Jump to content


Photo
- - - - -

Canon FF lens choice


  • Please log in to reply
46 replies to this topic

#21 craig

craig

    Full Moon Rising

  • Super Mod
  • 2826 posts
  • Gender:Male
  • Location:Austin, TX

Posted 12 January 2005 - 06:28 PM

Here's a quick shot using the 17-40 at f/4 with the 1Ds2. Edge softness isn't so bad but there's some CA. I'm going to play with that some more. I thought I'd post this to demonstrate the enormous barrel distortion this lens has. Yeah, so it's not a big problem underwater but you are paying for a rectilinear lens. There is noticable vignetting but the test chart is not carefully enough lit to conclude anything.
I love it when a plan comes together.
- Col. John "Hannibal" Smith

------
Nikon, Seatool, Nexus, Inon
My Galleries

#22 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 13 January 2005 - 03:34 AM

Craig, I would comment that with regards to edge sharpness, your test chart ends pretty much right where I see my lens starting to go blurry. I note that your bottom right and left corners are both softer than the top. This is what I observed on my lens too - it's a little annoying as for most shots I could live with some blurring in the top corners - in the sky or water. Perhaps the solution is to shoot with the camera upside down? :wink:

With regards to the LL forums, I was not suggesting that these were necessarily wonderful, just responding to the assertion that 17-40L is better than 16-35L because Luminous Landscape says so (Reichman tested one lens of each). Some full frame lens testers on RG's site, which does in my opinion have the best signal to noise ratio on the net (outside wetpixel, of course), have made similar comments ie that the 17-40L isn't worth testing/considering for FF. I disagree with this, but felt it was worth pointing out that my opinion was a minority dissenting one.

Here's my test shot from last week for barrel distortion and vignetting at 17mm wide open. Note the light source was nice and even, so this is true optical vignetting that you are seeing.

<edit> I should add, this is shot as JPEG as metered, not sharpened, but has had levels applied as I would for a normal photo, to set appropriate black and white points.

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com


#23 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 13 January 2005 - 03:49 AM

Edge softness isn't so bad


Well, my corners look pretty damn ugly wide open! Don't get me wrong, I really like this lens, but I can't think of many situations where I'd be willing to accept the kind of optical performance shown on these test shots. The vignetting and corner sharpness clears up pretty well by f/8, which is about as wide as I'd use this lens on a full frame camera.

100% crop bottom left corner, 17-40L @17mm, f/4. Being an 'L', the corner blur does have a magical painterly quality to it.

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com


#24 Alex_Mustard

Alex_Mustard

    The Doctor

  • Super Mod
  • 8375 posts
  • Gender:Male
  • Location:Peterborough

Posted 13 January 2005 - 04:12 AM

Don't get me wrong, I really like this lens, but I can't think of many situations where I'd be willing to accept the kind of optical performance shown on these test shots.  The vignetting and corner sharpness clears up pretty well by f/8.


I'd never condoning the poor performance of any lens - but careful composition and choice of subjects is one way to avoid optical problems if because of in situ light conditions you have to shoot wide open. As you say, Whitey, the problem would not be noticeable with blue/open water in the corners.

I think the top vs bottom of frame differences are because the camera is not parallel to the test chart - and not representative of an asymmetric problem with the lens? I'm not sure if you are joking!

Alex

One difference between your tests is one is I presume Craig's is shot UW, through a dome. Which is obviously a far more representative. But as he admits he has only just got the camera - and it was a quick shot. I'm sure Craig's proper tests will be more conclusive.

I have also heard the chat about the QC variability of these lenses. As James says - take your camera to the shop and try several!

Alexander Mustard - www.amustard.com - www.magic-filters.com
Nikon D4 (Subal housing). Nikon D7100 (Subal housing). Olympus EPL-5 (Nauticam housing).


#25 Kasey

Kasey

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 709 posts
  • Location:USVI

Posted 13 January 2005 - 04:19 AM

I don't think that the 1Ds MkII gave succh severe vignetting. Am I wrong? Perhaps Canon will work on this in software - boost the corners to match certain lenses.

These problems make me wonder what all the FF hubub is about! This is totally unacceptable for landscape work in my view.
Seacam F100;D2x; 60mm;105mm;16mm;17-35; 10.5mm;12-24mm
Sea & Sea strobes
www.underthecaribbean.com

#26 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 13 January 2005 - 05:28 AM

 

I think the top vs bottom of frame differences are because the camera is not parallel to the test chart - and not representative of an asymmetric problem with the lens? I'm not sure if you are joking!  

I have also heard the chat about the QC variability of these lenses. As James says - take your camera to the shop and try several!


No, Alex, just this once I'm not joking.

If you look at Craig's test shot, the 16:9 number is clearly softer at the bottom corners than the top corners - and the test chart finishes just where the dodgy bit of this lens starts. If you look at my brick wall, it's also definitely softer at the bottom corners than the top corners (I posted the 100% crops in a recent thread). I thought this was just my lens, but it's possible that there's a trend here. I'm aware of a grand total of 4 full frame 17-40L test shots published on the web, of which 2 are on this thread. And as I discovered on the 'dive medical' thread, just because data's not statistically significant, that doesn't mean it's not important! :roll:

I was joking about shooting upside down (although it would work!), and the bit about the L-quality blurriness. Just so you know. :wink:

I've got no reason to think this is a poor quality example 17-40L; I'd estimate I've shot 15,000 frames with it on my crop sensor camera, this particular lens is better than the 20/2.8 and 24/2.8 primes. Its performance in general is better than I expected on the 1Ds.

I don't think that the 1Ds MkII gave succh severe vignetting. Am I wrong? Perhaps Canon will work on this in software - boost the corners to match certain lenses.  

These problems make me wonder what all the FF hubub is about! This is totally unacceptable for landscape work in my view.


Yes, you're wrong! 1Ds Mk2 will give equal vignetting, and the appearance of worse corner performance (because the centre is relatively better).

Some would argue that this is because of angle of incidence of light in the corners and digital sensors not coping with this as well as film. Others would argue that this is rubbish and that we just used to spend less time photographing walls in the velvia days! :D

This lens is fine for landscape work - at f11, on a tripod - check my 17mm shot I linked to in my earlier post. The FF hubbub is justified! A sharp vignette-free 11 megapixel shot at 17mm is a beautiful thing, and this lens will deliver that.

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com


#27 Kasey

Kasey

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 709 posts
  • Location:USVI

Posted 13 January 2005 - 05:56 AM

Has anyone shot sunbursts with the mKII yet?
Seacam F100;D2x; 60mm;105mm;16mm;17-35; 10.5mm;12-24mm
Sea & Sea strobes
www.underthecaribbean.com

#28 Kasey

Kasey

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 709 posts
  • Location:USVI

Posted 13 January 2005 - 06:03 AM

I looked at the image linked above, and I thought that the vignetting was apparent and distracting especially in the foreground. Very nice image though, and I don't think it is spoiled by the vignetting, I'm just not convinced that this is acceptable performance.

To my eye - the 1.3X crop might be the sweet spot for sensor size.

Not trying to pick a fight, as I'm awed by the macro performance of the FF cameras, but FF for macro and crop sensor for WA is much the opposite of what I expected.

The DR in the above image looks very impressive. Has it been post processed or is this the performance out of the camera. Very nice indeed.
Seacam F100;D2x; 60mm;105mm;16mm;17-35; 10.5mm;12-24mm
Sea & Sea strobes
www.underthecaribbean.com

#29 james

james

    The Engineer

  • Super Mod
  • 9969 posts
  • Location:Houston TX

Posted 13 January 2005 - 06:27 AM

Maybe I'll have to shoot the Kodak FF with the Nikkor 17-35 before I send it off tomorrow. Now if I can only find a brick wall...

Cheers
James
Canon 1DsMkIII - Seacam Housing
Dual Ikelite Strobes
Photo site - www.reefpix.org

#30 scorpio_fish

scorpio_fish

    Orca

  • Member
  • PipPipPipPipPip
  • 1412 posts
  • Gender:Male
  • Location:Dallas, TX

Posted 13 January 2005 - 07:04 AM

You can get a used 1Ds for US$3700 or so, which is relatively cheap


It might be relatively cheap, but someone has pushed the slider bar way to the right. I never conceived of spending over a $1000 for a film body.

Whitey, you listed the Tamron 90mm as one of the lenses you've used. Could you compare that to the Canon macros you talked about? I like to hear a comparison to say the 100mm.

#31 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 13 January 2005 - 07:24 AM

Not trying to pick a fight, as I'm awed by the macro performance of the FF cameras, but FF for macro and crop sensor for WA is much the opposite of what I expected.  


Kasey, it's OK fella, I don't think this thread has bought me to fighting point yet... :wink:

The 17-40L has no real vignetting on this lens at the f/11 aperture I shot that photo at. I suspect what you see in the picture is just shadow. Check out this thread (the bit at the end on page 3)
http://www.wetpixel....-start-30.phtml
from last week - there's a very fine, non-vignetting photo of the very same wall at f11. I won't bore everyone by posting yet more wall shots (I promise!!! :shock: ), but the point of shooting them is that vignetting and distortion are way more obvious on a wall than a real life shot.

The point of the above f4 17mm wall shot was to show you this lens at its worst, because for me it does alter the way I use this lens - I'm not trying to discourage anyone who's thinking of going Canon full frame. On the 10D, I had no hesitation shooting at f/4, but the 1Ds just shows up too many flaws wide open to be useable for the majority of shots I do.

(BTW, The shot is post processed from two raw conversions from the same raw file - the dynamic range isn't that good straight out of the camera.)

James, the wall is on the side of my house. You'd be more than welcome to use it if you wish. Bring the Kodak and we'll shoot a comparison of Nikon v Canon. By the way (and I can't believe I'm trying to point out interesting features in a wall photo), do you like the vertical wire down the wall? That's for keeping the roof on when a category 5 cyclone hits.

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com


#32 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 13 January 2005 - 07:34 AM

You can get a used 1Ds for US$3700 or so, which is relatively cheap


It might be relatively cheap, but someone has pushed the slider bar way to the right. I never conceived of spending over a $1000 for a film body.

Whitey, you listed the Tamron 90mm as one of the lenses you've used. Could you compare that to the Canon macros you talked about? I like to hear a comparison to say the 100mm.


It's probably only cheap compared to the $7000 it was a few months ago. Or a P25 MF back. Still way more than I'd like to pay - particularly for something I may want to take underwater, given my recent propensity for flooding things!

I've had the Tamron 90mm macro since 1998, but I just don't really like this lens. It's very sharp, but AF is slow. I also like the look of the Ikelite port for the 100mm USM, which is the main reason I'm in the process of buying the 100mm USM myself. I'll let you know how they compare in a couple of weeks.

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com


#33 Craig Ruaux

Craig Ruaux

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 788 posts
  • Gender:Male
  • Location:Oregon, USA

Posted 13 January 2005 - 08:15 AM

Now if I can only find a brick wall...


I know you have a brick house James, so you're not getting out of it that easy :wink:

And some of the houses in your street have nice horizontal siding that will show barrell distortion quite well...
Why would I take a perfectly good camera underwater??
D300, D200, D70, 12-24 f4 AFS DX, 60mm f2.8, 70-200 f2.8 AF-S VR, 105 f2.8 AF-S VR, Tokina Wunderlens.

Photo galleries @ Ruaux.net

#34 james

james

    The Engineer

  • Super Mod
  • 9969 posts
  • Location:Houston TX

Posted 13 January 2005 - 08:52 AM

Good pt. She's a brick house!
Canon 1DsMkIII - Seacam Housing
Dual Ikelite Strobes
Photo site - www.reefpix.org

#35 james

james

    The Engineer

  • Super Mod
  • 9969 posts
  • Location:Houston TX

Posted 14 January 2005 - 09:55 AM

OK, ask and ye shall receive! I took a number of photos with the Kodak and the 17-35 before I ship today. I shot JPEG. I resized from 4500 pix wide to 500 and had to sharpen after that at 120 .7 0

[Edit: OK, it looks like this post got FUBAR'd when I moved my albums around. Images can be found at: http://www.reefpix.o.../album87?page=2 ]


Here's F2.8:

Posted Image

And a corner crop:

Posted Image

Here's F8

Posted Image

And a corner crop:

Posted Image

And for vignetting and sharpness comparison, I shot my garage door in deep shadow:

Posted Image

Vignetting is really easy to correct in the RAW converter, but I left it in here for comparison to other lenses.

My conclusions: Barrel distortion is MUCH better w/ the Nikon. Sharpness is a little better, but not drastically better. Vignetting is hard to compare without equal test photos. The brick wall at f2.8 shows very little vignetting. The garage door in shadow shows a LOT.

Your thoughts folks?

Cheers
James
Canon 1DsMkIII - Seacam Housing
Dual Ikelite Strobes
Photo site - www.reefpix.org

#36 UWphotoNewbie

UWphotoNewbie

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 857 posts
  • Location:Broomfield, Colorado

Posted 14 January 2005 - 10:20 AM

This is starting to get interesting!

Now that there are a whole set of WA DX lenses out for cropped sensors and a high megapixel dx camera due out soon from Nikon it would be interesting to test lens performance on both DX and FF cameras to see if there might not be some advantages yet with the cropped format.

Since the 17-35mm (or 16-35) on FF is basically the 12-24mm (or 10-22 more or less) on a cropped sensor, I wonder how these lenses stack up in terms of barrel distortion edge sharpness and vigenetting? How about the 10.5mm on dx vs the 16mm on FF?

It doesn't make much sense IMHO to say that the 17-35 performs better on a cropped sensor. This that is obvious. But does the cropped system afford any benifit in lens performance to counter the possible disadvantages in noise performance? (I'm not sure we are at the noise performance limit yet with the current megapixel range--maybe in consumer digicams)

Anyone with a bunch of gear to compare care to try?

UWPhotoNewbie: Not such a newbie to diving and UW photography.

Nikon D70: 60 mm, 11-16mm, 105mm, 15mm, 10.5mm

Ikelite iTTL Housing, dual Ikelite DS125

Nikon D600 topside 14-24, 28-300, 70-200, 35,50,85


#37 UWphotoNewbie

UWphotoNewbie

    Great Hammerhead

  • Member
  • PipPipPipPip
  • 857 posts
  • Location:Broomfield, Colorado

Posted 14 January 2005 - 10:31 AM

One more thing Re: test setup.

Rob, can you post your exposure info and test setup? Did you use a tripod? I agree that you should be more parallel to the wall. It looks as though you used a flash?

It looks as though you might have used a flash as it is very bright in the middle and less-so as you go out. One concern I have is that if you used a flash and the exposure was long and not on a stable tripod the corner sharpness could be due to motion blur in the areas not well illuminated by the flash. Maybe this is just the effect of extreme vignetting?

It would be better to do the test in full sunlight on a tripod where all 4 coners are not in the shade.

Thanks Rob for being willing to post your quick test results. I don't mean to pick apart something that was never intended as anything more than a quick snapshot. Its just that it got the engineering wheels turning!

UWPhotoNewbie: Not such a newbie to diving and UW photography.

Nikon D70: 60 mm, 11-16mm, 105mm, 15mm, 10.5mm

Ikelite iTTL Housing, dual Ikelite DS125

Nikon D600 topside 14-24, 28-300, 70-200, 35,50,85


#38 herbko

herbko

    Herbzilla

  • Super Mod
  • 2128 posts
  • Gender:Male
  • Location:Northern California

Posted 14 January 2005 - 10:55 AM

Since the 17-35mm (or 16-35) on FF is basically the 12-24mm (or 10-22 more or less) on a cropped sensor, I wonder how these lenses stack up in terms of barrel distortion edge sharpness and vigenetting? How about the 10.5mm on dx vs the 16mm on FF?  

It doesn't make much sense IMHO to say that the 17-35 performs better on a cropped sensor. This that is obvious. But does the cropped system afford any benifit in lens performance to counter the possible disadvantages in noise performance? (I'm not sure we are at the noise performance limit yet with the current megapixel range--maybe in consumer digicams)

Anyone with a bunch of gear to compare care to try?


This is an interesting question. Larger dynamic range of the FF sensor will be useful in some places, and packing 12M pixels on a 1.5x crop chip also has limitations. You'll run into diffraction and lens resolution limits at a wider aperture.

I have the 10-22. I can take a few test shot if I can find a brickwall. I live in earthquake country; we don't have brick buildings here.
Herb Ko http://herbko.net
Canon 5D; Aquatica housing; 2 Inon Z220 strobes; Canon 100mm macro, 17-40mm ; Sigma 15mm FE, 24mm macro, 50mm macro

#39 james

james

    The Engineer

  • Super Mod
  • 9969 posts
  • Location:Houston TX

Posted 14 January 2005 - 11:58 AM

Hi Herb,

Cinder blocks work if you back up further.

Cheers
James
Canon 1DsMkIII - Seacam Housing
Dual Ikelite Strobes
Photo site - www.reefpix.org

#40 whitey

whitey

    Manta Ray

  • Member
  • PipPipPip
  • 400 posts
  • Gender:Male
  • Location:Port Hedland in Australia's Northwest
  • Interests:All forms of nature photography. Dive medicine. The ocean.

Posted 14 January 2005 - 08:00 PM

One more thing Re: test setup.  

Rob, can you post your exposure info and test setup? Did you use a tripod? I agree that you should be more parallel to the wall.  It looks as though you used a flash?  

It looks as though you might have used a flash as it is very bright in the middle and less-so as you go out. One concern I have is that if you used a flash and the exposure was long and not on a stable tripod the corner sharpness could be due to motion blur in the areas not well illuminated by the flash. Maybe this is just the effect of extreme vignetting?


UWphotonewbie, shot at 17mm, f4, 1/4s, iso 100, tripod mount, mirror lock up, cable release, JPEG.

Light was indirect and diffuse in quality, and definitely not flash! (what kind of a dodgy wall tester do you think I am? :D ). It was just as or after the sun was setting, so there was no direct light - you would expect a more contrasty and more apparently sharp image if I'd shot this with direct sunlight. Sensor plane was as parallel to wall. I'm afraid I'm not an engineer, but I do think my lens testing methodology will stand up to close scrutiny! :D

The shot in this thread had manual levels adjustment as previously stated - this obviously makes the vignetting look worse, but it represents what I do to every shot I take so I thought it was more realistic. I'm not trying to make this lens look good, I just want to know its limitations so I can work around them. You can't compare vignetting to the Nikon shot unless it's also had levels adjustment applied.

James, I'm not sure about your conclusions re: the Nikkor 17-35's perfromance, my eyes don't entirely agree but then it's a bit hard to compare directly given the different lighting. I think both lenses are pretty horrible in the corners wide open. It would be interesting to test them side by side (I haven't seen this done elsewhere), and my offer for you to use my wall still stands. I do like your wall though, very nice.

:lol:

Rob Whitehead

Shooting with Phase One and Canon. EWA-Marine Factory Test Pilot.

www.pilbaraphoto.com