[Guest]

Hi there,



I ended up with a new Canon 15-85mm that is slightly decentered at 85mm wide open. Other focal lengths seem OK to me.



Below are two 100% crops taken with a 10 megapixel camera. The pictures show the 4 corners copied closer together for better comparison.



http://im.bilderkiste.org/1135246865645/test1.jpg

http://im.bilderkiste.org/2135246868663/test2.jpg



The photos are representative of the large number of shots I took. test1.jpg was composed of 4 shots, first focused at infinity with center AF and then recomposed each time to put the tower into one corner. The test chart setup (test2.jpg) was perpendicular and I used AF at the center.



I went through 3 copies of the lens and they were all the same (more or less).



Since the lens test on this site also mentions decentering I would like to ask: is this amount of decentering inevitable for this lens? Would you keep it or hunt for a better copy?



Thanks



Tom

[Rainer]

Hi Tom,



The slight decentering that you show in your images will quite likely

go away if you stop down to f/8 or f/11. I've seen images of the 15-85

showing much worse decentering (but typically more on the 15mm end.



So, if yours is ok at the 15mm end, especially, if the vignetting

shown in the corners of a wide open 15mm image is evenly

distributed, I would certainly keep the sample you have.



Just my thoughts .... Rainer

[Guest]

[quote name='Rainer' timestamp='1352545464' post='20913']

Hi Tom,



The slight decentering that you show in your images will quite likely

go away if you stop down to f/8 or f/11. I've seen images of the 15-85

showing much worse decentering (but typically more on the 15mm end.



So, if yours is ok at the 15mm end, especially, if the vignetting

shown in the corners of a wide open 15mm image is evenly

distributed, I would certainly keep the sample you have.



Just my thoughts .... Rainer

[/quote]

I tend to agree.

The corners on my lens show equal vignetting.

The blur shows only at 85mm and only in the leftmost 10%-15% of the frame.



On an interesting note, I found out this: When shooting upside down and also viewing the picture like that, the blur is still on the left hand side (instead of showing on the right). This is with IS on (I shot the test chart with IS off).



BTW the IS is really great. I got tack sharp images at 1/15s at 70mm hand-held :-)

But everything seems to have its price.

Maybe there is a reason the new 24-70L II still does not come with IS...

[Guest]

[quote name='tomri' timestamp='1352552921' post='20915']

I tend to agree.

The corners on my lens show equal vignetting.

The blur shows only at 85mm and only in the leftmost 10%-15% of the frame.



On an interesting note, I found out this: When shooting upside down and also viewing the picture like that, the blur is still on the left hand side (instead of showing on the right). This is with IS on (I shot the test chart with IS off).



BTW the IS is really great. I got tack sharp images at 1/15s at 70mm hand-held :-)

But everything seems to have its price.

Maybe there is a reason the new 24-70L II still does not come with IS...

[/quote]



If shooting upside down does not flip the softness, then the error could lie outside the lens. Did you check if this softness is due to slight difference in focal plane, i.e. can you focus left and right separately for optimal focus. Perhaps the lens focal plane is not completely parallel to the sensor plane, possibly the lens axis is not completely perpendicular to the sensor due to the tolerances in the the camera body (lens mount, sensor...)

[Guest]

[quote name='photonius' timestamp='1352630696' post='20917']

If shooting upside down does not flip the softness, then the error could lie outside the lens. Did you check if this softness is due to slight difference in focal plane, i.e. can you focus left and right separately for optimal focus. Perhaps the lens focal plane is not completely parallel to the sensor plane, possibly the lens axis is not completely perpendicular to the sensor due to the tolerances in the the camera body (lens mount, sensor...)

[/quote]

Well, I did that upside down test quickly and hand-held based on things I read on the net.

Maybe the result should not be counted as reliable/accurate.

There are just too many things that could go wrong that way.