Canon 24-105 vs 17-40 at 24mm

[markodarko]

Well, in my search for a 24mm lens to better my 24-105 I've tried the following:

1 x 16-35mm f2.8 II L
2 x TS-E 24mm II f3.5 L - the results of those can be found here:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=42108311
http://forums.dpreview.com/forums/read.asp?forum=1029&message=42120240

The 16-35 was mushy in the corners and the two TS-E lenses would not focus along the bottom edge of the frame, so after some looking around on The Digital Picture (after giving up on looking for a zoom to fill my needs after my 16-35 disaster) I thought I'd try out a 17-40 as by the looks of the results at f11 and above on those charts it really seemed to hold its own.

Anyway, it arrived today.

This is a very thorough review with shots taken both inside and outside so if you just want to know the conclusion and skip the rest, here it is:

CONCLUSION

I am astounded by this 17-40 at 24mm. You really will have to prize it from my cold dead hands. Bearing in mind that I did not buy this for its zoom range (although I'm sure that will come handy too), I bought it for use as a 24mm landscape lens so being fast was not a concern of mine as I usually shoot at greater than f11 for the most part.

The distortion is incredibly minimal, the CA is very well controlled at the apertures I'll be using it at (two qualities I was looking for which are lacking in my 24-105) and the biggest surprise is how well it resolves all over the frame - even in the extreme corners - way better than the 16-35 II copy I had which was terribly mushy - about the same as the 24-105 results you see below, and obviously better than the two duff copies of the TS-E 24 II that I tested.

As mentioned before, I'm not a hoarder of gear. I've used my 24-105 for landscape for years but lately it just started to feel lacking in certain situations - namely when a horizon was concerned - and required a lot of work in lightroom to correct the CA, some of which could never be removed properly without it affecting another part of the frame. I didn't really care how much the lens cost so long as I got the results that I wanted for MY needs - hence trying two copies of the TS-E 24 II - the most expensive lens I would have ever owned - had they not been duff. Alas, there seems to be a QC problem with the latest iterations of this lens:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=42120285

So there you have it. 4 lenses later and I'm as happy as a vicar in an all-boy's school. Can't wait to use it in earnest.

Thanks to everybody who commented and helped me out in the other threads. I appreciate it.

Mark.

And now for the results (in a reply as it's too long to post on here!) if you're interested. Make sure you're sitting comfortably though, there's a fair few images to click on. Enjoy!

 

[markodarko]

TEST SETUP

All tests carried out with the following:

1) RRS Aluminium tripod with RRS BH-40 ballhead
2) Canon 5D mk II with RRS L-plate
3) Lens manually focused with Live View at x10 magnification
4) Same exposure
5) 2 Second timed shutted

FIRST TEST - INDOORS

The focus point on this scene was a typical hyperfocal distance which happened to be on the text / black triangle on the tax document on the sofa. I picked this scene for a few reasons:

1) The sofa has a contrasty fabric with a visible thread detail

2) There's a patio window to the right of the scene so I could have side-light to emphasise any CA. It's usually at its worst in this scenario in my experience.
3) It looked like rain outside.





1) Frame Centre and Focus Point

In the following test the 24-105 is better in the centre at f4 and f5.6 but thereafter the 17-40 takes over in sharpness, contrast and CA. After f11 things start to go soft at the centre, however the 17-40 is almost as sharp here at f16 as the 24-105 is at f11!

























2) Extreme Right

This test will really show the CA as it's right near the patio window and right off the bat the 17-40 is sharper at the extreme edge even wide open and even in this extreme case CA is negligible. At f11 is practically a non-issue as in the real world this kind of condition wouldn't happen. The 17-40 is useable at f16 even with all that light diffracting it but it's game over for the 24-105 after f5.6 with it showing no improvement in sharpness by stopping down here.

























3) Extreme Bottom Right

It doesn't get any more punishing than this - it's the corner near the side-light in the absolute extreme. By f5.6 the 17-40 is already starting to show some detail, but no change for the 24-105. By f8 the cloth detail is starting to show clearly for the 17-40 but the 24-105 at f8 is not as clear as the 17-40 at f5.6! Any details it's trying to show are mushy in comparison. By f11 the 17-40 is practically crisp, only pushing a little more detail into the frame by f16 and not losing any contrast in the process. By f22 it has the same detail as the 24-105 has at f16 and the 25-104 has gone soft again.

























4 Extreme Bottom Middle

This part of the frame is extremely similar, so much so it doesn't really need commenting on. Suffice to say it's all good by f11 for both of them.

























5 Extreme Bottom Left

This is pretty much the same story as test #3 so I won't repeat myself.

























6 Extreme Left

As with the extreme right, the 17-40 is sharper even wide open and showing most detail at f8. After that it starts to soften up but is still better at f11 than the 24-105 is at f8. And no, I did not win the lottery.

























7 Top Centre - Behind The Sofa By 1.5 Metres

The 24-105's CA is a real problem here from the off. By f11 the millennium falcon is nearly in focus for the 17-40 and only loses a bit of detail on the cushion by f16. The 24-105 never attains the same detail.

























8 Extreme Top Right *

Again, the 24-105's CA really lets it down here but the 17-40 is incredibly controlled. The 17-40 is getting sharp by f11 but is even sharper at f16. It's downhill for the 24-105 after f11.

























And there ends the indoor tests. It's clear from that how much detail the 17-40 can pull out of the cloth and handle CA in comparison to the 24-105 so now comes an outdoor shot to verify the findings...

 

[markodarko]

TEST SETUP

All tests carried out with the following:

1) RRS Aluminium tripod with RRS BH-40 ballhead
2) Canon 5D mk II with RRS L-plate
3) Lens manually focused with Live View at x10 magnification
4) Same exposure
5) 2 Second timed shutter

SECOND TEST - OUTDOORS

This is the same setup that I used to test the 24-105 against the TS-E 24 II the other day so it can be compared against that test easily. The focus point is the corner of the wooden step as shown in the first crop.

As in the test with the TS-E the other day, these shots were taken at hyperfocal distance with an aperture of f13.

1) Centre Frame

There's really nothing in it here between them both.





2) Extreme Right

At first, as it's a bit dark here in the shadows it's hard to see any difference but if you look at the little stones the 17-40 is sharper here, albeit marginally.





3) Extreme Bottom Right

There's just no contest here. The 17-40 is sharp and the 24-105 is mushy.





4) Extreme Bottom Centre

These are too close to really tell them apart.





5) Extreme Bottom Left

As with test #3 there's no contest here again.





6) Extreme Left

Similar to test #2, there's not much in it but the 17-40 pips it if you look at the leaves and the grain at the edge of the frame.





7) Extreme Top Left

Even though there was a bit of a wind blowing the 17-40 can still be seen to be sharper here. The 24-105's CA lets it down again.





8) Top Centre

The 17-40 is clearly sharper here. Take a look at the drain pipe and the top of the chimney. There's also more detail in the roof tiles.





9) Extreme Top Right

Again, hard to tell as it's quite dark here but the 17-40 just about pips it. The 24-105 does well though.





CLOSING THOUGHTS

Bear in mind that absolutely zero sharpening or any manipulation has happened to these images, just converted from RAW and 100% crops. To be able to make an image sharp with the RAW converter's sharpening settings you first have to start off with good ingredients, and the 17-40 has it as it's the same sharpness across the frame at f13 in this example. Here's the top-centre shot of the house at the back with a tiny bit of sharpening in the RAW converter - as you'd normally do when processing:





I'm not sure how sharp the JPG is here is after it's been imported but on my Mac, in Lightroom / Photoshop it's incredibly sharp. The whole frame is sharp after sharpening. From right below the lens to infinity and in every corner and edge, and incidentally the distance form the ground to the centre of the sensor in these shots is only 15".

Hope that was useful to anyone. It certainly corroborates the findings that The Digital Picture has on their website in my opinion.

 

[Howard102273]

Thanks for the detailed test.

I have the 17-40 and the 24-105 (have had 2 copies). Although the 24-105 is my most used general-purpose lens (mainly for its focal range), paradoxically the 17-40 is my favorite. The 24-105s I had suffer from pretty severe vignetting at 24mm, but the 17-40 is much better in this regard even at 17mm. I never had the need for a 16-35 f/2.8 as I rarely even use the 17-40 at f/4 (I use it at f/8 or higher).

I also have the 24 TS-E (mark I), which is fun to use but I rarely use it nowadays. All these lenses have their merits and limitations, you just need to learn their particularities and use the most appropriate lens for the occasion.

--
Howard
cameras: 5DII, 50D, D60, Rebel 2000 (film)

lenses: 17-40 f/4 USM L, 24-105 f/4 USM L, 100-400 f/4.5-5.6 IS USM L, 24 f/3.5 TSE L, 35 f/2, 50 f/1.4 USM, 100 f/2.8 IS USM Macro L, 300 f/2.8 IS USM II, 430 EX II, Kenko Pro 300 1.4x TC, EF 2x TC III
personal website: http://www.travelerathome.com
blog: http://travelerathome.wordpress.com

 

[markodarko]

Thanks for the detailed test.

You are most welcome.

All these lenses have their merits and limitations, you just need to learn their particularities and use the most appropriate lens for the occasion.

I couldn't have said better. Totally agree.

 

[Richt2000]

Glad you are finally happy! Looks about right. I've had mine for 7 years and I certainly won't let it go. Tried a 16-35ii and it was no sharper but 3 times the price!

Use a pinch of salt with the digital-picture / image resourcing tests, don't forget they just have one sample too...

Rich
Landscape portfolio
http://www.pbase.com/richthompson/rich_thompson___landscapes

 

[007peter]

Wow, that is some impressive difference between 17-40L vs 24-105L. I wonder if this is a full frame thing. When I use 24-105L on my Canon 30d/XTi/XSi, I was actually rather impressed, and less so with 17-40L on the same camera. It seem like 17-40L really shine on a Fullframe 5D, where as 24-105L shines better on APS-C camera which hides its corner flaw with CA or Sharpness.

Have you done the same test with an APS-C camera body?

 

[bronxbombers]

24 1.4 II is very good, blows away all the 24-105 I tried whether f/4 or f/11, especially anywhere near edges

it definitely does quite better than your 25-105 or 17-40 test samples at 24mm

Well, in my search for a 24mm lens to better my 24-105 I've tried the following:

1 x 16-35mm f2.8 II L
2 x TS-E 24mm II f3.5 L - the results of those can be found here:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=42108311
http://forums.dpreview.com/forums/read.asp?forum=1029&message=42120240

The 16-35 was mushy in the corners and the two TS-E lenses would not focus along the bottom edge of the frame, so after some looking around on The Digital Picture (after giving up on looking for a zoom to fill my needs after my 16-35 disaster) I thought I'd try out a 17-40 as by the looks of the results at f11 and above on those charts it really seemed to hold its own.

Anyway, it arrived today.

This is a very thorough review with shots taken both inside and outside so if you just want to know the conclusion and skip the rest, here it is:

CONCLUSION

I am astounded by this 17-40 at 24mm. You really will have to prize it from my cold dead hands. Bearing in mind that I did not buy this for its zoom range (although I'm sure that will come handy too), I bought it for use as a 24mm landscape lens so being fast was not a concern of mine as I usually shoot at greater than f11 for the most part.

The distortion is incredibly minimal, the CA is very well controlled at the apertures I'll be using it at (two qualities I was looking for which are lacking in my 24-105) and the biggest surprise is how well it resolves all over the frame - even in the extreme corners - way better than the 16-35 II copy I had which was terribly mushy - about the same as the 24-105 results you see below, and obviously better than the two duff copies of the TS-E 24 II that I tested.

As mentioned before, I'm not a hoarder of gear. I've used my 24-105 for landscape for years but lately it just started to feel lacking in certain situations - namely when a horizon was concerned - and required a lot of work in lightroom to correct the CA, some of which could never be removed properly without it affecting another part of the frame. I didn't really care how much the lens cost so long as I got the results that I wanted for MY needs - hence trying two copies of the TS-E 24 II - the most expensive lens I would have ever owned - had they not been duff. Alas, there seems to be a QC problem with the latest iterations of this lens:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=42120285

So there you have it. 4 lenses later and I'm as happy as a vicar in an all-boy's school. Can't wait to use it in earnest.

Thanks to everybody who commented and helped me out in the other threads. I appreciate it.

Mark.

And now for the results (in a reply as it's too long to post on here!) if you're interested. Make sure you're sitting comfortably though, there's a fair few images to click on. Enjoy!

 

[Timbukto]

I like it for its size-weight-price. But IMO I didn't find those pixel-peeping differences to be that staggering at all...very close...looks like the 24-105 zoom would be just fine for travel/convenience.

However I prefer portrait primes + UWA zoom. Things like the 40mm 2.8 STM or even nifty fifty to my 85mm 1.8...all of which IMO will show significant differences in both max aperture and sharpness, etc.

 

[markodarko]

I wonder if this is a full frame thing.
Have you done the same test with an APS-C camera body?

It certainly could be, but no, haven't tested it on an crop-body as 24mm (which is the FOV I am most comfortable with) would be nearly 40mm on a crop body so not something I'm interested in I'm afraid.

Interesting though.

 

[markodarko]

it definitely does quite better than your 25-105 or 17-40 test samples at 24mm

I'm not being awkward or "funny" when I say this at all, but I'm not really sure you can say that your 24 II does any better or worse than my test shots objectively without having shot the same scene at the same focus points and aperture and on the same camera and at the same time of day with the same import settings from RAW then and comparing the results side-by-side as I have done.

Another thing to remember is that better is objective. For my needs I'm only interested in small apertures of f11 or more due to what I particularly shoot 24mm for. Of course, at larger apertures I would expect your 24 II to be far better than my copy of 17-40, there again I also expected both the 16-45 II and the two copies of the TS-E 24 II to be better here too, but in the case of the 16-35 II for example, some users have reportedly had fantastic results with this lens at small apertures yet the one I had was mushy.

For my own personal needs, having compared all four lenses the same way, I can objectively say that my copy of the 17-40 is better than the 16-35 II and the two TS-E 24 II lenses I tried and of course than my 24-105 - which is what the point of the test was, to replace the 24mm of my 24-105 with something better.

 

[markodarko]

Glad you are finally happy! Looks about right. I've had mine for 7 years and I certainly won't let it go.

Yes, I've had mine for a day and already feel the same!

Tried a 16-35ii and it was no sharper but 3 times the price!

Completely agree, except in my tests it wasn't as sharp at the corners at all, it was mushy - on par to the 24-105 of mine which is why I sent it back. Strange how there as seemingly so many variances in quality between certain lenses. Doesn't say a lot for Canon's QC. Well, either that or more likely I guess that we as photographers look for different particulars when trying out lenses and rave about them / dismiss them accordingly.

Use a pinch of salt with the digital-picture / image resourcing tests, don't forget they just have one sample too...

Oh for sure, but having tried 3 lenses already "blind" I had to start somewhere. It was only through those TDP tests that I could see how well / badly the lenses performed at f11 or greater and that was a big help for me to pick the next candidate for testing. Next on my list was a 24 f1.4 II if this one hadn't made me smile, but this one's got me smiling like a cheshire cat. :-D

Landscape portfolio
http://www.pbase.com/richthompson/rich_thompson___landscapes

Nice shots there, Rich. The light in your "Eastnor Park" shot is beautiful.

Here's a couple of mine - taken with my 24-105:

 

[Timbukto]

The 17-40L is nice on FF, its absolutely pedestrian on APS-C.

It's the more forgiving pixel pitch factor of FF...APS-C is more demanding on the glass it uses and obviously it ignores the glass it doesn't. The 17-55 2.8 IS is far better on APS-C than the 17-40L unless you have serious poor lens copy issues.

Same thing for 85mm 1.8...it already looks pretty good on APS-C, but the more forgiving pixel pitch of FF makes it even better on FF.

But then again, this improvement is still not staggering if you get used to it...it just ends up boiling down to what gets you the photos, etc. Really enjoyed using my t3i + 85mm 1.8 this morning to get indoor pool shots...I'll post some in the Rebel forum.

 

[markodarko]

I like it for its size-weight-price.

Yes, both of those were a nice bonus!

But IMO I didn't find those pixel-peeping differences to be that staggering at all...very close...looks like the 24-105 zoom would be just fine for travel/convenience.

The 24-105 is a great lens for travel and convenience, for sure - which I'm guessing is what its primary design was for - hence the image stabilisation. However, for landscape, while it has served me well over the years as my 24mm lens it should hopefully be clear from my tests how weak it is in the corners on my camera and edges at f11 or greater - which is all I'm interested in. Quite a specific set of needs really.

Take a look at the detail in the extreme bottom right of the frame at f11 of the inside shot for example:





And here again at f16:





The 24-105 can't resolve the cloth detail on the sofa here at all, even at f16 - which is why I was looking to replace it, not because it's not a great travel lens, but because at small apertures it is not sharp across the entire frame (and has horrendous distortion and CA which is sometimes a big problem for landscape but not necessarily for travel / day-to-day use). In the above crops and if this were a landscape shot and not a sofa and there happened to be small flowers at that point in the frame for example, they would be mush if taken with my 24-105 but they wouldn't with my 17-40.

However I prefer portrait primes + UWA zoom. Things like the 40mm 2.8 STM or even nifty fifty to my 85mm 1.8...all of which IMO will show significant differences in both max aperture and sharpness, etc.

That's great for your needs but for mine - as I mentioned in my post - the whole point of my search was quite specific - to obtain a better 24mm at small apertures for landscape. Something you'd have thought would be easy to do since "all lenses behave the same way at smaller apertures" as conventional wisdom dictates, but in my experience, in practice they most certainly do not.

 

[Timbukto]

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.

I show the difference between DPP + DLO and LR4 here -

http://forums.dpreview.com/forums/read.asp?forum=1029&message=41832269

 

[markodarko]

APS-C is more demanding on the glass it uses

Sorry if this sounds rude but this doesn't make any sense. It's the other way round.

Full frame is far more demanding on the lens because it has to use a greater image circle instead of a smaller crop. Things like vignetting and corner sharpness are real issues on full frame sensors when they're seldom an issue on a crop sensor. This is especially true for wide angle lenses due to the technicalities of getting the image sharp and bright at the edges of the glass.

The 17-55 2.8 IS is far better on APS-C than the 17-40L unless you have serious poor lens copy issues.

The 17-55 looks better on a crop body because it's an APS-C specific lens. It has been designed with the correct focusing distance from the sensor etc and will not work on a full frame body.

 

[Timbukto]

Yes I didn't elaborate...putting 18MPs on an APS-C frame is more demanding on the portion that is in frame compared to 22MPs on a FF frame. If you did a APS-C crop on a 5DMK II, you have far far less than 18MPs of image left. If you need to go to an event and need more telephoto performance, its better that you bring your APS-C compared to shooting your FF and cropping...you are putting more of the MPs to use in the framing that you want, but at the same time it also means its more revealing of the glass that is in frame.

So actually APS-C with its tighter pixel pitch tends to show you more flaws at 100% view...more pixels of CA, etc...

To say that APS-C only gives you the 'sweet' zone of glass is a huge simplification. It also exposes a lot of flaws in the entire frame that wasn't ever going to be resolved by FF sensors at 22MPs, etc.

 

[bronxbombers]

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.

it definitely does quite better than your 25-105 or 17-40 test samples at 24mm

I'm not being awkward or "funny" when I say this at all, but I'm not really sure you can say that your 24 II does any better or worse than my test shots objectively without having shot the same scene at the same focus points and aperture and on the same camera and at the same time of day with the same import settings from RAW then and comparing the results side-by-side as I have done.

Another thing to remember is that better is objective. For my needs I'm only interested in small apertures of f11 or more due to what I particularly shoot 24mm for. Of course, at larger apertures I would expect your 24 II to be far better than my copy of 17-40, there again I also expected both the 16-45 II and the two copies of the TS-E 24 II to be better here too, but in the case of the 16-35 II for example, some users have reportedly had fantastic results with this lens at small apertures yet the one I had was mushy.

For my own personal needs, having compared all four lenses the same way, I can objectively say that my copy of the 17-40 is better than the 16-35 II and the two TS-E 24 II lenses I tried and of course than my 24-105 - which is what the point of the test was, to replace the 24mm of my 24-105 with something better.

 

[bronxbombers]

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.

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).

but you seem to have a nice 17-40 copy

it definitely does quite better than your 25-105 or 17-40 test samples at 24mm

I'm not being awkward or "funny" when I say this at all, but I'm not really sure you can say that your 24 II does any better or worse than my test shots objectively without having shot the same scene at the same focus points and aperture and on the same camera and at the same time of day with the same import settings from RAW then and comparing the results side-by-side as I have done.

Another thing to remember is that better is objective. For my needs I'm only interested in small apertures of f11 or more due to what I particularly shoot 24mm for. Of course, at larger apertures I would expect your 24 II to be far better than my copy of 17-40, there again I also expected both the 16-45 II and the two copies of the TS-E 24 II to be better here too, but in the case of the 16-35 II for example, some users have reportedly had fantastic results with this lens at small apertures yet the one I had was mushy.

For my own personal needs, having compared all four lenses the same way, I can objectively say that my copy of the 17-40 is better than the 16-35 II and the two TS-E 24 II lenses I tried and of course than my 24-105 - which is what the point of the test was, to replace the 24mm of my 24-105 with something better.

 

[markodarko]

I see the difference plainly but they are rather extreme corners,

It's not just the extreme corners. The top centre is also not as sharp at smaller apertures. Take a look at the millennium falcon and the edge of the sleeping bag:

f11:





f16:





The 24-105 isn't as sharp at f16 as the 17-40 is at f11 in the above crops and never becomes as sharp as the 17-40. But it's not just the extreme corners or the centre, it's the edges too:





My point is - it's not as sharp at any point of the frame at greater than f11, full stop. And I only use apertures of f11 or more for landscapes so I don't care what the performance is below that. Perhaps I should have only included those f11 or greater crops in my tests.

but if that and the CA make a big difference to you, I think you need to try Canon's DPP + DLO right away.

With all due respect, any software used to correct CA, distortion and sharpness - and I mean any software cannot put details into the images if the detail is not already there to begin with. If you attempt to sharpen mush you just end up with sharp mush.

Not as big a difference between two lenses

Well I guess that's your opinion because your eyes and needs differ from mine, and that's ok. From where I'm sitting there are huge differences between these lenses for my needs. As I concluded in my outside shot test, the detail is there in the whole frame. From edge to edge and corner to corner at f13, and for my specific needs that's all I care about.

Most lens tests you will see or read about put a lot of emphasis on how good lenses are at wide apertures, (seldom will you see any data for f22) but like I said, I'm not interested in that. In my experience it's been difficult to find a lens that will provide corner to corner sharpness at f11 or more which is why I'm very happy with this one - no matter how small a difference you believe it is. Let's not forget that the 24-105 is an L lens and no slouch in its own right. You're not going to get a lens that's twice as good for example, the differences are going to be small when it comes to improving sharpness but sharpness isn't the whole picture here as I mentioned in my original post - it's sharpness, distortion and CA.