2nd copy of 17-40mm F4L, edge to edge inconsistency.

[gcams]

Hi folks,

Well this is my second copy of the 17-40mm F4L after returning the first one I had (thread http://photography-on-the.net/forum/...d.php?t=115418 refers). The new lens is much better than the last one I had and centre sharpness is excellent. In fact, so is edge sharpness, but only on one side. I did some tests this morning outdoors and started to notice that the right hand side of the frame is always softer than the left. This is particularly noticeable when shooting close up.

I did some newspaper tests under daylight and these are the results (all shot off tripod at F4 with mirror lockup, 100% crops):

Top Left:
img_3199_TL.jpg

Top Right:
img_3199_TR.jpg

I know the consistency of lighting across this shot is not ideal, but I have a number of other examples under different lighting conditions. Regardless of the situation, the results are the left is always sharp and the right is always soft.

My real question is, bearing in mind this is on a 1.6x crop factor camera (350d), is this inconsistency acceptible or should I be sending back my second copy and trying for a third??

I guess I'm a little hesitant to return this copy also, as at least it is sharper than the first, although the right hand edge of frame is getting quite soft. I'm not sure how it would look on a full frame camera which I hope to one day own.

Thoughts, comments?? Am I being too picky??

[gcams]

Here's another example at F4 with more consistent lighting across the whole page:

Left Hand:
img_3220_LH.jpg

Right Hand:
img_3220_RH.jpg

[BoySpot]

You really sparked my interest with this post. I have been stitching some panos taken with my 17-40 at the 40mm end. Using PTAssembler, you get a direct comparison between the sides (and top and bottom) of the lens on the same subject when looking for the control points. I definitely notice it is sharper on one side than the other and I think it is the same way around as yours from memory. I was wondering whether the filter was a factor but now you have got me thinking.

[gcams]

Hi BoySpot,

It certainly could be that you're experiencing something similar. I noticed the consistency between the left and right edge of frame becomes much better once the lens is stopped down a bit. Here's an example at F5.6 for example:

Top Left Hand Side:
img_3221_TLH_56.jpg

Top Right Hand Side:
img_3221_TRH_56.jpg

Also this is at the 17mm end. Again at the 40mm end the effect was much less (if at all noticeable), however I could imagine doing pano stitches you would really notice it even if it was marginal.

I also came across this old thread about a guy who was having the same problem (although in his case it was the left that was soft and the right that was OK).

It's hard to know what to do. Do I sent this lens back and take the risk I will get another possible worse copy (at least this one is sharp where it is sharp), or live with the fact this one is soft on one side at F4/17mm. Maybe doing these lens tests is a bad idea!!

[Jon]

First - are you absolutely positive you're set up with the camera square to the wall? Indoors and at wide apertures a slight angle can throw one edge outside the DoF.

[gcams]

Originally Posted by Jon

First - are you absolutely positive you're set up with the camera square to the wall? Indoors and at wide apertures a slight angle can throw one edge outside the DoF.

Hi Jon,

Yes indeed, been down that road unfortunately. I have tested this in about 5 different scenarios now (both indoors and out), and have even tried deliberately skewing the camera from parallel in either direction to see if that makes any difference. Regardless, the right hand side of frame is always softer, no matter what I try.

The first time I noticed it, I thought it must be my setup, so I repeated the tests (again and again). In every case and scenario, it's always the RHS that shows up soft. If it was my test setup, I would expect it to switch sides and change in severity depending on the scenario but it's just too consistent, indoors/outdoors, lit from the left/right, etc, makes no difference. The only way I have found to get the RHS close to the LHS is by stopping down, as shown above.

[Jon]

Have you tested the kit lens at the same time, just as a standard? Might be worth seeing if your camera might have something out of line. It does seem strange that it's happening on two random examples of the lens, which is why I'd try the other lens.

[gcams]

Yes I did try the kit lens on several of the scenarios above. The kit lens was soft at both edges () but consistently soft at both edges, unlike the F4L. Below are samples from the kit lens, F4 at 18mm (100% crops as usual):

Kit Lens Top Left Hand:
img_3203_kit_LH.jpg

Kit Lens Top right hand:
img_3203_kit_RH.jpg

By the way, in case I didn't make this apparent, my first copy of this lens was just plain soft all round. This copy is much better but I'm just a tad concerned about the edge to edge consistency.

[gcams]

Bump.. any other 17-40mm users wish to comment? Anybody else seen this with their lenses?

[MDJAK]

I own this lens and have been very happy with it, until reading this thread and taking a group shot this weekend. The right side of the photo was considerably softer than the left, though I chalked it up to not having the camera square to the group I was shooting. Now, I'm not so sure.

http://i.pbase.com/o4/12/533712/1/53440323._I1A5291.JPG

The little boy on the right is considerably softer than the others, though I know I wasn't square and it's possible he moved.

One more tidbit: My camera is FF so it shows corner softness more readily.

Should have gotten the 16-35 f2.8.

[Jon]

Originally Posted by MDJAK

I own this lens and have been very happy with it, until reading this thread and taking a group shot this weekend. The right side of the photo was considerably softer than the left, though I chalked it up to not having the camera square to the group I was shooting. Now, I'm not so sure.

http://i.pbase.com/o4/12/533712/1/53440323._I1A5291.JPG

The little boy on the right is considerably softer than the others, though I know I wasn't square and it's possible he moved.

One more tidbit: My camera is FF so it shows corner softness more readily.

Should have gotten the 16-35 f2.8.

Wasn't that the lens that accompanied your 1Ds II on its jaunt on the Sidewalks of New York though?

[MDJAK]

Shhhh. You don't have to tell the whole world, do you?

I may want to sell it one day.

Yes, in fact it was, but I truly don't believe it suffered in any way from the fall. May seem hard to believe, I know, but I see no difference in it from before. It was always a great lens.

BTW, I didn't even have the lens hood attached at the time of the fall.

[Jon]

Then I'd be inclined to look at some earlier shots and see if they show the same thing, is all.

[gcams]

Hi Mdjak,

I hope I don't end up being responsible for you becoming discontent with your 17-40mm F4L Your observations are interesting though. Have you tried the old newspaper test like I did above? It seems to be very good at showing up edge inconsistency. Just a thought!

[pfogle]

Just for the record, I tried a s/h 16-35/2.8L and returned it - same problem as described but much worse! Then the same thing with a second one. After that I got a 17-40/4L and it's absolutely fine - not as sharp as some samples I've seen wide open, but much better than the other two, and it is consistent.

From the samples, I'd say return it and keep going 'till you get a good one - they are out there!