Pages: [1]   Go Down

Author Topic: Canon Tilt-Shift Lens TS-E 24/3.5L Version II article  (Read 4552 times)

Luis Argerich

  • Full Member
  • ***
  • Offline Offline
  • Posts: 155
  • Astrolandscaper
    • http://www.luisargerich.com/
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« on: May 05, 2010, 07:11:52 PM »

I've been trying to diggest this article from the front page:

http://www.luminous-landscape.com/reviews/lenses/24-ii.shtml

I wanted to email the author but I'd probably learn more posting it here as he and many others might be reading.

Is the article technically right?

What I think:

I you want the image in focus from the FG to infinity the angle of tilt depends on the height of the camera.
Tilting all the way forward is not the way to achieve maximum DOF unless you have the camera very close to the floor.
If you get the right tilt for the camera height you can achieve focus from FG to infinity without needing to focus beyond infinity.

Is this correct?

Cheers,
Luis

Kirk Gittings

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1555
    • http://www.KirkGittings.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #1 on: May 05, 2010, 07:27:43 PM »

Yes that is essentially correct. Whether with view cameras, which I used exclusively for almost 30 years, or DSLR tilt shift lenses, I rarely need more than a slight tilt (1/4 inch or less on a DSLR T/S lense) to get everything in a landscape sharp (assuming everything is basically on approx. a flat plane) UNLESS the foreground is right under the lens, where I need more severe tilt.
« Last Edit: May 05, 2010, 07:33:33 PM by Kirk Gittings »
Logged
Thanks,
Kirk

Kirk Gittings
Architecture and Landscape Photography
WWW.GITTINGSPHOTO.COM
LIGHT+SPACE+STRUCTURE (blog)

Kirk Gittings

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1555
    • http://www.KirkGittings.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #2 on: May 05, 2010, 07:31:19 PM »

duplicate
« Last Edit: May 05, 2010, 07:31:59 PM by Kirk Gittings »
Logged
Thanks,
Kirk

Kirk Gittings
Architecture and Landscape Photography
WWW.GITTINGSPHOTO.COM
LIGHT+SPACE+STRUCTURE (blog)

JeffKohn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1668
    • http://jeffk-photo.typepad.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #3 on: May 05, 2010, 08:00:14 PM »

I don't think he stated anything incorrect (the guy knows his stuff when it comes to focus/tilts/etc, he's literally written books on the subject).

Reading the article I don't think he was trying to achieve focus focus through the entire frame his full-tilt examples, although I have to confess I'm not exactly sure what he _was_ trying to achieve. Nor do I see how the full-tilt 24 TS-E examples related to the 5x7 view-camera example. He may have needed to focus past infinity in the 5x7 shot, but that was presumably with a much longer focal length where greater tilts are needed. In the 24 TS-E shot, focusing past infinity to have only the top part of the driveway in focus doesn't strike me as a useful outcome.

uaiomex

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1190
    • http://www.eduardocervantes.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #4 on: May 05, 2010, 11:51:02 PM »

At the very end he said "But, take special care if you use a tripod!"
I don't quite understand why he refers to tripod use in this way.
A tripod gives more control of course and to certain points makes things easier.
But it sounds like a tripod makes things even more complexed.
Is there something I missed?
Eduardo
Logged

daws

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 282
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #5 on: May 06, 2010, 12:30:36 AM »

Quote from: uaiomex
At the very end he said "But, take special care if you use a tripod!"
I don't quite understand why he refers to tripod use in this way.
A tripod gives more control of course and to certain points makes things easier.
But it sounds like a tripod makes things even more complexed.
Is there something I missed?
Eduardo
I haven't read the article, but I'd guess he was referring to being mindful of getting your tripod legs in the bottom of the frame when you tilt, especially with WA T/S like the 17mm.

If handheld, your feet. (In my case, your paunch. Don't laugh -- been there, cropped that.)

Logged

Faintandfuzzy

  • Newbie
  • *
  • Offline Offline
  • Posts: 21
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #6 on: May 06, 2010, 08:39:14 AM »

I found his comment about color fringing interesting....that most of it was produced b the printer and not the lens.  That is the first time I've ever heard of a printer producing color fringing.  If there is color fringing showing, it's from the lens and nothing else.
Logged

uaiomex

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1190
    • http://www.eduardocervantes.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #7 on: May 06, 2010, 09:56:19 AM »

LOL, been there too.

Quote from: daws
I haven't read the article, but I'd guess he was referring to being mindful of getting your tripod legs in the bottom of the frame when you tilt, especially with WA T/S like the 17mm.

If handheld, your feet. (In my case, your paunch. Don't laugh -- been there, cropped that.)
Logged

Luis Argerich

  • Full Member
  • ***
  • Offline Offline
  • Posts: 155
  • Astrolandscaper
    • http://www.luisargerich.com/
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #8 on: May 06, 2010, 10:22:49 AM »

I'm not ready to believe the article is technically wrong so I wonder if someone can explain the idea or maybe contact the author for an explanation?

BartvanderWolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6699
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #9 on: May 06, 2010, 11:51:55 AM »

Quote from: Luis Argerich
I'm not ready to believe the article is technically wrong so I wonder if someone can explain the idea or maybe contact the author for an explanation?

It is not likely that Harold Merklinger is wrong, but he did make a remark that I find puzzling.

For instance, he mentions "Using Adobe Camera Raw or Photoshop to "fix" the fringing in one part of the image makes it worse in another". To me it sounds like he didn't add some canvas space to the image before applying the otherwise centered corrections. I find that hard to believe, but I have no other explanation.

He then mentions: "We apply full forward lens tilt. This brings the plane of sharp focus up just under the camera just six inches below the lens in fact. The angle the plane of sharp focus takes with respect to the film/sensor plane is now controlled by where the lens is focused. With the lens focused at infinity, and with the film/sensor plane vertical, the plane of sharp focus will angle slightly downward away from the camera, just at bit below the horizontal plane". That is mathematically impossible when the sensor plane is indeed vertical AND focus is set at infinity on the lens! However, when one sets focus 'beyond infinity' on the lens, then a virtual image is created and the tilt allows to focus on that.

The tilt also shifts the entrance pupil of the lens a bit in the direction of the tilt. That may be the cause of the diverging verticals he noticed, and it's a useful observation indeed.

Cheers,
Bart
« Last Edit: May 06, 2010, 11:55:58 AM by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2105
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #10 on: May 07, 2010, 03:56:08 PM »

Hi Bart, there is also the possibility that the lateral chromatic aberration on the 24 TS-E II is not well-approximated by a linear function of the image height. That is, the chromatic difference of magnification is better modeled using a higher-degree polynomial. (I have not personally studied the 24 TS-E II, so I cannot know for sure if this is the case, but I have studied a number of Canon wide angle lenses and found this to be true ... so it's a reasonable guess.) So even if the principal point coincided with the image center, Camera Raw's manual chromatic aberration controls may not be sufficient. That is, correcting a corner may cause the image areas 2/3 of the way between center & corner to become suboptimal, and vice versa.
Logged

BartvanderWolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 6699
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #11 on: May 07, 2010, 06:34:52 PM »

Quote from: madmanchan
Hi Bart, there is also the possibility that the lateral chromatic aberration on the 24 TS-E II is not well-approximated by a linear function of the image height. That is, the chromatic difference of magnification is better modeled using a higher-degree polynomial. (I have not personally studied the 24 TS-E II, so I cannot know for sure if this is the case, but I have studied a number of Canon wide angle lenses and found this to be true ... so it's a reasonable guess.) So even if the principal point coincided with the image center, Camera Raw's manual chromatic aberration controls may not be sufficient. That is, correcting a corner may cause the image areas 2/3 of the way between center & corner to become suboptimal, and vice versa.

Hi Eric,

I see, maybe he wasn't referring to a decentered CA but it varying by radius, he wasn't specific about that.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==

AJSJones

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 357
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #12 on: May 13, 2010, 08:03:06 PM »

Quote from: Luis Argerich
I'm not ready to believe the article is technically wrong so I wonder if someone can explain the idea or maybe contact the author for an explanation?

If you go to this page on his extensive web pages and open the various quicktime movies, you will get a feel for how "movements" like tilt affect the plane of focus.  Once you have understood the Scheimpflug and hinge rules, you'll be set for the T/S lens
Logged

JeffKohn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1668
    • http://jeffk-photo.typepad.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #13 on: May 13, 2010, 09:53:32 PM »

Quote from: madmanchan
Hi Bart, there is also the possibility that the lateral chromatic aberration on the 24 TS-E II is not well-approximated by a linear function of the image height. That is, the chromatic difference of magnification is better modeled using a higher-degree polynomial. (I have not personally studied the 24 TS-E II, so I cannot know for sure if this is the case, but I have studied a number of Canon wide angle lenses and found this to be true ... so it's a reasonable guess.) So even if the principal point coincided with the image center, Camera Raw's manual chromatic aberration controls may not be sufficient. That is, correcting a corner may cause the image areas 2/3 of the way between center & corner to become suboptimal, and vice versa.
Interesting. So out of curiosity, do the Lens Profiler-generated corrections handle this?

madmanchan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2105
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #14 on: May 13, 2010, 10:44:32 PM »

Hi Jeff, yes it does. I found this is more helpful on lenses with full-frame coverage, since the lens behavior is usually pretty consistent till you get near the farthest corners; then it can drop quite a lot. A simple scaling of channels for lateral CA fixups (and generic pincushion/barrel distortion correction) does not work well in these cases, because if you fix up the corners, then the areas in the middle image go bad, and vice versa.

I think the larger issue with the TS-E lenses in terms of profiling is that I'm not sure these lenses report the tilt & shift parameters electronically to the camera. Obviously these affect the optical configuration, and of course in the shift case which part of the image circle you're actually using. So in terms of processing an image to reduce residual aberrations, I don't know that there's an automatic way to "look up" the appropriate profile data based on the user's choice of tilt & shift parameters at the time of capture. I could be wrong about this (and would actually be very happy if that's true).
Logged

JeffKohn

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1668
    • http://jeffk-photo.typepad.com
Canon Tilt-Shift Lens TS-E 24/3.5L Version II article
« Reply #15 on: May 13, 2010, 10:56:40 PM »

Quote from: madmanchan
Hi Jeff, yes it does. I found this is more helpful on lenses with full-frame coverage, since the lens behavior is usually pretty consistent till you get near the farthest corners; then it can drop quite a lot. A simple scaling of channels for lateral CA fixups (and generic pincushion/barrel distortion correction) does not work well in these cases, because if you fix up the corners, then the areas in the middle image go bad, and vice versa.
Cool, good to know. This definitely provides some motivation for creating profiles rather than just using the manual controls, then.

Quote
I think the larger issue with the TS-E lenses in terms of profiling is that I'm not sure these lenses report the tilt & shift parameters electronically to the camera. Obviously these affect the optical configuration, and of course in the shift case which part of the image circle you're actually using. So in terms of processing an image to reduce residual aberrations, I don't know that there's an automatic way to "look up" the appropriate profile data based on the user's choice of tilt & shift parameters at the time of capture. I could be wrong about this (and would actually be very happy if that's true).
Yeah, there's no easy answer there. With the Nikon PC-E lenses the tilt/shift are completely mechanical, there's no metadata about lens movements recorded in the EXIF data. I'm pretty sure Canon T/S lenses are the same way.

I currently use PTLens for the PC-E lenses. What they did is to build a correction profile for the whole image circle (distortion only). They have a PS action you can run that expands the image canvas to the size of the entire lens circle. You drag your image to the appropriate place and then run the correction plug-in. This has some limitations: you have to remember to record how much shift you used, and you also have to approximate where on the canvas to place your image based on the shift amount. It's not perfect, but it does work to a certain extent as long as you're only using shift (it doesn't handle tilt, though).
Pages: [1]   Go Up