 also, just so I dnt have to thumb through all the points when composing.
OK... I just did some test images, some of my wife, some of my daughter, some of just random objects... then I photographed the same random object (the label on my daughter's Gund ball).
The various photos of my wife and daughter were much improved (I tried both CF13 set to 1 and set to 2 (9 point).
I took shots at 1.2, 1.8, 2.2, 2.8, and 3.5. On the Gund Ball label, it was good sharpness at 1.2, excellent at 1.8, but then softened up again at 2.2, then was incrementally sharper at 2.8 and 3.5. (I only previewed them on the LCD).
I will post some samples shortly as I am in the middle of downloading files from a different card. none of this still makes sense though as the problem also happened with manual focusing the lens.
tuesday or wednesday i'll try a more controlled test at various apertures like ed did above. but, so far mine seems to get sharper and sharper the more i stop down with no focus distance/shift issues.
OK... Here is my most recent hands on test. No tripod, no mirror lock up, just shooting the way 95% of us would normally shoot handheld (I think). I used one shot AF, custom function 13 was set to 1 (11 point AF), and the subject distance was about 18" or so, the minimum focusing distance and I used center point AF. Although not all images appear equally sharp, some of that could be attributed to movement while shoot. I wish I had more time to do more controlled testing, but this lens has to ship tomorrow (unless any of you can convince me I am out of my mind).
ed, i've found that at minimum focus distance this lens does funny things. i believe that it's optimized for f/1.2~2.0 at close focus and anything smaller puts it out of design spec (or into design limitations). from about 2.5' and further, i can get any aperture to be razor sharp in the center. at infinity focus, the whole frame is razor sharp at f/5.6~8. i haven't tried anything smaller than f/8 yet.
so, you might be pushing this lens beyond its design by asking for a smaller aperture at minimum focus. if you can get f1.2~2.0 to be sharp from 1.5' away and f/1.2~4.0 sharp from 2~2.5' away then i think you're doing well. but, this is just a guess as i have no idea what this lens is optimized for at a given distance and aperture.
it would be nice if canon were as technical as zeiss and leica in their descriptions of their lenses and in providing full disclosure of a lens' capability and limitations. this would let us know how far we can push this glass before it fails and save a lot of confusion on forums at the same time.
it might be worth calling your dealer and asking for one more day of testing. you very well might be crazy -- just like the rest of us.
-erik
Edited by flatdraft on Jan 23, 2007 at 12:07 AM GMT
I think your copy is fine. It looks the most soft at 3.5, but as you said, that could be because you made a slight move during exposure. If I were you I'd keep it. I'm past my 30 return period, so if I have a serious issue with mine it'll have to be settled with Canon. It looks like you're about to return a perfectly fine lens.
I tried to duplicate backfocusing with my 50mm f/1.2L as described by flatdraft (Erik) using close working distances. I was able to get the lens to backfocus with a 18 inch (46cm) working distance using both a Jackson 2004 FocusTestChart and a USAF 1951 chart. However, the lens autofocuses perfectly at longer working distances tested (129cm and 258cm).
I tested two other very reliable lenses that I know focus normally in most circumstance and have produced beautiful images for me. Using my EF 50mm f/1.4 at minimum focus distance with a Jackson 2004 FocusTestChart it showed front focusing in the test. With my EF 85mm f/1.8 lens, the Jackson 2004 FocusTestChart indicated it was backfocusing at minimum working distance.
Tests with the FocusTestChart at very short working distances tell you something about how a lens autofocuses under very close focus conditions. The results in the three lenses I tested tell me the FocusTestChart tells me nothing useful about camera-lens systems for most of the photographs I take at a longer working distance.
Edited by wcastleman on Jan 23, 2007 at 07:54 AM GMT
william, thanks a ton for the extensive test. this goes to show that the 50L is only as accurate as the mounted camera's AF system and that it's difficult to use at minimum focus distance for both the camera and the user.
given these results, i'd say there are less bad copies out there than originally thought. AF system error and user error have a lot to do with the outcome.
I just put the label on the box, and it'll be too much work to open it all back up and confuse myself some more. While there is a remote possibility this lens is OK, I've got way too many backfocused images (when using 45 point AF) to convince me otherwise.
Gary can certainly vouch as he's seen more shots of my family than I could have ever imagined ever sharing with a forum friend.
William, is this your third copy? I am interested to find out how it compares to your first two copies, particularly with regard to the sharpness falloff you experienced on your initial copies.
So bottom line is the lens will backfocus at working distances under 18", but will be fine beyond that when using the 1D and 1Ds bodies. I'd be interested to see the results using a 5D.
lord_malone..the lens minimum distance for focus is 18" approx...so there is no focus under that anyway. My second copy backfocused to subjects of almost 10 feet, but it may have just been a bad copy period...
Noted, Gary. Thanks. My my real world test shots are still not reflecting what I saw on the lens focus chart. And I haven't missed focus when aiming at an intended target. Oh well. I've seen stranger things. I stand by my assessment that my lens is good, but would still like to have it tested on a different body for a piece of mind.
eeprete wrote:
ed malone, did you ever wind up trying the lens with your 1D using the various AF points and settings to see how it was in comparison to the MK2n?
I did. And the results of the 1D focus test indicated that it was backfocusing. However, it performed problem free much like the 1D2N when shooting random objects around the house. AF seemed spot on.
In response to eeprete and lord_malone, I have used two EF 50mm f/1.2L lenses (code UU1000) on three bodies (5D, 1Ds mkII, ID mkII), and they have all focused fine at routine types of working distances. I have tested three lenses on two bodies with the Jackson 2004 FocusTestChart at close to minimum focus distance and aperture at f/4 to f/2. In every instance, the chart is telling me the camera_lens are either front focusing or back focusing in autofocus mode. The focus is fine if I focus via manual focus at these tight working distances.
The message I'm getting from this is: 1) When using a lens at close to minimum focusing distance, it is best to manual focus and/or stop the lens down to f/8 to f/16 if you want the sharpest results; and 2) the Jackson 2004 FocusTestChart at close to minimum focus working distance gives you false-positive misfocus results with your Canon lens_camera system.
I don't doubt that there are some miscalibrated lenses out there. I just think that using Jackson 2004 FocusTestChart at close to minimum focus working distance is going to detect more "bad" lenses and cameras than there really are.
For me, It's time to stop peeping at pixels and go take some photos with the EF 50mm f/1.2L.
Just something I have been pondering the last few days.I usually keep mine on CF 13-1 also, just so I dnt have to thumb through all the points when composing.
