Diagonal tilt - Ts-e lenses. Issue with panorama shift?

The PocketPANO frame has two tripod feet offset by 90°. Their function described at the homepage as this:

“...In combination with the two rotation elements of the TS-E lenses, all imaginable combinations of portrait and landscape format, horizontal and vertical shifting as well as tilting can be realized.”

https://www.pocketpano.de/english/rear-tilt-n-shift-frame/

Not quite sure but I think this solves a limitation of the Canon Ts-e lenses regarding diagonal tilt in combination with shift for panoramas/vertoramas.

I haven’t seen the issue fully described anywhere online. And it is a bit complicated to describe, but I give it a quick try: If you take 3 shifted shots for a panorama one only has the option of the diagonal tilted focus plane going left down or right up. If one wants the planes to go left up or right down, one has to rotate 90° at shifting rotation axis – and this rotation gives you a vertical vertorama instead of a panorama format in the 3 shots for stitching.

I suppose the way around it without the PocketPano frame is to make the 90° rotation at the tripod head, overhanging the camera to the left or right to the side of the tripods centre axis (and adding some instability to the tripod setup).

I might be wrong, diagonal tilt is rather difficult to get ones head around 😊

Here a link to one of Keith Coopers excellent videos on the topic: https://www.youtube.com/watch?v=HcwM8P6SNKw&ab_channel=KeithCooper

Other links visualizing the aspects of diagonal tilt on the Ts-e lenses much appreciated.

Did some tests, and I actually can't get a 90° rotation on the tripod head to solve the issue. I suppose that also rules out the PocketPANO solving it.

I suppose the way around it without the PocketPano frame is to make the 90° rotation at the tripod head, overhanging the camera to the left or right to the side of the tripods centre axis (and adding some instability to the tripod setup)..

I will take some test photos and post them later to illustrate the problem (which surely might be me not being able to get my head around the possible combinations of the lens rotation axes)

Here an attempt at a photo set up to illustrate the issue.

The red gradients approximate the tilted diagonal focus planes in between the tape lines (sorry just a quick setup)

The first 4 shots are single frame. And all is well and fine. All the 4 four diagonal directions of tilt are covered – by rotating the tilt axis ca. 45°, and importantly for my issue by rotating the shift axis by 90°:

But here comes the issue. When taking 3 photos with shift for a panorama, I can not rotate the shift axis the 90° (as it of course is in use for the panorama). And I end up with only 2 of the 4 possible directions for diagonal tilt:

Am I wrong? Or is this a limitation of the tilt axis on the Ts-e only being able to rotate 90°?

I am aware that using diagonal tilt in panorama may seem an unusual feature, but personally I like it a lot for creative purposes. And it is of course important to know what is possible and what’s not when standing in field trying figure out a shot. The diagonal tilt and the combinations of rotations on the ts-e lenses sure is complicated.

Well I hope I am wrong, and hope that someone here can show me the solution, in simple and practical terms (I have given up on the deeper math of Mr. Scheimpflug😊) .

I'd never bother with tilt and one of the lens holder contraptions, yet alone trying it for panoramic use.

Using tilt and stitching has a lot of potential gotcha's - with very few instances I've ever found a [good] use for it.

One day I'll find a really good example which needs it and use it for an article/video [suggestions welcome]

For basic tilt setting you just need to ignore the click stops...

I've an article about setting tilt on an arbitrary plane at

https://www.northlight-images.co.uk/setting-the-tilt-axis-for-tilt-shift-lenses/

also a video looking at it

Oh, and my book...

Keith I have read through all your articles on tilt and shift on your absolutely great homepage. Your articles and your videos on the topic for sure are the most clear and detailed online (and I will buy your book, promise 😉)

My purpose for tilt and panorama I suppose is basically creative. The tilt introduces an atmospheric sense of space which I find very interesting, and creatively highly useful - I am totally aware many will find it rubbish (used to that as a professional artist, mainly sculptural stuff and non-interest by most people is simply a part of the trade).

But that aside: from a purely technical point of view couldn’t it interest you if the Ts-e lenses actually have a 50% limitation in their use of diagonal tilt in panorama photos? Only being able to utilize 2 out of 4 angles? Just as a technical curiosity? Probably not enough for a whole video, but just as an interesting appendix on the limitations of the lenses which haven’t been described elsewhere?

Well, I will give it some more effort with the knobs and rotations axes trying to figure out if I am stupidly wrong or if the lenses actually have this limitation.

Why is there a limitation?

What is wrong with just rotating the tilt axis by 20º or 17º as desired?

Sure, it's harder to replicate, but I still don't really see what it is you're trying to do that can't be done?

Would one of these help?

https://www.northlight-images.co.uk/ppl-tse-adapter-lens-shift-review/

Not sure where you can get them any more...

The limitation I am struggling with only apply for panorama or vertorama photos made by multiple shots using shift.

In a single shot (with no intention for stitching to panorama) I have no problem getting any of 4 possible diagonal planes focused.

As far as I can see what is going on is this: The tilt rotation axis can rotate 90°, stetting the rotation around 45° and tilting down results in upper left diagonal focus plane. Tilting up results in lower right diagonal focus planes. Focus planes marked in red:

To get the two other diagonal planes in focus I have to rotate the shift axis 90°:

But the rotation of the shift axis 90° turns the shifting direction from horizontal to vertical. And this is of course not an option if the intention is to stitch photos to a final horizontal panorama.

Setting the tilt axis rotation less or more than 45°, as far as I can see, only refines the diagonal rotated plane within the total of the 90°. The 10° being closer to plain tilt, and 80° closer to plain swing.

The total image circle is 360° and using the tilt rotations axis of 90° in conjunction with “up/down” covers 180°. To cover the remaining 180° the shift axis has to be rotated 90° - this turning the shift direction from horizontal to vertical.

Hope this is a bit clearer, and sorry if my language and vocabulary is of - English is not my first language.

Martin Erik Andersen wrote:

The limitation I am struggling with only apply for panorama or vertorama photos made by multiple shots using shift.

In a single shot (with no intention for stitching to panorama) I have no problem getting any of 4 possible diagonal planes focused.

As far as I can see what is going on is this: The tilt rotation axis can rotate 90°, stetting the rotation around 45° and tilting down results in upper left diagonal focus plane. Tilting up results in lower right diagonal focus planes. Focus planes marked in red:

To get the two other diagonal planes in focus I have to rotate the shift axis 90°:

But the rotation of the shift axis 90° turns the shifting direction from horizontal to vertical. And this is of course not an option if the intention is to stitch photos to a final horizontal panorama.

Setting the tilt axis rotation less or more than 45°, as far as I can see, only refines the diagonal rotated plane within the total of the 90°. The 10° being closer to plain tilt, and 80° closer to plain swing.

The total image circle is 360° and using the tilt rotations axis of 90° in conjunction with “up/down” covers 180°. To cover the remaining 180° the shift axis has to be rotated 90° - this turning the shift direction from horizontal to vertical.

Hope this is a bit clearer, and sorry if my language and vocabulary is of - English is not my first language.

OK, I can see this limitation with the TS-E mount if you want the same shift direction (and why the internal cabling causes it)

My only solution would be to use a shift adapter to replace the standard EF->RF one I use - of course it needs electrical connections. No use of course on my native EF mount DSLRs

For my adapted (M645) lenses, with no electrical connections, it's not an issue, but then again I've no parallax cancelling mount for them, so it's a matter of manual camera shift to offset the shift of viewpoint.

Other expensive solutions spring to mind (involving adapters and medium format backs...

Fortunately (well, for me anyway   ), I have very little use for tilt when shooting stitched wide angle shots.

Maybe I'll try some experiments with my assorted adapted lens solution (35-210mm)!

Thanks Keith. Kind of nice to know this limitation actually is in the Ts-e lens - and not in me not being able figure out the settings. That is to say that I don't have to try to do something that is not possible out in the field.

Now what to do about it...can probably get very expensive very quickly. Well, all gear has limitations....and I find a lot to like in the Ts-e lenses....can't afford any upgrades right now, but I am surely tempted...

Thanks a lot. Shift direction in 45° for stitching panorama haven't crossed my mind

Setting the camera up on the tripod like this sure gives the possibility of getting the diagonals right for focus in any of the 4 directions:

I will give it a full try in a couple of days. And see how much cropping in the final stitched panorama it will result in compared to the plain horizontal camera orientation and 90° shift direction.

I have now tried the 45° stitch direction of five shots in a cross pattern. But after cropping with surprisingly little gain in the final panorama format compared to a single shot. Actually surprisingly next to nothing:

I could probably have been more precise in getting the 45° correct (the clicks at shift axis being at 30° and 60°), and then having to crop less, but still I suppose not terrible much would have been gained.

A bit better than a 3 shot stitch, but again not much compared to a single non-stitched shot:

As far as I can see getting more of the image circle into play would require a 90° rotation of the shifting direction, and that would again change the possible tilt focus planes to the opposing 2 possible diagonals. And the additional photos would not be able to stitch into a single photo as the tilted diagonal focus planes would not match.

I sure might be doing something wrong. The 45° degree shift direction for panorama certainly ads to the complexity of it all.

But anyway interesting to make exercises at a primitive test setup – I am getting more familiar with the settings and the possible focus planes, it will be useful when making decisions in real life. Hereby recommended 😊

Hi Brett, I think what you suggest only will give you 2 out of the 4 possible directions for diagonal tilt.

The issue is doing the shots for the stitch with consistent diagonal tilt direction - in any of the four possible diagonal directions. That is to say in four different variations of the same panorama view.

"Step 4: Turn lens clockwise 90 degrees on shift axis"

With your step number 4 turning the 90 shift axis the diagonal tilt direction also rotates to an opposing angle. This results in mixed diagonal tilted directions in the photos for the panorama stitch.

I suppose it might stitch together if the contrast between focused and unfocused areas isn't too strong, but it would then be an artificial digital blend of different diagonal angles of focus.

Well, difficult to describe in writing, easier to see it in praxis. But for the experiment one has to try to do the stitched material for all four diagonal directions - for the four different versions of diagonal focus of the same panorama view. It takes some time, but it is fun trying 😊

For a stitched horizontal panorama I still think only these 2 diagonal tilted focus planes are possible:

And these 2 directions are not possible:

One can of course make the less wide panorama format, by 3 stitched vertical shots. This will result in the mirrored options of diagonal focus planes.  Cropping this version down to the wider format could be said to limit the limitation.

(but I certainly will not rule out I am wrong, I have before been getting stuck in repeated errors in multiple variations )

I am aware the issue I am trying to get a hold on here may seem a very minor limitation.

But I have been out trying to get the shots for a landscape panorama while the light is fading and the rain clouds are setting in, and,,, "why the heck can't I get the focus plane to be where I want it to be?.." If I had known there were actual limitations in the options of diagonal tilt, I could more quickly have made compromises in for example distance or composition ...instead of fumbling with the knobs till the light was gone.