17mm TS-E Barrel Distortion

[CDD28]

Has anyone had a problem with barrel distortion on the 17 TS-E?

I've been using it for architectural photography and noticed some significant barrel distortion. I know these lenses are supposed to eliminate distortion and was wondering if maybe I have a bad copy.

 

[privatebydesign]

Has anyone had a problem with barrel distortion on the 17 TS-E?

I've been using it for architectural photography and noticed some significant barrel distortion. I know these lenses are supposed to eliminate distortion and was wondering if maybe I have a bad copy.

Mine has zero. Can you post an example?

 

[Loswr]

Photozone lists it at ~1% barrel. I don't really notice it on mine. Second the request for samples.

 

[RobertG.]

Without shift I didn't notice any barrel distortion. With maximum shift of 12mm barrel distortion is easily noticable as well as a lack of sharpness and resolution. Of course there is strong vignetting too.

See the attached out of cam JPEG as an example. The file was just reduced to 800 x 55 pixel. Full 12mm of shift was used and the lens was shifted to the right. It was a test shot, after my panorama shot was done.

The final picture taken with the TS-E 24mm in portrait orientation and later on stitched looks like this:

All shots were with the TS-E 17mm (attachment) and the TS-E 24 were taken with a tripod. Just the cam was moved on a panorama plate.

 

[privatebydesign]

Without shift I didn't notice any barrel distortion. With maximum shift of 12mm barrel distortion is easily noticable as well as a lack of sharpness and resolution. Of course there is strong vignetting too.

See the attached out of cam JPEG as an example. The oc was just reduced to 800 x 55 pixel. Full 12mm of shift was used and the lens was shifted to the right. It was a test shot, after my panorama shot was done.

The final picture taken with the TS-E 24mm in portrait orientation and later on stitched looks like this:

There is no straight edge to determine any barrel distortion in that image. Also it would be nice if you gave us the actual dimensions and position of this crop within the image circle.

 

[RobertG.]

For a reference see the horizon above the lake. The cam was leveled. No correction of the horizon was done in post. The attachment in my first post is not a crop. It is the whole picture taken with the TS-E 17mm as it came out of cam. I just reduced the size to 800 x 533 to fit it into the attachment. I'm too lazy to upload it to my server.

The final image was taken with the TS-E 24mm in portrait orientation. 26 shots were stitched with Kolor Autopano (Exif data: Kolor stitching | 26 pictures | Size: 11514 x 5757 | FOV: 89.90 x 41.51 ~ 9.90 | RMS: 1.88 | Lens: Standard | Projection: Cylindrical | Color: LDR ). It should be a refence to show how the forest on the right side of the attached ouf of cam JPEG should actually look like.

 

[privatebydesign]

Sorry, that doesn't demonstrate, or illustrate, barrel distortion.

 

[RobertG.]

OK, maybe I misunderstood what is meant exactly by "barrel distortion". English is not my native tongue. But my example shows that there is noticable distortion when full shift is used. Without shift everything is fine. But this is a shift lens.

 

[JoeKerslake]

It's difficult to see without any straight lines on the image.

Take a picture of a brick wall and post it, should be able to see whether the lines dip or not in that.

 

[privatebydesign]

OK, maybe I misunderstood what is meant exactly by "barrel distortion". English is not my native tongue. But my example shows that there is noticable distortion when full shift is used. Without shift everything is fine. But this is a shift lens.

Hi Robert,

No problem, we are all here to learn. This link illustrates barrel distortion, as well as the opposite pincushion distortion and the more complicated mustache distortion.

http://en.wikipedia.org/wiki/Distortion_%28optics%29

You are right about the shifted 17 having distortion on the outside edge, but it is yet another kind, projection distortion, and I agree on full shift it can be problematic.

 

[Loswr]

OK, maybe I misunderstood what is meant exactly by "barrel distortion". English is not my native tongue. But my example shows that there is noticable distortion when full shift is used. Without shift everything is fine. But this is a shift lens.

Here's an example of the 24-105L which has severe barrel distortion at 24mm (correctable with software), compared to the TS-E 24 II with very little barrel distortion.

 

[ejenner]

The TS-E 17mm definitely has some. I don't use it for architecture, so have never had a problem. But in testing in my home I did notice it (maybe it was with a bit of shift too).

I guess it is just a lot less than other options around that FL.

However, the op should show us what he thinks is 'significant' barrel distortion.

Then again when I first got this lens I noticed weird optical effects that apparently no-body else does so maybe I'm just on something.

 

[privatebydesign]

The TS-E 17mm definitely has some. I don't use it for architecture, so have never had a problem. But in testing in my home I did notice it.

I guess it is just a lot less than other options around that FL.

I can't see how someone saying it has 'none' can have looked very closely.

Then again when I first got this lens I noticed weird optical effects that apparently no-body else does so maybe I'm just on something.

I'd be interested to see an example. When I first got mine I went to the trouble of making a custom lens profile, it took me a few hours. When I used it it didn't make a pixels worth of difference and when I changed my computer I didn't even keep it.

As for the weird optical effects, without an example it is impossible to say if everybody else has missed it, but unless your lens is faulty I seriously doubt that is the case.

 

[mackguyver]

Here's a shot I did with the TS-E 17 while comparing it against the 16-35 f/4 IS. It's not a careful composition as I was just playing around, but it has lots of straight lines and was shifted around +4, IIRC. Other than not being perfectly level (building looks to be tilting back) or parallel to the building (horizontal lines aren't correct), I see very little distortion.

Edit: here's another of the same building with the same lens - this one was more carefully composed and was probably shifted ~+6 - still not seeing any noticeable distortion:

 

[privatebydesign]

If you put a one pixel straight line on the building edges there is no measurable distortion.

However on full shift you can get very noticeable projection distortion, but that is a function of the projection distortion, not a lens aberration. I am really looking forwards to see how the new 11-24 compares to the shifted 17 for projection distortion, I anticipate the new zoom lens will be much better than the shifted 17.

 

[mackguyver]

If you put a one pixel straight line on the building edges there is no measurable distortion.

However on full shift you can get very noticeable projection distortion, but that is a function of the projection distortion, not a lens aberration. I am really looking forwards to see how the new 11-24 compares to the shifted 17 for projection distortion, I anticipate the new zoom lens will be much better than the shifted 17.

I hope the new lens is, too, and I found another example - the infamous brick wall showing obvious distortion of the zoom lens in comparison. Once again, you'll have to forgive the somewhat sloppy alignment of the camera to the wall, but I think it's enough to see the difference. Both shot at f/8, 0.6s, ISO 100, +1 EV,

Top: 16-35 f/4 IS @ 16mm

Bottom: TS-E 17