EFS 17-55 USM corner blurriness: experts' opinion needed !

I have a 70d with a Canon EFS 17-55 f/2.8 USM.

After a few months of use, I noticed that the bottom right corner of my image was unsharp in some situations. The issue was clearly visible at 55 mm f/2.8, less so at 17 mm f/2.8. At f/5.6 it was still visible, and at f/8 almost gone (or at least reduced to an acceptable level).

In addition I had the feeling that the AF was inaccurate, so I sent the body and the lens to Canon for servicing.

I got it back and now:
- The blurry corner is still there.
- The MFA has to be set to W:+12 T:+15 to get correct focus (instead of W:+6 T:+9 previously)
- The AF seems to be more accurate, i.e. focusing on the same target several times produces less focus variations.

I called the technician, who told me that he adjusted the body AF, but did nothing to the lens, as he tested it and concluded that it was "within Canon specifications". I had also a long discussion with him about the MFA value being higher than originally and thus all modes but M, P, Av, Tv and B being unusable, but this would deserve another topic. My main concern was and remains: corner blurriness.

My understanding of the situation is that the lens has an optical problem (a centering defect ?), however Canon will not fix it because it is within their manufacturing tolerances.

So my question is: are Canon testing criteria too low or my expectations too high ?

Please give me your opinion.

Here is an example illustrating the problem:

The orginal image is here: https://www.cloud.sfr.fr/?shareObject=489d7ce6-a84a-14e1-ad70-9e2f130135ce

In my experience, Canon will only do things that are easy and fast - like electronic adjustments, etc. That is why I am trying to catch problems withing the return period.

Yes, this is decentering. I would try to convince them to send me a replacement. I owned that lens, the corners were not great but nothing like what you showed.

Do brick wall test on tripod with IS off.  Shoot, then flip 180 and shoot again.  Compare corners.  Right angle orientation to the wall critical.

70d IF wrote:

I have a 70d with a Canon EFS 17-55 f/2.8 USM.

After a few months of use, I noticed that the bottom right corner of my image was unsharp in some situations. The issue was clearly visible at 55 mm f/2.8, less so at 17 mm f/2.8. At f/5.6 it was still visible, and at f/8 almost gone (or at least reduced to an acceptable level).

In addition I had the feeling that the AF was inaccurate, so I sent the body and the lens to Canon for servicing.

I got it back and now:
- The blurry corner is still there.
- The MFA has to be set to W:+12 T:+15 to get correct focus (instead of W:+6 T:+9 previously)
- The AF seems to be more accurate, i.e. focusing on the same target several times produces less focus variations.

I called the technician, who told me that he adjusted the body AF, but did nothing to the lens, as he tested it and concluded that it was "within Canon specifications".

that in comparison to the 24-105, this lens has three sharp edges vs. none. "What are you complaining about." But seriously...

He could also mean that the decentering here is not severe enough to mess with. To center it. you may end up with both sides noticeable soft although the right side will show improvement.

My first 24-70 II was soft 1/3 of the way into the frame from the right side. This is a decentering issue they can't ignore.

I had also a long discussion with him about the MFA value being higher than originally and thus all modes but M, P, Av, Tv and B being unusable, but this would deserve another topic. My main concern was and remains: corner blurriness.

My understanding of the situation is that the lens has an optical problem (a centering defect ?), however Canon will not fix it because it is within their manufacturing tolerances.

So my question is: are Canon testing criteria too low or my expectations too high ?

Please give me your opinion.

Here is an example illustrating the problem:

The orginal image is here: https://www.cloud.sfr.fr/?shareObject=489d7ce6-a84a-14e1-ad70-9e2f130135ce

Rick Knepper wrote:

that in comparison to the 24-105, this lens has three sharp edges vs. none.

It does not.

Just another Canon shooter wrote:

In my experience, Canon will only do things that are easy and fast - like electronic adjustments, etc. That is why I am trying to catch problems withing the return period.

Yes, this is decentering. I would try to convince them to send me a replacement. I owned that lens, the corners were not great but nothing like what you showed.

Actually, this is already my second copy of this lens. The first one was worst, and it was on the same corner. I returned it to the seller a few days after my purchase and I got a replacement unit. Although I noticed that the second copy had a sligth corner blurriness at 17 mm, it was much better than the first one. Unfortunately, I did my tests at 17 mm only. Later on, I noticed the problem at 55 mm, but the return period was over. By the way, I don't know if the seller would have accepted another return.

MAC wrote:

Do brick wall test on tripod with IS off. Shoot, then flip 180 and shoot again. Compare corners. Right angle orientation to the wall critical.

I have no brick wall at hand right now, but I already did this test, with the same kind of result as the hedge shown here.

Rick Knepper wrote:

He could also mean that the decentering here is not severe enough to mess with. To center it. you may end up with both sides noticeable soft although the right side will show improvement.

This is exactly the reason why I am asking for opinions - and experiences - here. Sometimes the best is enemy of the good.

70d IF wrote:

MAC wrote:

Do brick wall test on tripod with IS off. Shoot, then flip 180 and shoot again. Compare corners. Right angle orientation to the wall critical.

I have no brick wall at hand right now, but I already did this test, with the same kind of result as the hedge shown here.

your dof is as thin as it gets with that lens f2.8 and 55, I can't tell if you are perpendicular or not

those settings might be good for a portrait where the corner makes no diff

I'd like to see the brick shots and 180 flip which will tell us the data we need

MAC wrote:

70d IF wrote:

MAC wrote:

Do brick wall test on tripod with IS off. Shoot, then flip 180 and shoot again. Compare corners. Right angle orientation to the wall critical.

I have no brick wall at hand right now, but I already did this test, with the same kind of result as the hedge shown here.

your dof is as thin as it gets with that lens f2.8 and 55, I can't tell if you are perpendicular or not

those settings might be good for a portrait where the corner makes no diff

I'd like to see the brick shots and 180 flip which will tell us the data we need

I promise that I was perpendicular ! The dof was not so thin, because the hedge was at a distance of approximately 25 feet.

Anyway, I will try to do your test, but please be patient, I won't be able to do it until next week-end.

70d IF wrote:

MAC wrote:

70d IF wrote:

MAC wrote:

Do brick wall test on tripod with IS off. Shoot, then flip 180 and shoot again. Compare corners. Right angle orientation to the wall critical.

I have no brick wall at hand right now, but I already did this test, with the same kind of result as the hedge shown here.

your dof is as thin as it gets with that lens f2.8 and 55, I can't tell if you are perpendicular or not

those settings might be good for a portrait where the corner makes no diff

I'd like to see the brick shots and 180 flip which will tell us the data we need

I promise that I was perpendicular ! The dof was not so thin, because the hedge was at a distance of approximately 25 feet.

Anyway, I will try to do your test, but please be patient, I won't be able to do it until next week-end.

not a problem to wait - it's the 180 flip that confirms the data.

clean the lens real good - both ends - before testing

in future, always do the brick testing under the return period.

I think that everyone at the school where my favorite brick wall resides think I'm positively crackers. I stand there for 20 minutes, half an hour or more with my camera on a tripod pointing at a wall. Meticulously taking shots and changing lenses - they must wonder what sort of pictures I wind up with.

Regardless, I take shots with every lens I get. Normally it's just going through the motions as I've never yet suspected one of my lenses of having decentering issues. If I did, I'd go back and rotate the camera as MAC mentions.

And, a hedge works well too. The brick wall can help show you the distortion, and unevenness of distortion for example, as well as giving you a grid to guide where softness starts. The brick wall also is most useful if the decentering is subtle - I don't think we're talking subtle here.A hedge is also good at showing detail, color, light and shadow and how a lens handles it. As you can see below - I use one in my backyard frequently. I'm about to go out there and try out my new Tamron 150-600 against my 70-300.

Honestly, I don't have any lenses with soft corners - or at least one corner that's noticeably softer than the others. Usually a quick set of snaps on the brick wall confirm there's no problem, and I've not gone out to it suspecting (knock on wood).

The 17-55 is an excellent lens for a reason - but Roger Cicila claims that it's also quite fragile. When it comes to lenses and lens issues - I'll call him an expert.

http://www.lensrentals.com/blog/2009/05/lens-repair-data-3-0

Anyway, I'd return it and hope for a good copy. It can happen.

I had 4 bad copies of the Tamron 70-300 VC. From non-operational, VC problems to front focus problems. I waited a year and a half, and then got a good copy my first try. When I got my 6D with 24-105 kit, I got an incredibly soft 24-105. But it wasn't so soft that I thought anything was out of the ordinary - especially when most people here said to MFA the thing, or that the Sigma I was using to compare (my only other FF lens I could use at the time) was a lot sharper. I finally figured, if it stank on my 60D, I couldn't MFA that, so it went back. The new one was so much better it wasn't even funny. With MFA it went from good to borderline excellent.

So don't feel bad at returning it again - especially based on your example. It happens, and it's your money.

24-105

24-70

16-35/4

17-40

MayaTlab0 wrote:

Regarding your lens, it seems to me that it stays decentered quite a while when moving towards to frame centre, but I wouldn't be able to be sure of it.

This also my opinion, because I have also non-symetric purpule fringing : it is much more present in the half right as you can see on my full res image.

rebel99 wrote:

a friend of mine owned a copy of 17-55 f2.8 tele and i got to handle it for a little bit. i noticed that the optic of this tele was almost "L" quality but its body and overall construction integrity was very poor. so, i honestly cannot recommend this lens to anyone i know its price is almost the same as a 70-200 f4.0IS or a 16-35 f4IS, which are outstanding in every aspect.

cheerz.

Of course there is always something better. But quality has a price : the 16-35 is 45% more expensive, it is also one stop darker, and has 36% less telephoto range. Of course, I would get a better image quality, and I guess that I would be more than happy regarding extreme corner sharpness with a full frame lens on an APS-C body...

I knew that the 17-55 construction was not so great, but my budget was also limited (I bought a 70d, a 17-55, a 480EX flash and a few accessories at the same time) and while I can accept this lens to have a shorter lifetime, I am also expecting it to work properly out of the factory.

I didn't see any other good quality (optically speaking at least), sub 1000€, stabilized, APS-C, low light (minimum f/2.8) and general purpose (wide angle to moderate telephoto) lens in the Canon lineup.

When Canon releases a 15-50 f/2.8 IS STM, I will call my banker

Let me give you my thoughts about this : the manufacturers are perfectly aware of the problem. However, we, customers, want the lowest possible cost. Whereas the manufacturers want to highest possible margin. To solve that equation, they do the following : they increase their manufacturing tolerances, thus they release lower quality products on the market. A few customers will complain and they will repair their equipment. The whole operation will cost them much less than if they would have fine tuned every lens manufactured...

70d IF wrote:

Rick Knepper wrote:

He could also mean that the decentering here is not severe enough to mess with. To center it. you may end up with both sides noticeable soft although the right side will show improvement.

This is exactly the reason why I am asking for opinions - and experiences - here. Sometimes the best is enemy of the good.

Can you call back and ask for an explanation of "within specification" relative to this repair?

Just another Canon shooter wrote:

Rick Knepper wrote:

that in comparison to the 24-105, this lens has three sharp edges vs. none.

It does not.

You saw this -  - didn't you?