New Autofocus RF mount tilt-shift lenses expected after next round of camera announcements

[SwissFrank]

If we were happy with keystone correction and AI interpolation, then no one would be using TS lenses in the first place.

???

There's no way keystone correction and AI interpolation can replace tilt. Your post makes so little sense I wonder if it was just a cut and paste editing error or something? I'm sure you know the subject well and just deleted some words or something.

That said I've already argued on this thread that shift is pretty much useless. You end up paying FAR more for a FAR more complicated lens that casts a huge image circle, and the more you shift, the farther you get into the corners of the MTF diagram and the further you have to deal with cos^4 and mechanical vignetting and weird highlight shapes. Combining left-shift and right-shift images to make one wider image was a nice trick before digital editing of high-resolution sensor images was possible. Now you can just stitch images together. A few times I used shift to avoid the camera's reflection in a window but that was when film looked so comparatively bad you didn't want to crop half a shot away after using a wider lens from the same standpoint. Now you can crop 80% away and still have a great image.

Really, I've had the 24TS and 90TS for years. I'm at a loss why anyone would use shift in the modern era, but tilt is absolutely irreplaceable.

 

[SwissFrank]

It is useful when training in karate but not really conversational

Well, I had basically a full university degree's worth and it ended up getting me a few jobs but I didn't really need it to do the jobs. But Tokyo's a great city to live in and for daily life of course it's critical here, unlike say Zurich (where I lived many years) where I'm certain that 20% of the population uses English exclusively and don't know and have no need to know the local languages.

BTW I checked out your flickr and absolutely mind-blowing.

 

[blurryeyedbeast]

Or alternatively (and this is a shot you will see a lot of if these lenses/cameras are capable of doing it), imagine setting the focal plane on two pairs of eyes and a third point in a scene. Then both they and the camera move around the scene, but the focal plane remains absolutely locked on the individuals' eyes and the third point...

I don't know how this will look as I don't do video, but you are exploring ideas I hadn't thought of (object tracking is supposed to be a new feature in R1?).

And this is just speculation based on what we can imagine could be done, never mind what thousands of others dream up!

Maybe this will just stay dreams or maybe Canon will turn it into reality.

 

[Del Paso]

I had a couple M6TTL bodies (literally never put batteries in though), a .58 and a .85, and had 35/1.4 and 75/1.4. I really loved that outfit. Photogs love to say "it's not the camera, it's the lens" but in fact you could get such great people shots with an M6 because no-one takes it seriously due to silence and non-imposing nature. An M6 with a nice point/shoot lens would get great pictures I think. I shot 400 and developed in Emofin, then scanned the film or a friend would do a couple prints a month for me.

I never heard of anyone with the big Leica telephotos. I didn't even know they had shoulder grips. I'm in awe.

These 400mm and 560mm Telyts were originally developped for the M + Visoflex, later on also for the Leicaflex SLRs. They were quickly and easily focused by sliding forward or backward the rail-guided fronttube which contained the lenses. (No helicoids).

 

[EOS 4 Life]

1) I've said I've done it. You're welcome not to believe me but why would someone lie about that?

Who said you were lying?
I just gave an opinion.
Are you that insecure?

 

[DSB]

True but even a smartphone would need to stitch to be wide enough for the architecture shots we are talking about. It also doesn't have a lot of pixels to play with interpolation although if the end result is only for web then maybe it is okay.

That's exactly my point - phones do have enough pixels to play around, especially if the end target is social media. In fact, I've come to realize that AI upscalers work much better on crappy phone shots, than on high res MILC / DSLR photos. Which makes sense, because they were probably trained mainly on phone-taken images...

But back on topic. The point is - it depends on what you intend to do with an image. I print large formats, so for me any loss of resolution is a no-go. If I was just spamming social media daily, then I agree, it makes no sense to bother with TS, just stick a 10-20 on there and crop whichever way you like.

and that is very true that the person paying for something will determine what quality is acceptable but there is no doubt that more stuff in post (whether DLO, interpolation, upscaling, denoise or genAI etc) will keep coming.
Recently on the website, there was a discussion about whether to use a great composition but low res on a cover vs a not ideal shot in better resolution. I think that composition won.
I remember the audio quality wars from vinyl to CD and then CD to mp3 but at the end, the usability of portable but poorer quality files won except for a tiny niche of users

I was thinking more along the lines of a client that is with you on a shoot and has a laptop or a tablet in hand monitoring your photos simply will not understand how the final image will look like. It's hard enough explaining to them how HDR works, I really don't see how they could imagine the final frame. More than once clients are at a remote location and I'm sending images directly to them as they're taken.

Completely agree that AI / upscaling / generative AI will keep coming more and more and more. I use all of it myself. But until there's a clear advantage of using automated tilt/shift lenses or being able to get same quality via keystone correction, I really am not that hyped about the new TS lenses. Obviously, we still have no official info on them, so maybe some revolutionary tech will end up changing my mind.

 

[DSB]

Interesting, but a bit of a non-sequitur there. Tripod usage has no impact on the potential value of automatically setting the focal plane. Talking of which...

How long does it take you to set the tilted focal plane manually when shooting at high magnification? Do you have experience of using extension tubes on the 90mm or 135mm to get magnification in the 1:1 ballpark, where the depth of field is going to be of the order of 1mm or less?

In the grand scale of things (working in a studio), the amount of time is negligible. I'd never shoot a small product handheld. It just doesn't make sense. I'd set up a camera on a tripod, frame the shot, set the tilt and then spend 20x more time on lights, reflections, props etc. What you're describing is more an "insect in a natural environment" shot, where I agree every second counts.

 

[David - Sydney]

I was thinking more along the lines of a client that is with you on a shoot and has a laptop or a tablet in hand monitoring your photos simply will not understand how the final image will look like. It's hard enough explaining to them how HDR works, I really don't see how they could imagine the final frame. More than once clients are at a remote location and I'm sending images directly to them as they're taken.

Completely agree that AI / upscaling / generative AI will keep coming more and more and more. I use all of it myself. But until there's a clear advantage of using automated tilt/shift lenses or being able to get same quality via keystone correction, I really am not that hyped about the new TS lenses. Obviously, we still have no official info on them, so maybe some revolutionary tech will end up changing my mind.

You are definitely the market that Canon is hoping to sell the Rf TS-E lenses too then

 

[David - Sydney]

Well, I had basically a full university degree's worth and it ended up getting me a few jobs but I didn't really need it to do the jobs. But Tokyo's a great city to live in and for daily life of course it's critical here, unlike say Zurich (where I lived many years) where I'm certain that 20% of the population uses English exclusively and don't know and have no need to know the local languages.

When working in China, it was more important at work to help the local speak better english than me learning Mandarin. Not to mention that they speak the local Shanghai language which is completely different to each other (same written language though).
Working in Antwerp was embarrassing as they spoke English (and multiple other languages) far better than I could ever expect. They replied in English every time we tried to speak netherlands.

BTW I checked out your flickr and absolutely mind-blowing.

You are very kind

 

[DSB]

???

There's no way keystone correction and AI interpolation can replace tilt. Your post makes so little sense I wonder if it was just a cut and paste editing error or something? I'm sure you know the subject well and just deleted some words or something.

Huh? I literally said they can't replace real tilt, because if they could, then no one would use TS lenses. I think you misunderstood something.

That said I've already argued on this thread that shift is pretty much useless. You end up paying FAR more for a FAR more complicated lens that casts a huge image circle, and the more you shift, the farther you get into the corners of the MTF diagram and the further you have to deal with cos^4 and mechanical vignetting and weird highlight shapes. Combining left-shift and right-shift images to make one wider image was a nice trick before digital editing of high-resolution sensor images was possible. Now you can just stitch images together. A few times I used shift to avoid the camera's reflection in a window but that was when film looked so comparatively bad you didn't want to crop half a shot away after using a wider lens from the same standpoint. Now you can crop 80% away and still have a great image.

Really, I've had the 24TS and 90TS for years. I'm at a loss why anyone would use shift in the modern era, but tilt is absolutely irreplaceable.

Sorry, but I don't consider 45 megapixels cropped to suffice as a replacement for left+right or top+bottom TS shiftstitching.

 

[DSB]

You are definitely the market that Canon is hoping to sell the Rf TS-E lenses too then

I already own all of the TS-E lenses, so I may not be a prime target. But I'm sure autofocus will probably bring a lot of photographers into tilt-shift world. Many feel intimidated by all the complex mechanics and movements and on top of that they're manual focus lenses which only further makes them look complicated. Having autofocus will surely help with popularity.

 

[gerald.d]

In the grand scale of things (working in a studio), the amount of time is negligible. I'd never shoot a small product handheld. It just doesn't make sense. I'd set up a camera on a tripod, frame the shot, set the tilt and then spend 20x more time on lights, reflections, props etc. What you're describing is more an "insect in a natural environment" shot, where I agree every second counts.

No, that is not what I’m describing.

I’m discussing a similar scenario to taking the studio image I shared earlier in the thread.

Still interested in your answer to the question I posed, even more so now that you state you spend 20x more time than that on the lighting.

 

[RobbieHat]

Apparently Canon will spend anything and deliver on almost any lens other than a fast UWA. Broken record, I know, but sure would be nice is they had some fast UWA for us astrophotography folks. I can’t imagine we are any more of a niche than the TS crowd. I too have a 24 TS and have used is sparingly for landscape and architecture shots but likely won’t replace it.

 

[Torsten]

I was being careful about handholding the TSE 17mm, since I do not own one. When I still was TSE inexperienced, I once borrowed one. My results were less convincing. I too I'm certain that I'll soon handhold the TS 14mm...hating tripods. They slow you down when visiting a city. But I'm no professional architecture shooter, so I'm not always looking for the 100% perfect results.

Back in the day when I just got the 17mm, my initial results with it were not that convincing either. For me it took a while getting used to what shift (and to some degree tilt) meant for my compositions took a while. Having taken the 17 TSE as the only lens I had with me a few times certainly helped in that regard as well as the handheld usage.
Reminds me of when I got the EF14/2.8 a few years before that in 2012. The impact that ultra wide angle lens had in terms of compositions for me needed some time to get used to. Even though I still use the old EF14/2.8 today, I use it less often because to me, it's a real special-purpose lens while the TSE 17 is more of a general purpose lens to me than the EF 14mm.

But then, I'm weird and I'm not a professional shooter of anything, leave alone architecture.

EDIT: Seems the forum can't properly handle emojis. I smell improper UTF-8 handling...

 

[Nemorino]

I don't know how this will look as I don't do video, but you are exploring ideas I hadn't thought of (object tracking is supposed to be a new feature in R1?).

A promo video of Schneider Kreuznach for their cine tilt lenses. Just two minutes but nice as they show the amount of tilt beeing used in a small graphic:
Story telling with tilt lenses
The sceen in a bar at ca 1:00 is maybe of Your interest.

 

[blurryeyedbeast]

Apparently Canon will spend anything and deliver on almost any lens other than a fast UWA. Broken record, I know, but sure would be nice is they had some fast UWA for us astrophotography folks. I can’t imagine we are any more of a niche than the TS crowd.

I suspect there is one very important difference: the number of people using the lens on well paying professional photography for said niche.

 

[Nemorino]

this would imply a completely new Tilt design, as the current TS-E lenses tilt all elements altogether.

Like this:
https://www.canonrumors.com/canon-patent-application-automated-tilt-movements/

 

[LDS]

I hate the tripods, too.

At least now there are lighter/smaller carbon tripods that can yeld good result, and WA T/S lenses don't require heavy tripods and heads. Shooting LV means also the mirror is already locked. I use less my Manfrotto 055/410 combo outside, most of the time it's the Beefree Advanced GT, although I'd like a geared head on it. Anyway this is not the kind of photography you do when traveling with friends... unless they are after the same photos. I carry it when I know I'm going to use it. My real issues, as said before, are places where they are forbidden. In Italy there is a old law that treats them as "public land use" (free translation of mine), probably made for the old "street photographers" (as Buster Keaton in the "Cameraman"), and now used to extort money from photographers - you need to apply for a permission and pay a fee. In other places it's much easier, I extensively took night shots in Doha a couple of years ago and police ignored me. In Milano in areas like City Life and Piazza Gae Aulenti is just like extracting a machine gun. Moreover many "celebrities" live around and everyone with pro gear is regarded a paparazzi.

In terms of value for money/size/weight/already has AF etc, using the RF10-20/4 and then keystone correction with associated interpolation would be a quicker and cheaper option than a TS-E lens

Given the price of the 10-20 it might be quicker but not cheaper. Anyway depends on the level of details of the subject, the amount of correction, and how large the image will be displayed.

 

[roby17269]

Like this:
https://www.canonrumors.com/canon-patent-application-automated-tilt-movements/

Oh I had forgotten that
So yes it'd be a new different type of TS design.
Well I guess we will see what Canon will deliver, if anything

 

[scyrene]

I do not think tilt/shift would work for that unless the butterfly would stay perfectly still.
Try stopping down your aperture.

Butterflies do sometimes stay still for a while. Stopping down makes the background busier.