Before and After; Image of a park.

[arith]

Lots of CA I can't get rid of but look at the distortion, or is it my eyes. Do I need to ditch the lens?

_IM_3912MarieDistort by Stephen Davis, on Flickr


_MG_3912Marie by Stephen Davis, on Flickr

[revi]

Looking at the screenshot, I see horizontal and vertical perspective correction, and rotation. Neither of these should be needed to correct a distortion caused by the lens you use, as these relate to camera orientation (pointing up or down, and not holding the camera level).

Tbh, I'm surprised by the need for horizontal perspective correction. Would it be possible to see the shot without it? (and perhaps without the vertical perspective and barrel/pincushion correction).

And, PSE doesn't have a correction preset for your lens? Those usually work quite well for barrel/pincushion, and might also work for CA.

[Dave Humphries]

I agree with Remco,

I see no controls on the Elements screen grab that will affect CA at all Steve.

You may not want to, but it might be time to move on to something like PS CC that will make these corrections a breeze.

[arith]

_MG_3912Raw by Stephen Davis, on Flickr

[Brev00]

You are just dealing with perspective distortion from shooting at an upward angle. It will happen with any lens. I had the very same issue recently shooting in Quebec City and Montreal. Tall buildings without much space to take a level shot. When I took the shots, I knew this was happening so I left space to allow for some correction when I got home. I use Elements as well and was able to do a fair job repairing the keystoning damage. I am sure other programs offer easier solutions. Elements does not have a ca correction feature so I use other programs to correct it (View NX 2 is the quickest and easiest--if it works). The workflow works for me. I have DxO Optics 8 and am doing a free trial of DxO Optics 11 but I don't think keystoning can be automatically corrected because it is not part of the lens profile. I am not going to keep DxO Optics 11 as, yet again, Elements finishes first for me.

[xpatUSA]

EXIF says it was shot at 17mm focal length, perhaps to get all the buildings in.

One solution (too late now of course) would have been to walk back quite a bit, holding the camera as level as possible, until the same scene + more is contained in the viewfinder but with the camera level. Then crop out the unwanted stuff in post.

That 50D has got enough pixels to allow cropping for monitor viewing or printing small, and you get perpendicular buildings free of charge!

A bonus might be that you'd be cropping out that wide-angle corner softness, too . . .

[arith]

You are just dealing with perspective distortion from shooting at an upward angle. It will happen with any lens. I had the very same issue recently shooting in Quebec City and Montreal. Tall buildings without much space to take a level shot. When I took the shots, I knew this was happening so I left space to allow for some correction when I got home. I use Elements as well and was able to do a fair job repairing the keystoning damage. I am sure other programs offer easier solutions. Elements does not have a ca correction feature so I use other programs to correct it (View NX 2 is the quickest and easiest--if it works). The workflow works for me. I have DxO Optics 8 and am doing a free trial of DxO Optics 11 but I don't think keystoning can be automatically corrected because it is not part of the lens profile. I am not going to keep DxO Optics 11 as, yet again, Elements finishes first for me.

Cheers Larry.

[arith]

Cheers Ted; I thought it was quite a lot of distortion and I did it 4 times the exact same. Pin and Barrel is normally around -10 at 17mm but here it is more than twice that; I found this image hard to do because from the water level in a previous photo the land is on an upward incline left to right, and the building is already perpendicular-ish in the middle of the image. There are no horizontal lines so I turned the building to make the top horizontal to adjust barrel/pin and then by iteration got the vertical perspective/rotation.

Normally I correct perspective in seconds.

Is it possible the IS (stability) alters the lens characteristics or is it just a very long building?

I was lost by the way, can't walk far without a rest and totally worn out when I happened to find this bench somebody had with exceptional consideration placed in the exact right spot. When I did move a million tourists turned up, even a lady who went inside an area forbidden for tourists and under the sign posed for her photographer partner.

[xpatUSA]

Cheers Ted; I thought it was quite a lot of distortion and I did it 4 times the exact same. Pin and Barrel is normally around -10 at 17mm but here it is more than twice that; I found this image hard to do because from the water level in a previous photo the land is on an upward incline left to right, and the building is already perpendicular-ish in the middle of the image. There are no horizontal lines so I turned the building to make the top horizontal to adjust barrel/pin and then by iteration got the vertical perspective/rotation.

Normally I correct perspective in seconds.

I'm no expert in distortion, 'cuz I never, ever get it with my Sigma lenses

But seriously, the Tamron appears little different to my 17-50mm at 17mm:



So, based on these samples, that amount of distortion seems unreasonable.

Is it possible the IS (stability) alters the lens characteristics or is it just a very long building?

Sorry, can't help with either question, I can only offer that the OS on my lens simply moves the image around in the focal plane but I've never noticed distortion while it is so doing.

I was lost by the way, can't walk far without a rest and totally worn out when I happened to find this bench somebody had with exceptional consideration placed in the exact right spot. When I did move a million tourists turned up, even a lady who went inside an area forbidden for tourists and under the sign posed for her photographer partner.

Dam' tourists . . .

I take it that the lens doesn't do it with similar shots made subsequently to one, i.e. the lens isn't suddenly dysfunctional?

How about printing out a grid test pattern and shooting it at home?