17mm on 1.6 crop body

[songtsen]

I have a Tamron 17-50/2.8 on my 400D and I've found the exaggerated perspective and distortion at 17mm somewhat disconcerting in relation to the FOV (28mm equivalent).

 

I've never used an ultrawide lens on my film SLR and would therefore appreciate advice on ways to deal with the unwanted ultrawide effects on my 1.6x crop DSLR.

 

Thanks,

 

Songtsen

[KenPapai]

Try this lens for much better optics: Canon EF 16-35 2.8L (you get what you pay for).

[tdigi]

I think your talking about distortion. At the wide end most lenses have some ( My 24-105 L has a lot at 24 ) If you use an

ultra wide say a 10-22 you would probably have some at 10 but at 17 you would have virtually none.

 

The 17-50 is a fine lens but its intended to be a standard zoom on a 1.6 body. If you want wide you may be happier with an

Ultra wide.

[bobatkins]

You are mistaken. The "perspective" of a 17mm lens on your 400D is exactly the same as the perspective of a 27mm lens on a full frame camera.

 

17mm is not ultrawide on a 400D. In fact it's not even superwide. It's just normal wide.

 

Better optics will in no way affect "perspective" or the type of distortion which is bothing you.

[songtsen]

Ken, wouldn't the 16-35 also behave like an ultrawide with a cropped FOV on the 400D?

[photo_dark]

oh no not again.

[songtsen]

Bob, I seem to have made a couple of wrong assumptions here, probably because I don't remember having similar issues at 28mm on my film SLR. I thought perhaps lenses retain the characteristics of their actual FL (except FOV) when used on a crop-sensor camera.

[geoffs1]

Just to inject some numbers into the discussion...

 

The Tamron 17-50mm has -2.51% distortion at 17mm http://www.photozone.de/canon-eos/289-tamron-af-17-50mm-f28-sp-xr-di-ii-ld-aspherical-if-canon-test-report--review?start=1

 

The Canon 16-35mm has -2.37% distortion at 16mm http://www.photozone.de/canon-eos/175-canon-ef-16-35mm-f28-usm-l-lab-test-report--review?start=1

 

The Canon 10-20mm has -1.25% distortion at 10mm, none at 14mm, and 0.4% at 22mm http://www.photozone.de/canon-eos/174-canon-ef-s-10-22mm-f35-45-usm-test-report--review?start=1

[j_smith6]

<i>Bob Atkins - Better optics will in no way affect "perspective" or the type of <b>distortion</b> which is bothing

you.</i>

<br><br>

I'm not English and maybe I just misunderstood what Songtsen said but I think he still means the regular "barrel

distortion". He's comparing perspective and distortion with the equivalent of the lens used on full frame camera.

[17mm@APS-C v ~27mm@135]<br>Better optics will affect distortion.

[g dan mitchell]

<p>Regarding the "perspective distortion," this is a normal "feature" of wide angle lenses. You'll see roughly the same thing

with any lens providing a similar angle of view coverage. (Fisheye lenses are even "worse" in this regard.) There are some

ways to deal with it: Level the camera carefully and put the horizon in the middle of the frame; don't go wider than necessary;

create a composition in which the main subject dominates and the "distorted" objects are at the periphery; adjust for the

"distortion" in post processing.

 

<p>Both of the Canon UWA (or WA on crop) L zooms (the 17-40mm f/4 and the 16-35mm f/2.8 II) can be great lenses and either could

be your best choice - it depends on several factors, including how and what you shoot. (Oh, and perhaps your budget?) Of course, if

you <i>need f/2.8</i> the 17-40 won't get you there, though the EFS 17-55mm f/2.8 IS will - more on that below

 

<p>When comparing the two L lenses in performance terms, the great strength of the 16-35mm f/2.8 is its performance at large

apertures: First it <i>has f/2.8</i>, which the 17-40 obviously does not have. Second, its performance at f/4 is better than

that of the 17-40 at this aperture. (Tests I've looked at suggest that the 17-40 at f/4 has roughly similar issues compared to those that

the 16-35 has at f/2.8.) At f/5.6 it is a mixed bag - one is probably a bit better in the corners while the other is a bit better in the center.

By f/8 the 17-40 seems to be a bit "better" than the 16-35, and at smaller apertures this trend continues.

 

<p>If your main use for one of these Canon L lenses is shooting wide at large apertures - perhaps handheld low light photography - the

16-35 may well be worth the additional cost and weight and can be the "best" choice. If your main use is small aperture shooting -

perhaps urban/natural landscape, etc - then the 17-40 will produce IQ that is at least equivalent and arguably perhaps a tiny bit better

in some respects, and at a lower cost and less weight/bulk.

 

<p>Again, both are fine lenses - just optimized for somewhat different strengths and weaknesses. The key is to match the lens

characteristics to your particular needs.

 

<p>One great way to visual the actual performance comparison is to use the "interactive blur charts" included in the lens tests at <a

href="http://www.slrgear.com/">slrgear.com</a>. Find the tests for the <a

href="http://www.slrgear.com/reviews/showproduct.php/product/31/cat/11">17-40</a> and the <a

href="http://www.slrgear.com/reviews/showproduct.php/product/1082/cat/11">16-35</a> and click on their respective 3D blur chart

images to open the interactive charts in two windows. With the two windows open and visible on the screen it is very easy to compare

the two lenses at the same apertures and at the same/similar focal lengths. As a crop sensor camera shooter, I think you would also be

interested to open one more interactive blur chart window for the <i><a

href="http://www.slrgear.com/reviews/showproduct.php/product/353/cat/11">EFS 17-55mm f/2.8 IS lens</a></i> - I think the results of

<i>that</i> comparison are very interesting.

 

<p>Dan

[songtsen]

Dan, thanks for the helpful advice.

[ulrich_brandl]

Songsten,

 

the 17mm lens retains its depth-of-field characteristics, but the field of view (and thus the perspective) is like 27,2mm. If you are focussing very close, the effective focal length can become shorter than indicated in some zoom constructions. I don't know if this is true for the Tamron. But I'd expect no more than 1-2mm of difference at the wide end. Maybe this together with the deeper sharpness could explain your obervations.

[equilibrium_360]

Exaggerated perspective and distortion are two separate things. Whereas perspective is depending on the focal

length, distortion is depending on the design of the lens. Both are characteristics of the lens, and sensor size

or camera type have nothing to do with either of them.

 

Perspective from a point, in strict technical terms, is the way in which two different objects appear with

respect to each other at a certain focal length. It does not change with sensor size. So, if two objects of equal

dimensions are placed such that one is closer to the camera than the other, then their RELATIVE size and shape in

the photo will be the same, irrespective of the sensor size (provided of course, both objects are within the

field of view). If the focal length is changed, it will change the field-of-view as well as the perspective.

 

Distortion, on the other hand, is a limitation of the design of the lens, either due to technical or cost-benefit

reasons. A lens would have same distortions at a particular focal length, irrespective of the sensor size or

camera type.

 

Perspective is not a limitation of a particular lens, whereas distortion is. If you want to learn more about

tools to get rid of distortions, try searching the post-archives on this forum.

[pgranone]

Get PTLENS at www.epaperpress.com/ptlens

It can correct distortion for you Tamron 17-50

 

With the Canon EF 17-40 you just shoot RAW & use the "tune" button in Digital Photo Professional. That makes it simple to correct distortion and other lens defects.

 

ALL ultra-wide angle lenses distort.

 

The Tokina 12-24 is not bad for distortion.

[markci]

<i>I thought perhaps lenses retain the characteristics of their actual FL (except FOV) when used on a crop-sensor camera.</i>

<p>

They do. But perspective as we speak about it in photography is a function of field of view, not focal length.

<p>

[markci]

<i>Whereas perspective is depending on the focal length</i>

<p>

Again, this is wrong. It is dependant solely on field of view. In particular it is a function of the field of view subtending a different angle when the photo is viewed than when it was taken.

<p>

<i>Perspective from a point, in strict technical terms, is the way in which two different objects appear with respect to each other at a certain focal length</i>

<p>

This is even more confused and wrong. The relative sizes of objects depend only on their distance from the camera. Focal length does not enter into it. If an object is twice as far away it will be half the size. It doesn't matter what lens you use to take the picture.

[songtsen]

"If an object is twice as far away it will be half the size. It doesn't matter what lens you use to take the picture."

 

So the distance between the objects appears to increase with a wideangle lens but the size of the objects remains unchanged ... did I get that right?

[equilibrium_360]

<i>In particular it (perspective) is a function of the field of view subtending a different angle when the photo

is viewed than when it was taken.</i><br>

<i>The relative sizes of objects depend only on their distance from the camera.</i><br><br>

 

Field of view covered in a particular photograph is a result of focal length and sensor size. Keep the same focal

length, same distance from camera, and change the sensor-size (crop to full, or vice-versa), and the

field-of-view in the photograph will change. However, the perspective, or the relative size and shape of the

objects, would remain the same.<br><br>

 

Consider a photograph of two footballs (at different distances from the camera) on a lawn taken using a crop

camera at a particular focal length from a particular distance. Note the relative size of the two footballs in

the photograph. Now, replace the camera with one that has full-frame sensor, and take a photo again at the same

focal length and from the same distance. The second photograph will have wider field-of-view, in the sense that

you will see some more stuff, around the two footballs, that got cropped in the first photograph. The two

footballs will appear to be of different sizes in the second photographs. However, there size relative to each

other will be the same. What that means is, if you crop the second photograph to match the field-of-view of the

first one, the two footballs would be the same size as they were in the first photograph. So, perspective did not

change, just the field-of-view did.<br><br>

 

However, now if you change the focal length, and take another photograph with the first camera from the same

distance, the relative sizes of the two footballs as well as the field-of-view would change. Even if you crop one

of the photographs to get the same field-of-view as the other, the relative sizes of the two footballs would be

different. So, perspective and field-of-view both changed this time.<br><br>

 

Since perspective only changes with change in focal-length, it is a function of the focal length.<br><br>

 

This can be observed when taking panoramic photographs as well. What you are saying, Mark, is that even if you

change the focal length between two shots for a panorama, they will overlap and align exactly right. However,

that is not true, as soon as you change the focal length, the whole perspective is shifted, and parallel lines on

the two shots would not line up.

[equilibrium_360]

<i>If an object is twice as far away it will be half the size. It doesn't matter what lens you use to take the

picture.</i><br><br>

 

Your skepticism is well-placed, Songsten. The statement is wrong.<br><br>

 

With a long telephoto lens, the distance between the objects does not affect their relative sizes, as much as it

does with a wider or fish-eye lens.<br><br>

 

Try taking a picture of any two objects from a particular point at the narrowest and longest focal-length you can

manage, and you will know what I am talking about.

[markci]

<i>With a long telephoto lens, the distance between the objects does not affect their relative sizes, as much as it does with a wider or fish-eye lens.</i>

<p>

My statement is correct, and you cannot even have tried the experiment you propose or you would have discovered it.

<p>

Long story short: you're full of shit.

[markci]

And as a display of precisely how full of shit you are, kindly explain why digicams, with tiny sensors and lenses with focal lengths typicall something like 6 or 8 millimeters, show no such perspective effects?

[markci]

<i>What you are saying, Mark, is that even if you change the focal length between two shots for a panorama, they will overlap and align exactly right. </i>

<p>

Uh, no. That's not what I'm saying. You clearly don't understand what I am saying.

<p>

Perspective distortion is a matter of viewing angle at taking not matching viewing angle when viewing. Period.

[markci]

<i>So the distance between the objects appears to increase with a wideangle lens but the size of the objects remains unchanged ... did I get that right?</i>

<p>

I'm not even sure what that means.

<p>

If you don't move the camera, changing the focal length of the lens just crops the photo differently. But a shorter lens will, if you frame the main subject the same, force you to move closer. Moving closer is what actually changes perspective. <i>Perspective itself</i> -- the size relationships of different objects in the photo is ONLY a function of camera position. Nothing else. <i>Perspective distortion</i> which is that weird exagerated effect you get with wide angles (or the compressed effect you get with long lenses) is caused by a mismatch between viewing angles when you take and when you view the photo.

[geoffm]

Listen to Mark:

 

"If you don't move the camera, changing the focal length of the lens just crops the photo differently. But a shorter lens will, if you frame the main subject the same, force you to move closer. Moving closer is what actually changes perspective. Perspective itself -- the size relationships of different objects in the photo is ONLY a function of camera position. Nothing else."

 

This has got to be one of the most misunderstood concepts in photography. It's funny, because anyone spending half an hour with a zoom lens could prove beyond doubt that what Mark says is so.

 

Thanks, Mark.

[sudhish_verma]

Mark is right