Tilt Shift Lenses - Looking for Advice

[Hector1970]

Hi All,
I'm really tempted to save all my money and buy a Canon TS-E 24mm II tilt shift lens.
Alot of people like the 17mm version and there is a 45mm and 90mm version.
I had a few questions.
a) Is 24mm more useful than 17mm or would the use of an extender actually make 17mm more useful
b) What would you use a 45mm TS-E for - is it designed with a particular type of photograph in mind
c) What would you use a 90mm TS-E for - A few of the minature videos or photographs I've seen have been done with this focal length but I assume thats not the real purpose of that focal length

What would I be using if for - I love sharpness I would like to achieve back to front sharpness for landscapes.
I might do a bit of architectural photography.
I do like the tilt (minature) effect (although its got quite common and the examples are getting worse).
If you use an extender with a 17mm is it then a bit restrictive as the maximum aperture is reduced?

I'd be interested in any advice.
The 24mm TSE would be my favourite at the moment.
I have a 50mm Flektogon Pentagon 6 on a Arax tilt mechanism.
It only tilts - it's like a hard to control Lensbaby.
I'd love a real tilt/shift lens.
I'd be interested too in those who have a tilt shift lens.
They are so expensive - do you actually use it alot afterwards.
It's a bit like a holy grail object.
I was like that with the 70-200 IS II - it even exceeded my expectation and I'm delighted I saved up for it.
Will I have the same sort of moment with a tilt shift lens.

 

[Menace]

Would it be possible for you to rent it for the week end and really give it a good go? See how you like it.

 

[shutterwideshut]

Would it be possible for you to rent it for the week end and really give it a good go? See how you like it.

I Agree. Tilt shift lenses are not for everyone. These lenses are manual focus, non zoom lenses which require care and attention to use. These lenses require slower workflow than the usual auto focus zoom/prime lenses. It's much better for you to rent and see for yourself if you and ts-e lens can get along in the real world.

But if I were to suggest one focal length, I would suggest you take the Canon TS-E24mm f/3.5L II.

 

[Hector1970]

I've no option to rent. I envy those who can. I'm used to manual lenses I have a few already.

 

[rs]

The relatively new 17 and 24 II both have independent rotation of tilt/shift axis, which makes them much more useful that the previous generation TS-E lenses. While landscape and architecture can be taken with a wide range of focal lengths, usually you're looking at the wider end for those applications. The 24 TS-E is optically the best of the bunch, while the 21 year old 90 TS-E is also optically excellent, although it is typically used for product photography. The 17 TS-E isn't all that sharp wide open, and doesn't respond too well to TC's. Even stopped down to f11 with the 1.4x TC, its nowhere near as sharp as the 24 TS-E II is wide open:

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=487&Camera=453&Sample=0&FLI=1&API=4&LensComp=486&CameraComp=453&SampleComp=0&FLIComp=0&APIComp=0

Obviously ultimate sharpness is less important than having the correct focal length for your application, but the 24 TS-E is possibly the pick of the bunch for your requirements.

 

[Peter Hill]

http://www.luminous-landscape.com/forum/index.php?topic=63827.5
"Once you shift or tilt a lens you have moved the entrance pupil and this will create all sorts of parallax and spatial rendering problems when it is time to line up frames.* A mount that lens you move the camera body instead of the lens solves the problem but there are only a couple of very expensive alternate lenses that allow you do this directly with a DSLR."

I'm sorry but that particular comment is referring to shifting a normal lens, when a rail for panos is definitely required to avoid the different perspectives from moving the camera. I'm not sure you are getting the idea about what the shift function on a tilt+shift lens actually does. The camera stays immobile, but the lens element is shifting across the sensor and utilising the full image circle. You manually focus your first shift shot. On shifting the lens for the next shot you do not touch the focusing ring as the focus remains constant.

 

[Vivid Color]

Thank you for your extensive post, Peter.

I had a problem with some of your links. Removing the punctuation at the end fixed it.

Thanks for the heads up. I edited the post to remove the full stops. Doh.

I tried the edited links and still could not access your pages so something may have happened in the editing. Thank you for posting and I would really like to see your posted pics and articles.

 

[Loswr]

I'm sorry but that particular comment is referring to shifting a normal lens, when a rail for panos is definitely required to avoid the different perspectives from moving the camera. I'm not sure you are getting the idea about what the shift function on a tilt+shift lens actually does. The camera stays immobile, but the lens element is shifting across the sensor and utilising the full image circle. You manually focus your first shift shot. On shifting the lens for the next shot you do not touch the focusing ring as the focus remains constant.

Sorry, Peter, but you're incorrect. A shift-pano with a TS lens, where the camera isn't moved but the front of the lens is shifted relative to the camera, will result in parallax. Whether or not you can see that in your image depends on the relative distances to and between the subjects in the image. Old view cameras shifted the camera back with the lens fixed in place - no parallax. TS lenses used as you describe shift the lens with the camera fixed in place, and that means parallax. To avoid that, you need to either fix the lens in place for an optimal solution (DIY collar or the expensive Hartblei one), or use a macro rail to move the camera an equal-but-opposite distance as you shift the lens.

It's also worth mentioning that for architecture shots using a shift pano, for example, if 24mm isn't wide enough to capture the full height of a building, negates perspective correction - and that may be the very reason you used a TS lens. So, to the OP - the primary determinant of 17mm vs. 24mm should be the focal length you need. A sharper lens (TS-E 24L II) doesn't help if you have to cut off the top and sides of a building.

Personally, I went with the TS-E 24L II as that focal length works best for my style.

 

[rs]

The shift function on the 24mm TS-E II allows for up to 12mm shift in any direction. Although the corners will be missing, you can stitch together multiple images from a 36x24mm sensor to cover the area covered by a 60x48mm sensor (minus the corners).

If a single image was captured by such a large sensor behind that lens (which, after all, is what you're trying to simulate), the lens would obviously be in one fixed place. The sensor is the bit which covers all of those places the conventional FF sensor can't reach.

So the only logical way to fully replicate that with stitching shifted images together is to keep the lens fixed and move the sensor around. Anything else is just an approximation.

 

[Loswr]

Any idea why ts lenses does not have integrated collars, so this could be done by simply attaching the lense to a stand, then shift the camera around?

Probably because the main intention of shift is to correct perspective, and in that case, you want to shift the lens relative to the camera. As I stated above, using shift to create a panoramic shot negates the ability to correct for perspective in that shot.

 

[rs]

So the only logical way to fully replicate that with stitching shifted images together is to keep the lens fixed and move the sensor around. Anything else is just an approximation.

Thank you.

Any idea why ts lenses does not have integrated collars, so this could be done by simply attaching the lense to a stand, then shift the camera around?

Would there be a use for a 60x48mm MF camera, specifically designed around the Canon ts-lenses (corners would be bad, but you would have maximal area for subsequent cropping).

-h

No point in creating an EF mount camera with such a large sensor - first of all, the corners would be out of the imaging circle of the TS-E lenses, secondly it couldn't have a mirror as one that big wouldn't have room to flip, and thirdly the only lenses which would half work would be the TS-E lenses, and this would render the shift part of their name useless. Just get MF for anything like that.

As for the lack of tripod collars, I guess its partly a design which doesn't offer every single feature its users are demanding (but as Neuro said, there are options out there, home made or third party), and stitched panoramic photos aren't the main intended purpose of these lenses. After all, Canon do make a 14mm lens and an 8-15mm fisheye.

 

[Random Orbits]

Any idea why ts lenses does not have integrated collars, so this could be done by simply attaching the lense to a stand, then shift the camera around?

If you're going to do much stitching, it's a much better idea to get a pano rig than a lens collar. The rig will allow you to use all your lenses for panoramas.

 

[Hector1970]

Thanks for your replies everyone. They are really interesting and whetting my appetite.
Your photos Peter are very impressive, it's a great advert for TS-E lens.
Congratulations on your rewards well deserved.
A tough call between 17 and 24mm.
Maybe 24mm would be more flexible for me.
I will be very interested in the Samyang version.
I haven't seen any hands on reviews yet only announcements.
I think it's supposed to be around March.
I have their 14mm and 8mm lens and they pretty good for the price they are.
The 14mm is very good for wide field astrophotography.

I've read that the 24mm TS-E is really easy to break by overtightning.
Is this a concern or would you want to be really doing something stupid to break it?
I believe it you over tightnen the knobs and they break it's a big and expensive job to repair the lens.

 

[rs]

Yes, but you would have an optimal starting-point for cropping the final image (lens limited rather than sensor limited).

Why use such a dedicated and single purpose camera, just to crop later on? That is all a TS-E on a bog standard Canon FF camera does, and it does it in camera.

It is an impressive camera, but I'm failing to understand why it was built, other than someone using the ethos 'because I can'.

Hector1970 - if you have any other lenses covering the 17-24mm range, try using those in your expected photo situations to work out which gives better framing for you. If you don't have any lenses like that, try out TrumpetPowers cardboard cut out trick (comment borrowed from another thread):

There's an easy way for a photographer to get a mental grasp on the meaning of focal length.

Take a sheet of cardboard and cut a hole in it the same size as your camera's sensor (or film or whatever). Hold it 50mm away from your eye, and you'll see exactly the same field of view as your camera captures with a 50mm lens. Hold it 24mm away and you see what you would with a 24mm lens.

You can even shift the cardboard off-axis to simulate the shift effect

 

[rs]

Why do people use FF instead of crop, or MF instead of FF? Because they think that larger sensor area has some advantage, I guess.

Why do people use shift lenses to capture stitched panoramas? Partially because they think that larger sensor area has some advantage, I guess.

Why would people want to connect a Canon TS-E lense to a camera of larger than 24x36mm sensor? Because they think that larger sensor area has some advantage, I guess.

There is an awful lot of image circle that cannot be used (in one shot) of a FF camera:
-h

Yes, it clearly works, and is capable of producing good results. But then you've got a very expensive camera with only two lenses available, each manual focus with tilt as the only usable function. Why not just use a MF back/body with MF lenses (which are numerous), and many are available with tilt and shift working on that large sensor. The Canon lenses are more than up to the job of resolving well for such as sensor, but I still wonder why bother when production MF lenses, bodies and backs all exist.

If you really want to use a Canon TS-E lens to produce what looks like an unshifted large format wide angle panoramic image, just do what we've already been talking about - mount the lens to a tripod and shift the Canon FF body, and stitch the images together. It's no good for action shots, but the Hartblei camera isn't either.

 

[Loswr]

Why do people use shift lenses to capture stitched panoramas?

I'm not at all sure that most people buy TS lenses for the primary purpose of capturing stitched panorama. While a TS lens can be used in a limited way for pano shots, it seems more of an opportunistic and incidental use. As stated above getting a real multi-row pano setup offers much more flexibility for creating panoramic images.

 

[KyleSTL]

...The 17 TS-E isn't all that sharp wide open, and doesn't respond too well to TC's. Even stopped down to f11 with the 1.4x TC, its nowhere near as sharp as the 24 TS-E II is wide open:

In comparison to the 24mm the 17mm is slightly less sharp, however, the 17mm TS-E is widely considered to be one of the best ultra-wide lenses of all time and easily out-resolves any Canon or Nikon UWA zoom or prime. So your comment is very misleading. Both lenses are among the very best in there respective focal lengths, and what should be discussed is whether a wide angle or UWA FL is needed. The ability of the 17mm to act as a 24mm TS with the TC is just a perk, if 24mm is the FL that is needed, then the obvious choice is the 24mm II.

 

[rs]

...The 17 TS-E isn't all that sharp wide open, and doesn't respond too well to TC's. Even stopped down to f11 with the 1.4x TC, its nowhere near as sharp as the 24 TS-E II is wide open:

In comparison to the 24mm the 17mm is slightly less sharp, however, the 17mm TS-E is widely considered to be one of the best ultra-wide lenses of all time and easily out-resolves any Canon or Nikon UWA zoom or prime. So your comment is very misleading. Both lenses are among the very best in there respective focal lengths, and what should be discussed is whether a wide angle or UWA FL is needed. The ability of the 17mm to act as a 24mm TS with the TC is just a perk, if 24mm is the FL that is needed, then the obvious choice is the 24mm II.

Yes, you're right. The 17 TS-E is pretty much without peers at its focal length. But to call my post misleading is wrong - I did end the post by saying that the most important factor isn't sharpness, but choosing the right focal length for your application - which you conveniently missed out of your quote.

 

[noisejammer]

I think I can comment on using a tc with a TS-E 17. My tests yielded an unexpected result.

I tested my TS-E 17 plus a 1.4x Mk III against my friends TS-E 24 II. I stopped both down to f/8 effective - in other words the 17 was stopped down to f/5.6 while the 24 was stopped down to f/8. I used magnified live view to focus at the center of the frame, and then magnified live view to focus at 2/3 of the way out and in the corners.

My results
1. In the center there is no difference at all.
2. At the 2/3 position, there was no difference at all.
3. In the corners, the 17/1.4x combination was BETTER. If memory serves, point sources were rendered on half the pixels.
4. There was no difference in chromatic aberration between the two systems.
5. Focusing precisely enough to extract everything from this lens is extremely difficult - even magnified live view with a 3x loupe left me with some uncertainty.

Ok, so I've looked at the results on TDP.

Bryan and I disagree on our results however Bryan was shooting a flat calibration target. I refocused the lenses to eliminate the effect of focus curvature. My conclusion is that if was are shooting a brick wall, the 24 TS-E may well be sharper. If I was shooting something else, my results would depend entirely on how well the subject matches the focal curvature of the lens.

The other caveat is that I had exactly one 17mm, one 1.4x tc and one 24mm. I do not have any idea of whether this is representative or not. Nevertheless, I was convinced enough that I decided I did not need to purchase a 24 TS-E and picked up a 25/2 Zeiss instead.

Other thoughts
6. Using the TS-E 17 at f/8 is not an issue - if you are shooting a landscape, you are probably going to use f/8 to f/16 anyway.

7. Although I've found a way to use my Lee filters on my TS-E 17, this comes at the price of reduced shift before I hit vignetting. If memory serves, I can move the lens about 6 mm each way so that the effective image is about 48x24 mm (115 deg diagonal field of view) with the filters on. Fotodiox has a solution for this but I'm not willing to buy (or carry) yet another set of filters.

8. On reflection, I have not tried to see how the tc responds when the lens is exercised in shift. I'll try that when the lens comes back from Canon.

9. I strongly recommend you secure the adjustment knobs with a blob of nail polish / thread locker. Loosing one can be inconvenient... been there, done that.

10. Using the 17mm requires a lot of discipline. The front element is entirely unprotected and replacing it is expensive.

 

[KyleSTL]

...The 17 TS-E isn't all that sharp wide open, and doesn't respond too well to TC's. Even stopped down to f11 with the 1.4x TC, its nowhere near as sharp as the 24 TS-E II is wide open:

In comparison to the 24mm the 17mm is slightly less sharp, however, the 17mm TS-E is widely considered to be one of the best ultra-wide lenses of all time and easily out-resolves any Canon or Nikon UWA zoom or prime. So your comment is very misleading. Both lenses are among the very best in there respective focal lengths, and what should be discussed is whether a wide angle or UWA FL is needed. The ability of the 17mm to act as a 24mm TS with the TC is just a perk, if 24mm is the FL that is needed, then the obvious choice is the 24mm II.

Yes, you're right. The 17 TS-E is pretty much without peers at its focal length. But to call my post misleading is wrong - I did end the post by saying that the most important factor isn't sharpness, but choosing the right focal length for your application - which you conveniently missed out of your quote.

Fair enough, I didn't include your entire post in my quote, however, this part of your comment is factually incorrect:

The 17 TS-E isn't all that sharp wide open