Open Photography Forums  
HOME FORUMS NEWS FAQ SEARCH

Go Back   Open Photography Forums > Digital Camera Discussion > Lenses: DSLR and Rangefinder, MF adaptions to 35mm such Zoerk

Reply
 
Thread Tools Display Modes
  #1  
Old July 7th, 2009, 09:16 AM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default Canon TS-E 24mm f/3.5L II - test images

Hi folks,

I've been testing my new TS-E Series II lens, and I'm starting to get a grip on its specifics. I'll share a couple of images, but I've not spent any time on color or tonal adjustments. You'll basically see a straight Raw conversion, without CA correction or special sharpening (other than a uniform sharpening to recover from diffraction at f/11). There are some mazing artifacts and moiré, which basically are easy to prevent with a little attention and a good Raw converter. The fact that there is aliasing/moiré visible is a sign that there is more resolution than the sensor can resolve, despite diffraction blur !

Okay, without further delay, here is a regular non shifted image:


When you click on the image, you'll get the full size (9+MB !!) version, so it may take a while if you have a slow connection, you have been warned.

Again, the colors/tonemapping are far from what I'd like and the light was horrible and it was hot and a bit hazy (UV-index warning), so use the image to inspect for sharpness and CA when only using a bit of tilt to increase DOF (typical landscape scenario).

I'll post a few other examples later.

Bart
Reply With Quote
  #2  
Old July 7th, 2009, 11:07 AM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default

And here's another example. This time I tested the result of an orthographic stitch with 2 images shifted in opposite directions relative to the camera shift, and then simply overlayed them at the overlap in Photoshop. That will tell how much of the +/- 12mm shift will still have a usable image quality. To help in determining the image edges with the various shift amounts, I've marked them in the image.

Here's the result:


When you click on the image, you'll get the full size (9+MB !!) version, so again it may take a while if you have a slow connection, you have been warned.

Do note that no special processing was done for the corner regions, other than the general sharpening of the entire image. That general sharpening could be replaced by something that does more justice to the specific aberrations towards the edge of the image circle. Also, the light fall-off can be compensated in post-processing, but there will still be more noise towards the corners due to relatively lower exposure levels, so that should also be fixed in PP.

To me, as far as we would accept this low level of post-processing, the image tends to drop off rapidly beyond 8mm shift in the landscape direction. That would suggest that a vertical shift in landscape image orientation (e.g. in architecture) would allow a full 12mm shift with only the extreme corners being compromised. I have an example of such a shift which I'll post later.

Also in this image pair, despite the diffraction blur from the f/11 aperture, there are Raw conversion artifacts (moiré and false color artifacts) due to excess of resolution. Tweaking the Raw converter, or a double conversion with the appropriate settings, should reduce the visibility significantly.

Bart
Reply With Quote
  #3  
Old July 7th, 2009, 12:27 PM
Michael Fontana Michael Fontana is offline
Senior Member
 
Join Date: Dec 2006
Location: Switzerland
Posts: 1,557
Default

Hi Bart

thanks for sharing your tests, adding the shift scale and already a first question:

where did you have the focus - on the flatstitch - the 2nd image?
The curvature of field (hopefully the correct word in english) is interesting:

at the center it looks like the focus was on the center windmill, while at right border (especially) the focus seems much closer, meanwhile at the left....

Correct me if I'm wrong, but I don't think that its reason is the tilt of 1 degree, only, but the curvature of field, unless you moved focus accidently by shifting.

The unshiftet image looks nice, as does every shift lens, using the sweet spot - the inner circle only.

While CA is surprisingly well controlled - even at the big, 12mm shifts of the flatstitch - I was hoping for a larger °sweet range° at each side.

Could you provide a °brickwall°-image, please?
__________________
http://www.proimago.net
Reply With Quote
  #4  
Old July 7th, 2009, 01:25 PM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default

Quote:
Originally Posted by Michael Fontana View Post
Hi Bart

thanks for sharing your tests, adding the shift scale and already a first question:

where did you have the focus - on the flatstitch - the 2nd image?
The curvature of field (hopefully the correct word in english) is interesting:

at the center it looks like the focus was on the center windmill, while at right border (especially) the focus seems much closer, meanwhile at the left....

Correct me if I'm wrong, but I don't think that its reason is the tilt of 1 degree, only, but the curvature of field, unless you moved focus accidently by shifting.
Hi Michael,

Yes, it is a combination of curvature of field and other sherical aberrations, together with the tilt. The curvature of field is the reason why a flat brick wall doesn't tell the whole story. That's why I also wanted to test things on subjects with predictable 'depth', and see if the curvature could be put to good use. I focused the near end of the focus plane on the waterlilies (but in Live View the distortion was already clearly visible), and at the far end I focused at the windmil in the center.

Nothing was changed between shots, other than the shift and the camera position in the opposite direction on a horizontal bar.

Quote:
The unshiftet image looks nice, as does every shift lens.
Yes, it's the best unshifted 24mm result I've ever seen. There's a recent comparison with 2 other lenses at FM forums. The Nikon 14-24 G lens does pretty well, even though its not at its best at the 24mm end, but even better at the 14mm end.

Quote:
While CA is surprisingly well controlled - even at the big, 12mm shifts of the flatstitch - I was hoping for a bigger °nice range° at each side.
Yes I was hoping that also, but that's why I wanted to share these test images. This lens is not perfect as some would like to make you believe, but it's much better than any 24mm lens I've seen before, and it has tilt (which allows to use it at wider apertures and still get lots of depth of field), and still a decent amount of shift if you're willing to do some postpocessing.

Quote:
Could you provide a °brickwall°-image, please?
Yes, I have an example that's interesting, so I'll post it later. I have also more details about how diffraction affects center and shifted corner performance. Interesting fact; the lens has its highest center resolution at f/4.0, while the far corner performance keeps improving (but to a lower peak level) until f/11, maybe f/13. It may be useful for some subjects to shoot 2 images at different apertures and blend them.

Bart
Reply With Quote
  #5  
Old July 7th, 2009, 03:46 PM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default

And here's another type of subject, same approach to test for the usable shift boundaries of the extrteme horizontal shift in landscape orientation:



When you click on the image, you'll get the full size (11+MB !!) version, so again it may take a while if you have a slow connection, you have been warned.

In this image the CA is a bit easier to spot, but still easy enough to correct in post-processing. It also demonstrates the benefit of the tilt capability. However, this lens also has a beautiful bokeh, amongst others due to the almost round aperture. It is therefore also possible to use the tilt in an opposite direction to force a very shallow depth of field.

Bart
Reply With Quote
  #6  
Old July 7th, 2009, 01:06 PM
Asher Kelman Asher Kelman is offline
OPF Owner/Editor-in-Chief
 
Join Date: Apr 2006
Posts: 31,748
Default



"When you click on the image, you'll get the full size (9+MB !!) version, so again it may take a while if you have a slow connection, you have been warned."

Let me put aside the architectural necessities of the lens for a moment. The picture itself is so beautiful with the vignetting at the edges being exactly what I'd work for. I cannot stand uniform illumination. Artifacts really don't matter much in this picture.

If I can get this in two shots of a landscape with the center well drawn, I'm happy.

Now would this also work for a picture of an entire orchestra?

That really needs the pixels this would offer.

For architecture, having a little robot like the Gigapan cannot give the tilt's this provides. Certainly this is more convenient than putting a digital back or camera on a LF setup.

This study is going to be very helpful. I'm looking forward to more of this but right now, it appears to be a very simple way of getting a superwide detail rich landscape.

Thanks, I'm enjoying the scene!

Asher
__________________
Follow us on Twitter at @opfweb

Our purpose is getting to an impressive photograph. So we encourage browsing and then feedback. Consider a link to your galleries annotated, C&C welcomed. Images posted within OPF are assumed to be for Comment & Critique, unless otherwise designated.
Reply With Quote
  #7  
Old July 7th, 2009, 02:16 PM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default

Quote:
Originally Posted by Asher Kelman View Post
Now would this also work for a picture of an entire orchestra?
Hi Asher,

It would work for large groups, with the same limitations for edge sharpness. Besides, with the full L/R shift the image simulates a 60x24mm sensor, which equates to a 103 degree horizontal angle of view. That's a bit too much, also for people (unless you put the skinny ones at the edges ;-)). As can be seen from the waterlilies, geometrical distortion becomes promeninent towards 100 degrees coverage.

Having said that, there are solutions for that with appropriate software. DxO apparently can compensate, but also some pano stitching software like PTAssembler. And talking about stitching software, when this lens is simply rotated around the entry pupil, you'll get super (center of lens) resolution faster than shifting camera and lens in opposite directions. Also nice is that when using center of lens resolution, the lens can be used at a wider aperture, and the lens is very sharp between f/4.0 and f/8.0 (before diffraction starts to have its effect on resolution).

Quote:
For architecture, having a little robot like the Gigapan cannot give the tilt's this provides. Certainly this is more convenient than putting a digital back or camera on a LF setup.
A T/S offers a very flexible way of shooting, with lots of control. I got the lens because I needed something good in the 24mm range, with the added flexibility of tilt and shift. The tilt is something that currently cannot be done easily (if at all) in postprocessing without resorting to focus-stacking and lots of post-processing. When time is money, tilt is king.

Quote:
This study is going to be very helpful. I'm looking forward to more of this but right now, it appears to be a very simple way of getting a superwide detail rich landscape.

Thanks, I'm enjoying the scene!
It's a journey which I am glad to share. The scene BTW is from a UNESCO World Heritage site in the Netherlands, called "Kinderdijk". It's a collection of 19 windmills in close proximity, 16 of which were built in 1738 and 1740 to pump out the water through the canals that were dug by hand in 1366 and 1369. I'll go back sometime when the light is better and the temperatures are below 35 degrees Celcius (95 F).

Bart
Reply With Quote
  #8  
Old July 7th, 2009, 03:23 PM
Cem_Usakligil Cem_Usakligil is offline
Senior Member
 
Join Date: Jun 2006
Posts: 5,947
Default

Quote:
Originally Posted by Asher Kelman View Post
... If I can get this in two shots of a landscape with the center well drawn, I'm happy.

Now would this also work for a picture of an entire orchestra?

That really needs the pixels this would offer.

For architecture, having a little robot like the Gigapan cannot give the tilt's this provides. Certainly this is more convenient than putting a digital back or camera on a LF setup.

This study is going to be very helpful. I'm looking forward to more of this but right now, it appears to be a very simple way of getting a superwide detail rich landscape.

Thanks, I'm enjoying the scene!

Asher
Hi Asher,

You do not necessarily need this lens for the purposes you state IMO. Even the "lowly" 17-40L does a great job when it comes to landscapes. If Bart doesn't mind me doing so, I can show an example from the same locality (since I was there with him during this shoot ;-)). Or I can send it to you by PM, whatever.

@Bart: This is yet another fantastic thread, full of good info as we came to expect from you time after time. But we should not take it for granted. I, for one, appreciate a lot the valuable knowledge you share with us. Thx.

Cheers,
__________________
Kind Regards, Cem

flickr
website
Reply With Quote
  #9  
Old July 7th, 2009, 03:40 PM
Asher Kelman Asher Kelman is offline
OPF Owner/Editor-in-Chief
 
Join Date: Apr 2006
Posts: 31,748
Default

Quote:
Originally Posted by Cem_Usakligil View Post
Hi Asher,

You do not necessarily need this lens for the purposes you state IMO. Even the "lowly" 17-40L does a great job when it comes to landscapes.
For sure! I can use a pinhole in a lenscap too! The 17-40 is no lowly lens and is a great choice. However, I'm interested in detail-rich landscapes not "landscapes", LOL! I need to use the same lens for stitching for architecture indoors and outdoors.

I am impressed that the center of this lens is very sharp and according to the FM review, is outstanding compared to the 16-35, although the 17-40 wasn't mentioned. If you have a link to a 17-40 comparison, that would be helpful for my quest.

Asher
__________________
Follow us on Twitter at @opfweb

Our purpose is getting to an impressive photograph. So we encourage browsing and then feedback. Consider a link to your galleries annotated, C&C welcomed. Images posted within OPF are assumed to be for Comment & Critique, unless otherwise designated.
Reply With Quote
  #10  
Old July 7th, 2009, 03:47 PM
Kathy Rappaport Kathy Rappaport is offline
Senior Member
 
Join Date: Aug 2006
Location: So. California
Posts: 1,793
Default Tilt Shift

Right now the Tilt Shift lenses are very trendy for portrait work along the lines of using the tilt/shift for an upmarket lensbaby. I was able to use one at a workshop put on by Canon a couple weeks ago and they really do have a learning curve.
__________________
You can call me ChatKat
********************
I created this piece of fine art. It's Fine Art because it's mine, I made it and I say it's fine art...
Reply With Quote
  #11  
Old July 7th, 2009, 03:55 PM
Cem_Usakligil Cem_Usakligil is offline
Senior Member
 
Join Date: Jun 2006
Posts: 5,947
Default

Quote:
Originally Posted by Asher Kelman View Post
For sure! I can use a pinhole in a lenscap too! The 17-40 is no lowly lens and is a great choice. However, I'm interested in detail-rich landscapes not "landscapes", LOL! I need to use the same lens for stitching for architecture indoors and outdoors....
Asher sorry to say this but this comment of yours is totally off the mark imo. I use the 17-40 and i have very detailed and sharp landscapes to show for it, no need to wave them off as just landscapes. Almost all my recent pictured which you have enjoyed are taken with that lens. fwiw...
__________________
Kind Regards, Cem

flickr
website
Reply With Quote
  #12  
Old July 7th, 2009, 03:58 PM
Bart_van_der_Wolf Bart_van_der_Wolf is offline
OPF Moderator
 
Join Date: Jun 2006
Location: Netherlands
Posts: 4,046
Default

Quote:
Originally Posted by Cem_Usakligil View Post
Hi Asher,

You do not necessarily need this lens for the purposes you state IMO. Even the "lowly" 17-40L does a great job when it comes to landscapes. If Bart doesn't mind me doing so, I can show an example from the same locality (since I was there with him during this shoot ;-)). Or I can send it to you by PM, whatever.
Hi Cem,

I have no problem with a comparable image taken with another lens, as long as it is informative with respect to the corner performance. That corner performance is probably what people are interested in when they read this thread.

Quote:
@Bart: This is yet another fantastic thread, full of good info as we came to expect from you time after time. But we should not take it for granted. I, for one, appreciate a lot the valuable knowledge you share with us. Thx.
Thanks, you're welcome.

Bart
Reply With Quote
  #13  
Old July 7th, 2009, 04:12 PM
Asher Kelman Asher Kelman is offline
OPF Owner/Editor-in-Chief
 
Join Date: Apr 2006
Posts: 31,748
Default

Cem,

No, I'm not dismissive of your remark on the 17-40. If you have some images that would be good to see what the corners look like at different apertures. I have already put this lens on my wish list. It's a very versatile design as a walk around lens and lightweight too. However, I'm interested in getting the most resolution in architectural scenes too.

Bart,

I'm impressed with both your work and the pictures in the FM review you referenced. There the 24mm TS II outshines the others tested in the corners and the center. That's important to me in knowing what will be overlapped in each stitch or what can be used alone. I can see that the lens is a fine optic. If it's a good as it seems at the center, it will already be an asset. I could either use it as a TS for architecture with no stitching or on my Gigapan just using the center. My impression, so far is that it appears much sharper than the competing lenses except, perhaps, for the 17-40 which is not shown.

I'm enjoying your test pictures!

Asher
__________________
Follow us on Twitter at @opfweb

Our purpose is getting to an impressive photograph. So we encourage browsing and then feedback. Consider a link to your galleries annotated, C&C welcomed. Images posted within OPF are assumed to be for Comment & Critique, unless otherwise designated.
Reply With Quote
Reply

Bookmarks

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are Off
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
News: Canon announces new 500D Ivan Garcia Canon Eos Mount DSLRs 10 May 1st, 2009 10:55 PM
FS, UK various Canon Items Ivan Garcia Buy and Sell Photo Equipment: Excess gear by participating members. Pictures please! 1 September 23rd, 2008 04:49 PM
Canon 1D Mark III test report by Arthur Morris, Online! Asher Kelman Breaking News 9 July 9th, 2007 06:17 PM
Canon 24mm f/1.4L softness - and a clue! John Maio Lenses: DSLR and Rangefinder, MF adaptions to 35mm such Zoerk 5 January 13th, 2007 12:22 AM
Desire to write a new, free raw converter Daniel_Hyams Image Processing and Workflow 13 August 1st, 2006 01:43 PM


All times are GMT -7. The time now is 10:31 AM.


Posting images or text grants license to OPF, yet © of such remain with its creator. Still, all assembled discussion © 2006-2014 Asher Kelman (all rights reserved) Posts with new theme or unusual image might be moved/copied to a new thread!