Canon 24-105 vs 17-40 at 24mm

[wskb]

Thanks for the write up, I found it interesting given I'm also in the market for a 24mm. Any reason you didn't consider the 24L 1.4 though?

 

[markodarko]

I'm just saying that at landscape apertures my 24 1.4 II looked better than the 24-105s I personally tried to a noticeably larger degree than your 17-40 looks better than your 24-105.

Ah ok. Gotcha.

I used to have a 17-40, on aps-c only, but going by how it did there i suspect my 24 1.4 II does better on FF (although it is true that in aps-c days i mostly focused on seeing how it did at 17mm and 40mm).

I wonder if it's related to what 007peter was experiencing on his APS-C with regards to the 17-40 "lacking" on a crop body...?:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=42138204

but you seem to have a nice 17-40 copy

Yes, I think so. I'm happy with it anyway - especially considering the journey I've had getting here with the other lenses. I thought it was going to be an easy task to replace the 24-105 at 24mm for my needs!

 

[markodarko]

Thanks for the write up, I found it interesting given I'm also in the market for a 24mm. Any reason you didn't consider the 24L 1.4 though?

No, no reason whatsoever. It was next on my list if this one didn't give me what I was looking for. Luckily, it did.

Bear in mind that I'm only interested in f11 or greater, so your needs may vary, but for my specific needs I picked the 17-40 over the 24 f1.4 II to test first because it seemed to exhibit less CA than the 24 II (and CA is one of the things that bugs me about the 24-105) in these tests:

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=100&Camera=453&Sample=0&FLI=2&API=4&LensComp=480&CameraComp=453&SampleComp=0&FLIComp=0&APIComp=8

...but with similar sharpness. It also seemed to exhibit less distortion at 24mm:

http://www.the-digital-picture.com/Reviews/Lens-Distortion.aspx?FLI=2&FLIComp=0&Lens=100&Camera=453&LensComp=480

It also exhibited less flare:

http://www.the-digital-picture.com/Reviews/Lens-Flare.aspx?Lens=100&Camera=453&FLIComp=0&APIComp=6&LensComp=480&CameraComp=453&FLI=2&API=3

Of course, at f5.6 or wider the 24 II is clearly sharper but I don't care about wide apertures for this particular need of mine so that's why I went for this one first and I'm glad I did.

 

[markodarko]

I see the difference plainly but they are rather extreme corners, but if that and the CA make a big difference to you, I think you need to try Canon's DPP + DLO right away. Not as big a difference between two lenses, but I think you'll find what it does with whatever CA you still have left is under-rated on these forums...most people still stick with LR, etc.

Knowing the 24-105 and how the image degrades after correction in a real world test, I will post the differences in image quality once both have been corrected for CA, distortion and sharpness so you may better see what I mean and why I began my search in the first place.

I understand what you mean by only seeing a small difference in what I've posted on here because as mentioned, these are just "as is". I think a real-world example will better illustrate the differences as they are amplified after correction. Obviously, in my mind I know what the differences will be like but you're right, the differences here may be too slight to see until you factor in the corrections.

Give me a sec to correct and collate and I'll post some crops again.

 

[markodarko]

IMO I didn't find those pixel-peeping differences to be that staggering at all...very close...

This is where I didn't convey what I know of how the 24-105 behaves after correcting and how much a difference not having to correct as much makes to the image. I kinda see your point if you don't take the corrections into account so here's a real-world test.

TEST CONDITIONS

The images are the same images taken from the outdoor tests above except this time I've applied sharpening, CA correction and lens correction to the images.

1) Both images had the same amount of sharpening applied in CS5 Camera RAW of 65, and Detail of 40. Both up from the default of 25.

2) CA adjustments:
24-105: -20 Red/Cyan correction
17-40: -8 Red/Cyan correction

3) Selective exposure set in Camera RAW to tone-down the brick exposure and balance out the images.

4) No additional changes were made in Camera RAW. Images were opened in Photoshop and Lens Correction of +3.15 applied to the 24-105 image with rescale image disabled so it's not stretched more than it needs to be. 17-40 did not require any correction. Here's the difference in distortion between the two lenses at 24mm:

http://www.the-digital-picture.com/Reviews/Lens-Distortion.aspx?FLI=2&FLIComp=0&Lens=100&Camera=453&LensComp=355

5) Crops taken at 100%.

CONCLUSION

Hopefully these images will greater illustrate how much of a difference not having to correct the vast lens distortion and CA of the 24-105 makes to the quality of the image in a real-world scenario. I have purposefully included near-centre crops and not just extreme corners to illustrate the differences.

As mentioned in another post, correcting the barrel distortion isn't always necessary for the 24-105 and I actively avoid doing it because of the degraded quality, but when there's a horizon involved it's unavoidable.

I'm also fully aware that there are "better" lens correction tools out there including DXO, PTLens and Canon's own DPP etc., but ultimately, not having to move pixels around will always be preferable over positional correction - which is why I started my search in the first place.





































If these were images I'd want to keep I'd add some USM in photoshop after the Lens Correction but as mentioned before, you can't add detail when it's not there and you just end up sharpening things that can't be sharpened.

Hope that helps explain my previous tests and why I saw such a huge difference between the 24-105 and the 17-40.

 

[Timbukto]

That is a bigger difference made more apparent with midframe crops. I don't dispute the quality of the 17-40L used as it should, i.e. Stopped down. I'm keeping mine, I just need a better tripod.

 

[Robert Sheldon]

Your post confirms what I "felt" about using the 17-40 in photographing covered bridges. The 17-40 is my work horse. I switch to a 24-105 if I need more than about 35mm (just to give me zoom flexibilty) or to a Sigma 12-24 if I need wider than 17mm. Oh yes, in photographing covered bridges, some times even 12mm is not wide enough. Otherwise I use the 17-40 about 85% of the time and I use f13/16 aperature.
--
Bob Sheldon
Photo Gallery at
http://www.bobsheldon.com

 

[markodarko]

Your post confirms what I "felt" about using the 17-40

That's great, Bob. I'm glad it was of use to you. I'm glad I did such a thorough test for myself at different apertures because I now know the lens' limitations and sweet spot (f11) and know how far I can push the aperture if need be - which in other tests I've done can be up to f18 and still retain detail. Things like that are good to know when you're out in the field so you're not wondering "Hmm, I wonder if this will be sharp if I choose f16" etc.

 

[markodarko]

That is a bigger difference made more apparent with midframe crops.

Yes, I should have explained that better in the first place really (about how the quality of the 24-105 degrades drastically after correction) so I apologise about that.

Quite a difference, isn't it. Hence my elation at the 17-40 at 24mm with not having to correct it.

 

[rwbaron]

Thanks for an very indepth test. You remind me of myself when I buy new equipment and frankly I have a tendancy to drive myself a bit looney.

This is helpful as I'm about the buy a 5D3 and have the 24-105 which I 've used with my 40D, 50D and now 7D. I've been very happy with the lens and it's much better on the crop camera than the 15-85. I will need a replacement for my 10-22 and my first choice would be the 17-40 as I sold a 16-35 series 1 a few years ago due to non use. I just did not like the range on APS C.

Your test validates my idea of the 5D3, 17-40, 24-105, and my 70-200f4LIS along with my 300f4L and 400f5.6L. I think I should have most situations covered with that line up and the 7D.

Thanks again for posting.

Bob
--
http://www.pbase.com/rwbaron