Patent Application: Canon RF-S prosumer lenses

[SwissFrank]

This sounds like a great lens, and if Canon did this I would immediately be purchasing an RF-S camera body to go with it.

Are the cameras a lot smaller than the R5? For me the R5 is borderline too small. I haven't tried it but doesn't the R5 make a smaller file from a smaller sensor area when run with a small-sensor lens?

 

[DhlcadR6]

Are the cameras a lot smaller than the R5? For me the R5 is borderline too small. I haven't tried it but doesn't the R5 make a smaller file from a smaller sensor area when run with a small-sensor lens?

If you put RF-S/EF-S lenses on, of course you get smaller files. For R% you are dropping back to 18MP~approx

 

[Walrus]

Are the cameras a lot smaller than the R5? For me the R5 is borderline too small. I haven't tried it but doesn't the R5 make a smaller file from a smaller sensor area when run with a small-sensor lens?

Yes, definitely; R5 is like my R6, I have the RP which is considerably smaller then them, and I had the R10 which is definitely smaller then the RP; R50/100 should be slightly smaller then R10 but not to the point of really being more portable, the real difference would be an R body with the form factor of the M6II (which I also had), that thing with the EF-M 22 was a real pocket camera, while today you cannot "pocket" any R camera, even with the 28mm pancake.

But yes, excluding R7, any other current R crop body is WAY smaller then a R5, between my R6 and R10 was night and day.

But I agree with you, for my tastes R5/6 are, if not borderline too small, just the perfect size (considering that I use the R6 with the vertical grip); I would deem RP/R8 ok but borderline small (I bought the little grip extension for the RP, with that it's just perfect size for a small FF camera), while the R10 (even if control layout, including the joystick, is WAY better then the RP) was really too small for my hands, which are normal size if not slightly small for the average man, and too small for the lenses I was using with.
R10 was fine with the 16 and 35 RF (and so would be fine with a 50 1.8 RF), with 85 f2 RF was already borderline unbalanced, and was really undersized with any of my EF adapted lenses, excluding the EF-S 24, which was indeed really ugly with the adapter, but in the end was still compact and balanced enough.

 

[EOS 4 Life]

How so? It's taking a scene and producing a rectilinear file, right? How can you even detect that the intermediate format of the information--an image projected on the sensor--had distortion, once you let the algo correct it? How is that not OK for landscapes in some situations? How not for astro?

If you don't have an actual side-by-side comparison to show how this is bad, can you explain even what a person would look for to tell the difference? Between say the Canon RF16/2.8 that apparently uses this technique big-time, and some other 16mm lens of your choice that is perfectly rectilinear, or even a theoretically perfectly rectilinear one?

A stretched star is a bad look even in the corners.

 

[janhalasa]

I would vote for constant aperture for video and 15mm as starting point would be very welcome, especially with good correction in the distortion department. But maybe two lenses for PHOTO and video + VIDEO and photo

I guess you prefer constant aperture to zoom without changing it. I think today's variable aperture lenses allow that. The only exception is when it's not physically possible - when you choose the lowest aperture on wide end and zoom in. So I think a f/2.8-f/4 lens is better than a constant f/4 for everyone. But I may be missing something, since I don't do video.

 

[neuroanatomist]

A stretched star is a bad look even in the corners.

I get it...you've decided you don't like algorithmic correction of barrel distortion. It's becoming clear that you've made this decision without any evidence including that already shown, and without even bothering to apply logic to the situation. It's a common occurrence these days for certain people to ignore facts and reality and just stick with their own misguided opinion, but commonness doesn't make it less sad. So I'll try one more time.

Perhaps the problem is you are just interpreting the word 'stretched' literally. This is where logic can help. Yes, there is 'stretching' going on...but before the stretching, there is a 'squishing'. The aberration being corrected with the stretching is barrel distortion. The purpose of the stretch is to algorithmically correct the squish. If the idea of stretching bothers you, think of it as it really is...unsquishing.

Here's a diagrammatic example of what's happening in the corner of an image from such a lens:


Barrel distortion makes straight lines into curves and circles into ovals, and correction restores them to straight lines and circles. In a lens that's optically well-corrected, lens elements do that job.

Here's a real example:



The top panel shows heavy barrel distortion, the middle panel is that image with digital correction of the distortion. The bottom panel is a different lens with essentially no distortion. Are the optically-corrected lines straighter than the digitally-corrected ones? Is sharpness/detail meaningfully affected (as asked previously, with examples)?

Since you are hung up on stars, I will borrow from Bryan's (TDP) reviews. Here's the RF 14-35 at 14mm, corner with distortion corrected:


Before you cry, "Vindication! Look at that stretching!," be aware that you're seeing coma. The shapes of the stars are actually pretty good. For comparison, here's the EF 11-24 at 15mm, where there is essentially no geometric distortion to be corrected:


As with my building examples, I don't see any meaningful difference between the two as far as distortion goes, though both have a fair bit of coma (as do most Canon lenses).

So, I'll ask again...can you show us an example of this 'stretched star bad look' from the geometric corrections? Or are your objections just due to an opinion unsupported with actual evidence?

 

[SwissFrank]

A stretched star is a bad look even in the corners.

How can you tell it was stretched? Can you show an example?

 

[SwissFrank]

If you put RF-S/EF-S lenses on, of course you get smaller files. For R% you are dropping back to 18MP~approx

Of course, but I get smaller files also if I use a small-sensor camera, no? So aren't we facing a decision: if we want a more portable solution, we can use the smaller RF-S lenses either 1) with the camera we have for free, which produces a smaller file, or, 2) buy a second body, for a fair amount of money, which also produces a smaller file, and for which we have to learn a somewhat different feature set, way of operating, lose advantage of muscle memory and so on? Maybe I'm stupid or lazy or just poorer than everyone else but I'm wondering if sticking with the body I have isn't the cheapest, simplest, easiest solution?

 

[SwissFrank]

too small for the lenses I was using with ... 85 f2 RF was already borderline unbalanced

Thanks for your note which was quite informative. Only comment I have is I've never felt cameras were unbalanced, even with the EF50/1.0, EF85/1.2, EF70-200/2.8s, etc., as I just hold them more by the lens and less by the body as they get bigger.

Yes, I quite by accident bought the EOS M on like the fourth day of sale: entering Bic Camera from the subway you pass the Canon counter and the thing just jumped out and grabbed my wallet. And I have a picture of it in my jeans pocket with the 22/2 mounted. But I used all my EF glass on that without issue. I didn't pick it up with the 70-200 mounted by the camera any more than I'd try to pick up the 70-200 by the rear lens cap.

 

[Walrus]

Thanks for your note which was quite informative. Only comment I have is I've never felt cameras were unbalanced, even with the EF50/1.0, EF85/1.2, EF70-200/2.8s, etc., as I just hold them more by the lens and less by the body as they get bigger.

Let me rephrase with unbalanced I meant "slightly disproportionate"; my EF lenses are all pretty big and heavy, and even more so thanks to adapter extra length, so of course I'm used (since film days) to grab the camera and the lens correctly.
My note was like "up to RF 35, or 24, and of course anything smaller like 16 and 50, on a R10/50/100 you grab the camera to support the lens; starting from 85 STM you start having a balance in size/weight between camera and lens, but the lens starts to be borderline long and much less portable considering the camera size, and anything longer and/or heavier then you start grabbing the lens to support the camera".
RP has some more leeway (especially with the small extra grip), the 85 STM is still totally at home; but now (having sold the 85 STM) I directly jump from the native 50 STM to the adapted 24-70 2.8 II, and the difference in portability is of course pretty staggering; but then, I mostly do ceremonies nowdays, so the portability and compactness for me (referring to travel/nature/family/etc) is a non-issue.

 

[Tom W]

Image stretching needed at the widest angle focal lengths. Makes it not useful for astrophotography. Questionable for architecture, but the jury is still out on that.

 

[neuroanatomist]

Image stretching needed at the widest angle focal lengths. Makes it not useful for astrophotography. Questionable for architecture, but the jury is still out on that.

I'll ask the same things I asked of @EOS 4 Life – what evidence do you have that distortion correction is a problem for astrophotography or architecture. Can you post a picture showing how the distortion correction renders images from a lens that requires it unusable for astro or architecture? For the latter, can you see a difference and identify the digitally corrected image vs the optically corrected image in this example?

 

[SwissFrank]

Image stretching needed at the widest angle focal lengths. Makes it not useful for astrophotography. Questionable for architecture, but the jury is still out on that.

Why so? How can you possibly tell, looking at the resulting photo? The lens itself compresses a bit, so when stretched back out, everything ends up, to the pixel, being exactly where it was supposed to be, no? The image in the file you get from the camera, post-correction, is literally more rectilinear than any (uncorrected) lens ever made in the history of mankind. No lens is naturally as rectilinear as a software-distortion-corrected lens (at least, assuming the algo knows how the lens magnification varies, and I haven't seen any evidence that Canon doesn't know this.)

 

[neuroanatomist]

Interesting that when those making claims about lenses requiring distortion correction being unusable for astrophotography or architecture are asked to support their claims with evidence, they go silent...and when presented with evidence that refutes their claims, they ignore that as well. I guess I shouldn't be surprised, people keep wearing red hats, too.

 

[dcm]

Image stretching needed at the widest angle focal lengths. Makes it not useful for astrophotography. Questionable for architecture, but the jury is still out on that.

NASA might have a different viewpoint...

 

[ghalib737]

Nikon will cross the sales of Canon within few years due to publishing their lens tree and allowing third party lens. M50 had only six or seven third party lens, still it was the most popuyy Apsc. R10 and R50 has potential but Canon will not learn lesson.
Worried after buying R10. As I need a overall travel zoom lens like Sigma 17-50mm f/2.8 (is).

 

[neuroanatomist]

Nikon will cross the sales of Canon within few years due to publishing their lens tree and allowing third party lens. Worried after buying R10. As I need a overall travel zoom lens like Sigma 17-50mm f/2.8 (is).

Sure, Canon has been on top of the market for 20 years and currently has nearly 50% ILC market share, while Nikon used to be a very close second but over the past few years has dropped to a distant third with less than a 14% market share. But they've published a roadmap, and they allow 3rd party lenses. That will make all the difference.

Canon's lack of a roadmap and blockade of 3rd party lenses have been the case since before the R10 launched. Yet you bought an R10 anyway, so if you're worried you have no one but yourself to blame.

 

[Walrus]

Worried after buying R10. As I need a overall travel zoom lens like Sigma 17-50mm f/2.8 (is).

I can suggest the Tamron 17-50 f2.8 (NON-VC) which could be found super cheap, and on my former R10 was working like a charm, both in quality and compatibility. I tried (still on R10) a sigma 17-50 2.8 OS and the quality was exactly like the Tamron in the middle, while was terrible on borders and corners, I put it back on eBay the day after I received it.

 

[ghalib737]

Sure, Canon has been on top of the market for 20 years and currently has nearly 50% ILC market share, while Nikon used to be a very close second but over the past few years has dropped to a distant third with less than a 14% market share. But they've published a roadmap, and they allow 3rd party lenses. That will make all the difference.
View attachment 210880
Canon's lack of a roadmap and blockade of 3rd party lenses have been the case since before the R10 launched. Yet you bought an R10 anyway, so if you're worried you have no one but yourself to blame.

I brought R10 as my wife use my camera for her product photography (she need a simple menu camera as I had M50 before R10). For her work 18-45 works well as the IS of this lens allow her to go up to 1/30 shutter speed hand held. But for my YouTube work I need an fast zoom and a fast prime which I do not have and also Canon not offering. If Canon go for 15-70 f/2.8-4, it will be a relief for me. And for studio video now am using EFS 24mm f/2.8.

I wish Canon will allow Sigma or Tamron for their 17-50 f/2.8. Or they can release like those one.

 

[ghalib737]

I can suggest the Tamron 17-50 f2.8 (NON-VC) which could be found super cheap, and on my former R10 was working like a charm, both in quality and compatibility. I tried (still on R10) a sigma 17-50 2.8 OS and the quality was exactly like the Tamron in the middle, while was terrible on borders and corners, I put it back on eBay the day after I received it.

Yes. I tried Sigma 17-50mm and it is too heavy for travel video work along with Viltrox speed booster. But having IS is an advantage but it sounds a lot. I never tried Tamron 17-50mm--- although it is not available in my country (Bangladesh)