Not sure if I'm not picky enough... or if I'm one of the lucky ones..
I seem some small signs that contrast is not 'great' when wide open and focused near MFD, but in general from what I've seen so far I'm fairly happy...
These are all taken hand held IS on, 200mm at or very close to minimum focus distance.
First a resized version, then a 100% crop from near the focus point, processed using DPP, zero sharpening, all other settings at 0....
This here is more like the kind of shot I really want to be able to take anyway... 160mm f4, somewhere close to MFD, 1/25th second... if my daughter stays still enough I can get them... If I ever get some real light, should be even better, to get I had to go to ISO1600 to get 1/25th second for this.
DyNoMohum, that's not looking sharp to me... at least, that level of sharpness is also what I am seeing under the same conditions. It is no were near as sharp as it is at 135mm MFD f/4. Maybe you can shoot an object from an angle (3D effect). That way it is easier to detect if there's some kind of backfocus going on.
mfurman, thanks for your reply. I don't know if it means anything, but it would be interesting to see if owners of a 1D series camera are having the same problems. I use a 5D.
Now, here are 100% crops of the results of last nights test (no PP, sharpness off, tripod, IS on, cable release):
REMOVED LINK
I focused on the HP logo. You can clearly see that beyond 135mm FL the focus is moving backwards... At 200mm f/8 there is something in very sharp focus... But it ain't the HP logo for sure.
BTW Ken Rockwell also mentioned a backfocus at f/4 at 200mm MFD in his review of the 70-200L 4IS. By f/5.6 he says it is gone, but that's probably the DOF hiding the backfocus.
Edited by Daan B on Feb 28, 2008 at 03:48 PM GMT (Reason: removed link)
Thank you Daan for creating this thread, besides Mfurman and Jammy Straub (and about 2 other people, sorry I have forgotten your names )on this forum, not many others have seemed to report this in the past and it makes me think I'm going a bit insane. Intially I thought I had a "soft" copy because of this issue, then one day I purchased a 40D which made it focus better, and went to the zoo and took some shots of far away things and WOW... (For some reason it is really crap on my 400D, which works fine with my other lenses)
Also sorry Dynomohum your 100% crops at MFD 200mm are VERY soft, just like mine...actually exactly like my lens
Firstly, demonstration of this lens's resolving power e.g. This is a 100% crop showing typical sharpness (out of camera JPG) (not the sharpest I have ever got, but a typical "sharp" shot): http://farm3.static.flickr.com/2061/2184735891_c873bcf419_o.jpg
f/8 1/100 200mm 100% centre of frame crop
__________________________________________
Ok, now onto the discussed issue, the two below shots aren't very sharp examples of what I am demonstrating, but it is the disparity that is the point I am trying to make. Yes, this lens is capable of better sharpness under both scenarios below, but this is the "typical" difference in daily use.
Here is sharpness at 100mm MFD (again, not the best sharpness I can muster, but a typical result hand holding in not too good lighting) - note thin dof: http://dawei.zenfolio.com/img/v3/p183174585.jpg
f/5 1/500 100mm 100% crop
Now, this is at f/9 MFD 200mm. Admittedly it is NOT the absolute BEST result at 200mm MFD I can muster (due to 1/200 speed and hand holding and in bad lighting etc.), BUT it is a typical result, and the disparity between this and the second shot is typical. (Yes I know, razor thin DOF blah blah etc..but this happens even when I take about 100 frames on tripod refocusing each time) http://dawei.zenfolio.com/img/v3/p21140687.jpg I'm editing this message to again emphasize that the two previous shots are not the best sharpness this lens can muster at the given settings, it is the DIFERENCE I am trying to make a point about. The 100mm MFD shot is actually quite soft for this lens, and so is the 200mm MFD shot, but sharpness disparity between the two is similar in my better, sharper shots. It's just that I was so excited to post this I didn't spend much time looking for better, sharper shots but I think my crops make my point though.
I will post up some more controlled examples if anyone is interested, but I think other posters have made the point well enough
RE: Ken Rockwell comments, my copy also back focuses at 200mm I agree.
I have only used mine for about a hour. All my shots were taken in poor lighting (which I guess is where you would use IS)... My camera is a Canon 40D, which some say has a very strong AA filter. I have taken less then 100 photos so far, and really have not compared my MFD to anything further away, etc... My previous 70-210mm f3.5-4.5 USM is like 15 years old, so to some degree what I have to compare to is a very old lens...
Also... my shots are with absolutely ZERO sharpening, not even the small amount that all in camera JPEGs would get. All at F4 as well.
I do know and/or have evidence that the contrast improves when I increase the aperture, but as of yet, I haven't had a chance to take any photos where i had enough light to actually increase the aperture... So, my testing is still a work in progress...
One question... Is there ANY better alternative to this lens? From what I've been lead to believe, all of the Canon 70-200mm lenses suffer from this sort of thing to one degree or another, is this not true? By 'alternative' I mean another zoom in this range... I'm not married to IS.... and would consider a f2.8 Canon... but again, do all Canon 70-200mm suffer this same issue?
Dawei Ye, is your lens calibrated? If not, maybe it is worth a try.
Also the difference between your 400D and 40D could be because of the difference in the AF itself, but it could also be that your 400D wasn't calibrated properly (to match the 70-200L).
The thing with these f/4 lenses (as opposed to f/1.4 or f/2 lenses) is that the large DOF easily hides a possible backfocus. The point where a possible backfocus would be most visible, is at 200mm f/4 at MFD, because there the DOF is at it's thinnest. But of course other FL at f/4 at MFD could also show the problem. As you can see in my test sequence, the backfocus starts to arise at and beyond 165mm at MFD. f/4 @ 135mm @ MFD is sharp but could be sharper IMO. But then again, when you look at a backfocused image, where it is IN focus, it is very sharp.
Since there aren't any users with MFD problems yet (in this thread) that have had their lens calibrated, it is unknown wether calibration will eliminate the entire MFD problem. But I am convinced that all the MFD problems are related to some kind of front or backfocus issues. Which in theory can easily be solved by calibration. In everyday shooting these issues remain mostly undetected because the f/4 DOF will cover things up. It would also explain why some users have issues and others have not (even with the non-IS version).
But I could be wrong: maybe it is just a design issue...
DynoMoHum, you could do some new tests in better light and by using a tripod. But if the issue remains, I would let Canon calibrate your copy to your body before giving up on this lens... Can't hurt to try...
I do not believe these issues are totally related to mis-focus. It's easy to see that there is a loss of contrast, you can also see that even at the point of actual focus, the contrast is not what it is when focusing at further distances. I haven't done it yet, but it should be easy with live view to get around any issues of AF mis-focus and then verify that the contrast is NOT as good, regardless of how accurate the focus is or isn't.
DynoMoHum wrote:
I do not believe these issues are totally related to mis-focus. It's easy to see that there is a loss of contrast, you can also see that even at the point of actual focus, the contrast is not what it is when focusing at further distances. I haven't done it yet, but it should be easy with live view to get around any issues of AF mis-focus and then verify that the contrast is NOT as good, regardless of how accurate the focus is or isn't.
It could be... But the loss of contrast could also be a side effect of a possible backfocus. When shooting 3D subjects at 200mm MFD I can see a backfocus very clearly. Where it is in focus, I don't see a loss of contrast (besides the normal loss of contrast when shooting wide open).
I haven't had mine calibrated, but you could be onto something, I have noticed better AF performance stopped down and I usually shoot at 5.6
The thing is my copy is so sharp otherwise, that I don't want Canon fudging it up if I send it in. I may have to because my 40D is going in eventually (I've been saying this for 3 months) to get its issues sorted out, so we'll see how it is afterwards. I'm tempted to throw all my Canon gear in, but then I don't want them coming back misfocusing with my Tamron 17-50
Over on POTN, a poster mentioned that the MFD issue only affects a handful of units and is not within spec, but reading Mfurman's links it looks like there are many users of this lens who find this issue.
Actually after searching and reading through so much, I still have NOT found 1 acceptably sharp pic from anyone at MFD and 200mm. Unfortunately Jammy's thread's pics are offline now
I've done more testing, I'm pretty convinced my camera does have this issue. However, I'm also pretty convinced that it's a design issue and that it's probably pretty much like that for all of these lenses. I have no way to know that for sure but that's what my gut tells me. I haven't decided yet if this issue is enough to throw the baby out with the bath water so to speak... The lens as a whole seems to be quite nice, and if this is it's only flaw, then it's still overall a very good lens. I also do not know if there is any other zoom of this focal range that is any better. Now, if this issue is something that can be fixed or it is possible to get a copy that does not have this issue, then perhaps the lens would be more or less perfect. I personally would hate to think that my lens is less then it should be, that is... if it can be made to work better, then I would like to have it work better. If this is as good as any of them ever get... I'm not sure if I will be willing to give up on all the other good things that come with it...
Here are crops from two different shots taken at 200mm f4 MFD, the first was a attempt to use autofocus, the second I used my Canon 40D's Liveview and 10x zoom to manually focus. In the autofocused shot, not a single point in the photo of my work cube's wall was in focus, in the manually focused shot, about half was the other half was not, this being due to the fact that the camera's sensor was not perfectly parallel to the wall(user error, not the camera's fault). The second crop is clearly much more in focus then the other... The point of this is to show that if the lens does get focused correctly at MFD, it can be reasonably sharp... (perhaps not perfectly sharp, but reasonable none the less)... The problem... AF doesn't seem to be able to focus that well at MFD and 200mm.
I should also add... that assuming all of this is related to SA, it's probably virtually impossible for the AF system to correctly AF given the SA problem.
Thanks DynoMoHum for posting these pics. It clearly shows that (reasonably) sharp pics can be made at MFD 200mm. This is what I've experienced also. The fact that you can achieve this by manually focusing instead of AF points towards a possible AF problem. My first thought then would be that there is a back/front focus issue. But if this is a non correctable design issue (focus shift/RSA) or something that can simply be solved by calibration is not yet clear to me. Like I said, nobody with MFD problems had their lens calibrated yet (at least who participates in this thread up till now). I hope to get some answers after my visit to Canon Service coming Tuesday
Of course it does not help that every user has a subjective opinion on what is sharp or not. Maybe, the level of sharpness that you've achieved by focusin manually is within specs.
If the MFD problem isn't solvable, I will keep the lens anyway. I only came across this problem by doing test shots on flat surfaces. In real world shooting, doing portraits and other distant stuff, it performs excellent. It is very sharp, has nice contrast, neutral bokeh, killer IS, etc... So, I agree with you, if this MFD flaw is it's only weakness... well, there are worse things in life
I am done with it for now... It's time to enjoy this lens
This thread reminded me of the 50L focus shift discussions. In the case of the 50L the plane of focus shifts backward up to a couple cm when shooting close in and stopped down a few stops (my copy is worse at about 3 ft and f/2.8). The sharpness of the 50L is just fine in this case it’s just that the actual focal plane is behind the intended focal plane. Popular belief is that uncorrected spherical aberration is the cause and that this was deliberate in the design to increase the quality of the bokeh.
It seems as if something similar is happening here. I don't think the MFD is shifting between 135 and 200mm but that perhaps uncorrected spherical aberration is degrading the image quality close in at the upper end of the zoom range. At least with my copy the lens focuses fine at the MFD at all FL but as with others the IQ is reduced at 200mm compared to 135mm. Thus I guess one could say that the focus distance needs to be greater at 200mm to get the same sharpness as at 135mm.
I didn't test this directly but with my lens the sharpness at 200mm is fairly close to that at 135mm by 6 ft and a bit better than that at 10 ft. I did do a quick test for posting here at MFD to show that the IQ is reduced at 200mm compared to 135mm. Here is my set up (upper region in center is a flat focus target with a lower angled region to show the DOF):
Here are two 100% crops at about the MFD, both sharpened to better show the IQ differences. They show that IQ is indeed degraded at 200mm at MFD compared to 135mm. These were manually focused and with the standard testing precautions. A link to the unaltered original (other than conversion from raw with 0 sharpening) is below each crop.
I did do a few additional shots with this setup to see if the IQ at 200mm improved as the focal plane was moved backward. Within the limits of the FOV for this test (just a backward shift of several inches) I didn't notice any difference in the image quality. Based on my observations above I would say that the improvement in IQ at 200mm occurs gradually from MFD through 6ft with some increase still beyond this.
)on this forum, not many others have seemed to report this in the past and it makes me think I'm going a bit insane. Intially I thought I had a "soft" copy because of this issue, then one day I purchased a 40D which made it focus better, and went to the zoo and took some shots of far away things and WOW... (For some reason it is really crap on my 400D, which works fine with my other lenses)
but I think my crops make my point though.