Canon 24mm f/2.8 vs. 17-40mm f/4 L Test Results

[sheldonnalos]

Hello All -

 

I was one of several parties who raised concerns about the level of

performance of the Canon EF 24mm f/2.8 prime lens about a month ago.

I was considering purchasing a Canon 17-40mm f/4 L lens at the time,

and promised that I would do a comparison and post the results for

all to review. Here it is!

 

Methodology was: Canon 10D, Large Fine JPG, manual white balance, ISO

100, tripod mounted, self timer, MLU, checked for focusing accuracy,

matched lighting conditions and exposures, no post processing (other

than levels), 100% center crops.

 

I pick the 17-40mm as the clear winner. I tested several copies of

the 17-40mm before purchasing this one. It was sharper than a 16-35mm

in an in store test (although a different 16-35mm was sharper than a

different 17-40mm in a third test...). After reading endless reviews

and tests regarding all these lenses, my only conclusion is that

sample variation plays a big part in any sort of generalization about

which lens is sharper than another. I don't think we can make blanket

statements without making some sort of "average lens" consideration.

 

In summary, no, my 17-40mm doesn't meet the high optical standards

set by my 50mm f/1.4 or 70-200mm f/4 L. However, it is still an

excellent lens and is now my primary lens on the 10D.

 

Hope this helps!

 

Sheldon<div></div>

[benb]

Did you use manual focus to ensure focus was correct?

 

There is barely any visible difference anyway. The zoom won't let you use f/2.8 if you need it, and the prime is showing slightly less saturation in your pictures.

 

Hardly a big enough difference in either direction to matter. Use whatever you feel like.

 

I suppose if you always shoot a controlled situation on a tripod like this these tests are meaningful. But there is a lot your test doesn't show.

 

Did you test flare?

[shambrick007]

Damn! All variables aside, I woulda expected better of the prime. Hope it's sample variation.

[robert_reiss]

I'm a novice, but why isn't this a significant result? Based on these images, I think I'd have to think twice about using the 24mm prime at f2.8. If I had a digital camera with an f4 lens, I'd increase the ISO. The color (or is it contrast?) and sharpness of the 17-40mm seems to be real.

[icephoto]

Good images, good comparison.<br><br>

What was left out was:<br>

24mm f2.8: <b>$290</b><br>

17-40mm f4.0 L: <b>$690</b><br><br>

 

Is there $400 difference between the lenses? Maybe to a professional. <i>Maybe</i>. I'll keep the $400 difference and spend it on other toys.<br><br>

 

Shoot'em if you've got'em... tfj

[bbc]

<i>"no post processing (other than levels)"</i>

<br><br>

Surely the fact that you adjusted the levels means that the results cannot be considered

valid?

[www.philwinterphotography.]

This test is similar to my recent comparison of an EF 20mm f2.8 to my 28-135 IS. The 28-135 beat the daylights out of the prime, and they're about the same price. Now tell me, who was it that said primes are better than zooms?

 

Phil

[chiswick_john]

"There is barely any visible difference anyway" - just the kind of thing that make personal opinions invalid - the difference is obvious to my eyes.

[jim_larson1]

I shoot the 24/2.8 a fair bit. I have not noticed this softness wide open.

 

Almost makes me want to shoot comparison shots against my 17-40/4L. almost. . . .

[gareth_harper]

I'm not surprised to hear that somebody finds the 17-40 performs better than the 24mm f2.8

I just got my 24mm f2.8 back from Canon, I'd put it in for a health check. No fault found.

So I've gone and purchased a 24mm f1.4L. A serious outlay but as a film user I love the 24mm image and 24mm f2.8 was a big disappointment in terms of detail and sharpness.

I've just processed my first rolls of film shot with the L lens, I'll let you know how I find the initial prints just as soon as I find some darkroom time. Fingers crossed the L delivers.

[tan]

<< What was left out was: 24mm f2.8: $290. 17-40mm f4.0 L: $690 >>

 

What was left out was ..... for $690 or less ..... you get 3 lenses ..... 17mm,

20/21mm, and 35/40mm. Not bad a deal at all if it's as sharp as primes!

[tan]

Let me correct myself - you actually get the following lenses with the 17-40:

 

14mm; 20/21mm; 24mm; 28mm; and 35/40mm.

 

So that's like having 5 lenses for $690 ... or less!

[tan]

Sorry for the typo ... first one should read as 17mm.

[tony_san2]

Hi Sheldon,

 

You conducted an excellent test. Your results corroborate with other on-line tests. The

17-40 is an excellent modern lens design. I own one. The wide angle Canons (non L)

never compared that well against the wide angle zooms.

 

Tony

[bob_bell]

the 24mm is a lot sharper than that. I think there is a focusing issue with the 24mm or it was not focused correctly. Not to discount your test because I know it is a learning tool but you almost have to depend on MTF graphs for a base starting point and then experience and use to make the call between close lenses.

[josh1]

Try "focus bracketing". Focus could have been off. Also, try at different focussing "distances". Lenses perform differently at different focussing distances. Some are great at infinity, yet horrible at closer focussing distances. Has to do with field curvature of the lens often.

[michael focus97.com lee]

For my money, I'd rather have my 17-40 than all the 4 or 5 primes I could have. Even if the sharpness was a bit better, I'd still choose the zoom.

 

The lens test is great, and is certainly still valid for sharpness purposes anyway, despite levels adjustments. It would be interesting to see the non-level adjustments, though, to get an idea of contrast, saturation, etc. From what I see, though, the saturation is better with the zoom.

[PuppyDigs]

I own both lenses and indeed the 17-40 is a very sharp zoom. However, it is prone to extreme distortion, especially at the wide end, whereas the zoom is virtually distortion free. Plus the zoom costs 2.5 x more, is 3 times bigger and blocks the popup on my 10D.

 

Pixel peeping and distortion aside, 8 x 12 prints are a virtual tie between the 2 lenses.

[yakim_peled1]

<p> <i>Now tell me, who was it that said primes are better than zooms? </i> </p>

<p> I am one of those who said (and still says) that. <b>My sample</b> of the 17-35/2.8 USM L was much worse than <b>my samples</b> of the 24/2.8 and 35/2. Sharpness, distortion, flare control, you name it. Also, <b>my sample</b> of the 70-200/4 USM L was much worse than <b>my samples</b> of the 85/1.8 and 200/2.8. Admittedly, I suspect that <b>my sample</b> of the 70-200/4 USM L was not 100% fine as I was not able to get a single sharp picture with it. </p>

 

 

<p>Happy shooting,<br>

Yakim.</p>

[sheldonnalos]

To add to Yakim's post, and to emphasize my comments about sample to sample variation...

 

I've tested my 24mm f/2.8 against a 17-35mm f/2.8, and the 17-35mm was better than my 24mm. This contradicts even Photodo.com. I don't think that either my methodology or Yakim's methodology or Photodo's methodology were wrong, I simply think that the sample to sample variation is a big deal, big enough for these two lenses to beat each other in different tests with different copies.

 

And to everyone who somehow feels I've slighted their 24mm lens - please don't think that's my intention. I'm just adding my experiences and tests to the wealth of information that's out there regarding these lenses. Besides, I think Puppy Face is right, the differences in my test (from f/4 on) probably wouldn't show up in an 8x10 print.

 

So the moral of the story is... test your lens for your purposes when you acquire it. If it's not good enough, get a different one.

 

Hope this helps!

 

Sheldon

[nico_smit]

Sample variation? I recently tested MY copy of the Canon 20 /2.8 against MY copy of the Canon 17-40 /4 L @ 20 mm and the prime was significantly sharper than the zoom. Even at f8/f11 the zoom was significantly less sharp than the 20 mm wide open. Just to make sure I repeated the same test about 5 times on different focus distances (solid tripod, MLU, manual focus) with the exact same result, which exclude operator error in the test. One thing that I noticed was just how much easier it was to manual focus the 20 mm prime versus the zoom. That one stop difference do make a significant difference with regards to manual focus (on a 10D).

[chiswick_john]

If sample variation is a reality - and I suspect it is - digital has given us the oportunity to test of a selection of lenses virtualy on the spot and select the best (if you have an accomodating dealer). Additionaly if their are indeed lemons out there they are more likely to turn up as used items - something to think about?

[carl_weller]

Guys, there's one thing that you're all missing.

 

We're looking at a huge blow-up of a row of books that would form an absolutely tiny

section of a 24mm by 36mm negative. The 24mm 2.8 has a close focus of about 25cm.

The books would be less than 5% of the total image captured [i'm guessing here, so don't

sharp shoot me].

 

You do not use a 24mm lense to photograph half a dozen books, where you then check to

see which shot has the sharpest text along the spine! The damn thing is for landscapes,

wide sweeping vistas, yada yada yada...

 

It's this sort of nothing-to-do-with-real-world-use test that gives us all a bad name. We

are not nerdy tech-heads who argue about the last 0.00002% of 'quality' are we?

 

Photography is about capturing the magic of a scene, communicating with people through

a two dimensional representation, not this sort of arcane techno 'my lense is sharper than

yours, but you can only see it with an electron microscope' bollocks.

 

Given that you wouldn't be able to read the text on the spine of either image when the

whole image [not a massive enlargement of a single section] was printed at a normal size,

from a normal viewing distance, I'd be happy with either lense. In fact I have both of them.

 

As others have pointed out, the 24mm is small, light, of very high quality, doesn't have

marked flare or distortion problems, and is pretty cheap for what you get. What is there to

complain about? Bugger all, but you still find people who will find something to complain

about anyway.

 

You see this in so many photography books. They start with huge sections on gear.

Cameras, lenses, flash units, blah, blah, blah. Anybody can buy the best gear, but there

are people out there with truly marginal equipment pulling off absolutely amazing work.

 

You've got to get past the gear and concentrate on the pictures or you'll end up as some

sort of weird collector of photographic equipment first, and photographer second.

 

that's my $0.02

[chiswick_john]

My 24mm 2.8 is sharpest at near distances - ask it to do some wide open landscapes and the lack of definition is all to obvious. If all people want to do is 6x4 or 10x8 prints then any lens is going to give acceptable results and even look fairly sharp to some eyes - if you want to do big enlargements and get the very best out of DSLR sensors and conventional film the differences between lenses, either sample variations or design, become much more obvious.

[michael focus97.com lee]

Carl, a photo of books is just as valid as a 'sweeping vista', for the purpose of testing sharpness. It's not that we're all missing something, it's that we are perfectionists. The more refined our eyes get, the more critical we are of our photos. Will my girlfriend care about the two-thirds rule, a tad bit of vignetting, or lines per mm? No, but I do, because I like to contiually raise the bar for myself, so that I can capture and accurately represent something beautiful that I saw.

 

Lance Armstrong uses the lightest, most aerodynamic bikes he can get ahold of. That's the tool he needs to really show what he can do. I love a sharp, high quality lens so that I can show what I can do, particularly with enlargments. I agree: capturing is what's important; but it's *how* it's captured that is also important, to me, and many others...