Is 50mm distortion free?

[chris_shawn]

Hello,

 

I use a EOS 1Ds MKII with a Canon 50mm 1.4/L lens and photographed pictures of a

horizon of the sea (from a tripod). Now in photoshop it looks for me that the

horizon is slightly arched in the corners.

 

My question:

- Does even a 50mm lens have distortion (I always thought it was distortion-free)?

- Am I experiencing an other effect, e.g. is this already the arching of the

globe (I doubt it)?

 

 

Thanks!

[ronaldo_r]

"50mm" is not a magic number so it doesn't have to be completely distortion-free.

 

The Photozone review of the 40 1.4 says: "...as expected for a fix-focal lens the level of distortions is very low (but existent)..."

 

www.photozone.de/8Reviews/lenses/canon_50_14/index.htm

[ymages]

it is not a L lens

 

why dont you try EPTLens plugin in photoshop ?

 

here a test : http://www.photo.net/equipment/canon/ef50/

[ronaldo_r]

"...it is not a L lens..."

 

Are your saying L lenses are distortion free?

[chris_shawn]

oops, you are right. it is not a L lens...

[stb]

The 1.4 is known to have more distortion than the 1.8.

 

http://www.lumieredargent.com

[h._p.]

Depending on how far away the horizon is, and looking out to sea it's pretty far, I'd expect to see some curvature. Take a look here... http://en.wikipedia.org/wiki/Horizon

[tal_marcu]

Your lens should have a negligible distortion.

Do you use a CRT monitor ? Is it possible that your monitor is not well calibrated ?

[w.smith]

There is no such thing as a distortion free lens! EVERY lens distorts. Only some lenses

distort less than others. The curvature you noticed is a kind of 'wide angle distortion'. It

gets less as the focal length gets longer.

 

Minimizing that wide angle distortion is the main reason why portraits are preferably shot

with short telephoto lenses.

[beauh44]

Hi Chris, The Canon 50mm f/2.5 Compact Macro is probably as close as most lenses can get but I'm sure has a wee bit of distortion. The Hasselblad SWC (camera/lens combination) was/is really good too in that department. I think it's been discontinued but can still be found on Ebay. It's the nature of the beast, I'm afraid.

[skipd]

So far, nobody has really touched on the right answer.

 

Most "wide-angle distortion" is not true optical distortion, but a problem with perspective. When you have a camera close to a subject and are using a wide-angle lens to capture the subject, you will generally see some "distortion" of the closest elements of the subject that are at the edges of the image. This is NOT distortion in the true sense.

 

Perspective is TOTALLY controlled by the distance between the viewer (or camera) and the subject.

 

The best way to create images is to find a place to photograph a subject from that provides a good perspective. You can visualize the perspective without a camera by creating a frame with your hands or possibly a pair of carpenter's squares. Once you've found the position to take the photograph from, then - and only then - you should choose a focal length to allow you to fill your camera's viewfinder with the portion of the subject that you want in your image.

[beauh44]

Hi Skip<P>

 

Good advice! However, I believe the original "meat and taters" of Chris's question was:<P>

 

<I>Does even a 50mm lens have distortion (I always thought it was distortion-free)? </i><P>

 

And, as has been pointed out, that's not the case.<P>

 

BTW, it's also quite helpful to use a level. Best wishes...

[vrankin]

Apparent perspective distortion and optical distortion within an individual lens design are different topics. In an optical bench test a lens without distortion would render a straight line straight.

[kerry_grim]

Are you sure the distortion you are seeing is not just curvature of the Earth? However, that is my speculation, not that I am saying it is fact. I would try photographing a subject that you know has true horizontal lines.

[chris_shawn]

Thanks for the answers. They were all very helpful. I assume that my photo has a bit of it all:

 

a) perspective distortion, as a part of the wall i photographed in front doesn't run parallel to the sensor (how i held the camera). I fixed it in photoshop.

 

b) 50mm is not distortion free. Sounds very logical what you all described. Some teachers and some books are just not very exact and.

 

c) the original picture had a slight curve in the very edge of the horizon. It might have been the earth (which I doubt), but probably it was just the 50mm. I simply cropped the image.

[gdanmitchell]

As far as I know, there is no "distortion free" lens. Some lenses exhibit various types of

distortion to a greater or lesser extent, but no lens is perfect.

[m_barbu1]

<p><i>Are your saying L lenses are distortion free?</i></p>I'm pretty sure he was clarifying the original post.

[robin_sibson1]

No real-world lens is perfectly rectilinear, but there are some that come so close you'd never see the difference. According to PhotoZone, the 50/1.4 has a "very low" distortion of -0.438%, whereas the EF-S 60/2.8 has -0.078%, the CM 50/2.5 has -0.045%, and the 100/2.8USM has -0.033%, the last three being described as having "absolutely negligible" distortion. At the other extreme, the EF-S 17~85 at 17mm has distortion of -4.02% - no difficulty about seeing that!

[ted_marcus1]

Is there any lens that doesn't have some form of distortion?

[bobatkins]

You wont see earth's curvature with a 50mm lens shooting the horizon from sea level.

 

It's lens distortion. Pretty much all lenses distort, some a tiny bit that's not even noticable, some a lot.

 

If you want to avoid it, make sure the horizon runs exactly along the middle of the frame, then crop. Images of straight lines runing through the center of the image will not be curved, even with a fisheye lens!

 

The better way is to compose however you want, then correct in an image editor.

[William Michael]

>You wont see earth's curvature with a 50mm lens shooting the horizon from sea level<

 

Hello Mr Atkins,

 

I agree with this statement, and also with what you further wrote.

 

However, with respect to you, and expanding on your answer for Mr Shawn`s benefit:

 

The original question does not specify the camera position, hence: Mr Shawn can rule out the earth`s curvature being involved, if the shot was taken at, or near sea level; as the elevation of his viewpoint increases, so does the likelihood of the earth`s curvature being seen in this scenario.

 

I believe, this is the meaning of your statement of fact.

 

(We both make the assumption that his viewpoint was around sea level.)

 

Regards WW

[William Michael]

Opps

 

I am a bit of a dill - just saw the JPEG attached later in the thread, after I posted my answer. Obviously viewpoint was at sea level.

 

Regards WW

[fourfa]

photozone only tests on 1.6 crop cameras, I believe, which isn't entirely reliable for full-frame users like Chris. I find the distortion with the 5D + 50/1.4 to be objectionable close up and merely noticeable at infinity. post processing is currently the best answer other than trading for a 50/2.5

[kelly_flanigan1]

Special high buck variants of reprographic Process lenses once were made with ultra low distortion specs of 0.002 to 0.005 percent, these were used in precision map making. Theses variants had lens curvatures selected and matched to build up a lower distortion version of the already low distortion process lenses. Actually the prime design criteria for a process lens is low distortion, since one is often matching panels of maps to a precision scale. <BR><BR>Many times the lowest distortion "50mm" lens for 35mm is a slower design like a f2 or F1.8; not a fast F1.4 version. <BR><BR>In making panaromics with digital cameras typically one measures the distortion for a lens, and has experimental data points one uses to force "correctness".

[robyn_s]

A corollary question then... inspired by Kelly's comment above:

 

Where can one go to find a comparison of the rectilinearity of 50mm lenses available for the EF lens mount?

 

This is not a specification found in Canon's EF Lens Works III.

 

For example, I wonder if the 50/1.2 might be designed to have less linear distortion than the 50/1.4, despite the former being faster, and if that could be one of the reasons the L version costs so much more?

 

In fact, why not compare the 24-70/2.8 at 50 also, which would give us four currently available 50mm EF lenses that are 2.8 or faster to compare the linear distortion of. I'd be really interested in seeing this data for a full frame, as well as 1.3 APS-H frame sensor.