This Canon patent application delves into the operation of a motorized tilt-shift lens for what appears like the Canon RF mount.
Tilt-shift lenses have complicated and fragile mechanics, and removing some of the fiddly nature from these lenses would help in the field if the ergonomics of controlling the lens were fair.
From the patent application, Canon feels it would then be possible to remember positions and reset and go to remembered positions (perhaps saved as well on EXIF data I would imagine), so it’s possible to look at a saved photo, and then decide you want to retake it with the same tilt shift settings, it’s just a matter of clicking a button. Or if you have a favorite setting (for example atypically shifting 4.5mm to take photos of buildings), having a saved setting feature can completely optimize your time in using a tilt-shift. This seems to be the intent of the patent application which is to streamline the amount of time it takes to effectively use a tilt-shift lens.
Canon places the control of the tilt-shift either on the lens itself or in the camera. To be honest to me, it seems to be better suited to be in camera than on the lens. Bluetooth control would be swell too.
This patent could be expanded on and include some form of automation on shift that would look at your perspective distortion and shift until it’s gone automatically. It could also lead to Canon doing an “auto-stitch” which is to shift the lens or instance all the way to the right, take a picture, center, take another picture, and shift all the way to the left and take the final image and combine them together.
Now that could all be me because I’ve used shift lenses through the years far more than tilt-shift, so I immediately thought about the practicality of the shift mechanism on what a full camera-controlled shift lens could do if Canon put their minds to it.
Outside of removing the manual nature of tilt, I could see it having a finer control and possibly some interesting video usages as you move from 0-degree tilt to miniaturization in a smooth movement as another potential benefit. Tilt I think is the hardest aspect to grasp, so to be able to watch the back LCD, control the tilt, and watch the effect could be a bonus. Also if you are using tilt to take photos with a much deeper plane of focus, the camera could automatically adjust the tilt until foreground and background elements are in the plane of focus.
Craig was personally excited to have something that didn’t break as often in the field. Apparently, that was a big problem when he was renting out lenses.
This was a patent application where at first, I was “meh, I guess it’s cool.” and then got more excited as I wrote this up. That happens sometimes. Oh and if someone from Canon reads this and it gives them ideas for a future patent application … I do accept cheques and paypal ;)
As with all patent applications, this may never lead to an actual patent or a product. However, it does certainly give us ideas about what Canon is researching.
Via: Asobinet.com
Source: Japan Patent Application 2023-135457
|
When you purchase through links on our site, we may earn an affiliate commission. Here's how it works. |
In my generation, only smart people had fast cars and we looked after them and were very careful. By the amount of broken and bent out of shape expensive Mercs and BMW's I see in major accidents these days...I can only conclude that a) stupid pople now have access to machinaery that far exceeds thier skills, b) some people have more money than they care for or c) some people hire this stuff and don't treat them like they should and don't fear the consequences.
Also i guess when you are renting maybe you don't have the same regard for how expensive and delicate the optics and mechanisms are.
I've personally never damaged one either.
I do put most if it down to human silliness... but they shouldn't have been so easy to damage.
Only the original 100-400 had a higher repair rate for us (some of which was done in-house).
The first time I used my TS-E 24mm f/3.5 II, I saw how fragile the knobs and locks are. When I use the lens I am constantly aware of it and turn carefully. That's the only thing you can do. But Canon can indeed make some improvements.
While I personally don’t see any need for the movements to be motorized, if they are then they could also be encoded (of course, that’s possible without a motor). Encoding the movements and recording values in the EXIF would be a high interest to me, since it would enable profile lens corrections to be applied for the movements (e.g., correcting asymmetrical peripheral shading).
When ever I've rented any kit, it goes back in the same condition that it arrived to me. I treat it as my own as and as a precious thing. Unfortunatly, some people don't even have the concept of a precious. I tended to use rental as an opportunity to try out kit that I'm not sure of before buying. I did this with my 24-70 II L, 35/f1.4 L, 70-200 f2.8 II L and my older 400mm f2.8 LIS.
I'm not sure motorized movements are a good idea, for the same reason one might not want a motorized zoom. It's quicker and more precise to set them manually - especially shift, tilt requires more training. With motorized movements one might start to "hunt" to get the exact placement, depending on the movements speed. I wouldn't like a UI that requires to switch between T and S each time, it would be easier to forget the switch and move the wrong setting.
I agree that some automatic settings could be nice, but using these lenses means "slow photography" anyway. I'm sure it would be better for renters, less wear on parts by less careful people.
Encoding would be useful, especially if the positions are shown on the display, and recorded on photos.
I did buy a used TS-E 90mm used almost 10 years ago. Played around with it, found that I wouldn't be using it on any type of regular basis, and sold it at a profit(!) about 1-1½ year later.
The issues of damage are more related to dropping the lenses or "torquing" the knobs rather than rotating them. I have seen a friend drop her 24 and she had to send it to CPS as it obliterated her shift mechanism. However, never thought of the lens mechanisms to be fragile.
I see the validity of finer control and electronic actuation. But that will accelerate battery usage and then comes the question of price, since the TS-E are already dearer, with Canon adding the "R multiplier", these lenses will surpass $3K. Also, weight may increase. Will it reduce time while focusing? It really depends on how Canon makes the system user friendly.
Of course, as you wrote, if you overtighten the controls with sheer bestiality...
PS: tell your friend I'm sorry for her, I dropped mine too and know what it costs. ;)
Lately, I've been considering the TS-E 135/4L Macro. I certainly don't need it, and I wouldn't use it frequently. But like the MP-E 65, I think it would be a fun lens to play around with for small items, especially with winter coming up and outdoor opportunities becoming more limited. Not a lens typically available used locally and not one I'd likely be able to sell locally, either. I suppose I could rent it, but more likely I'd just buy it. It's been discontinued by Canon and isn't available at B&H or Adorama, but Amazon still has it (sold by them, not a 3rd party). I'll keep thinking about it...
I do find it interesting that in the TDP review, at the top Bryan states, "Note: The Canon TS-E 135mm f/4L Tilt-Shift Macro Lens has been replaced by the Canon RF 135mm F1.8 L IS USM Lens." Ummm, no. Granted, they're about the same price (the TS-E is $100 more), but a fast AF prime is not a replacement for a TS lens. Oh, and they both take 82mm filters. That reminds me, a few months ago I commented, "In the box on the left is an extra 82mm XS-Pro clear filter just waiting for some future RF lens to protect." Maybe I was telling myself I should get the RF 135/1.8 instead of the TS-E 135/4? o_O
(It'd also require a motor to rotate the plane of tilt.)
The other main trick of tilt is to minimize field of focus, but you can just swing the front manually because you're not trying to just exactly nail some specific amount of unfocus--we don't even have a term or units or way of measuring it, much less specifying it.
There's no particular reason I see that the motors for these functions would be more susceptible to breaking in a fall than focus motors on non-internal-focus lenses.
The lens resolution on such a wide image circle will fall off dramatically towards the edges.
You'd probably get similar or superior results simply by using about 0.6x the focal length (assuming 11mm shift) and shooting a single shot.
Also, stitch software is able to stitch shots where the camera's not even facing the same direction, so you can just rotate the camera on the tripod a bit to take your second shot. If you do this you're not using that dark and blurry outer ring of the image circle, but rather the sharper center portion.
Actually it works quite well - and you can do it not only in a single plane.
Given the 45MP R5, though, and how sharp today's RF glass is in the center, I suspect cases where say a center half of a 45MP 35mm image (22MP) is notably inferior to two 45MP 50TS images stitched together (call it 75MP) would be very very small. I like resolution IN THEORY, and of course 3.4x the pixels sounds great, and has a certain intellectual satisfaction...
... but what use case would actually USE those pixels? I mean, I usually edit personal photos to 1500x1000. That's 1.5MP. No photo on the internet seems to be so big. Billboards are too far away to really see more pixels. Magazine printing probably doesn't have the resolution. Maybe doing a big panorama on a wall in an office that people CAN get close to? And how many times does someone on this forum make such output?
(Really the great thing about 45MP, to me, is that you can crop the heck out of photos and still have something totally useful. But by definition, in this panorama proposal, you are not cropping out details.)
And yet, while you have 3-4x the pixels, you also have substantially less sharp the lens performance. Lenses with a large image circle have a lot of fall-off of sharpness towards the edges, and in an apples-to-apples comparison you're comparing the 43mm-wide image produced by the normal lens with a 63mm-wide image produced by the shift lens. A defect in the lens manufacturing will simply be magnified 1.45x on the shift lens. You don't magically get a lot more lines of resolution at a given contrast just because the image circle is bigger.
And swinging a shift lens from max left to max right shift is hardly a chore anyway!
In short, you and the OP are arguing:
-- this ultra-rare media usage is nonetheless somehow common enough,
-- and having the extra MP despite no increase in IQ is somehow so valuable,
-- that we will make the first motorized shift lens in human history to satisfy this ultra, ultra, ultra, ultra niche.
In contrast my argument is:
-- there are enough people who use AF for single points in a scene,
-- that I project there would be enough people who'd also like AF for 2 or 3 points in a scene,
-- to make this valuable.
I mean, how many photos have two-three objects at different depths that you'd like focused? I'd guess something like half of them, no?
Ultra-rare? I use it quite often to get images that looks well in a 16:9 (or similar) layout compared to the 3:2 sensor size. Pixel peeper will notice something at the edges, most people won't - especially when the edge detail is not so important. These lenses anyway are not so bad even when fully shifted. It's also not about having more MP - it's to simplify the shooting of panoramic images.
What is niche for you is routine for others. Some people still require fairly large images, and can't crop too much. Sometimes you can't back-off enough. And these images are easier to stitch, Anyway, I don't care if it is motorized (I said I wouldn't like it), doing it manually just require moving the lens manually instead of sitting down and letting the camera do it itself (as I do now focus stacking with Helicon Remote doing it for me). Still, once the lens is motorized it's quite simple to get that - more than pixel-shifting the sensor.
Tilt allows to rotate the focus plane, but it's not magical - depends where those points are, and what focus depth for other points is needed. Sincerely, focus stacking allows to get much more in sharp focus, although it doesn't work with moving subjects within the frame, but even tilting require fairly static subjects - unless the camera tracks and predicts the trajectory of each focused object. Sure. less issues with leaves and the like.