The 2020 RF lens roadmap, up to 8 new lenses coming in 2020

Feb 28, 2013
1,612
272
70
Like a 50mm 1.2 that is bigger and heavier than the EF one, or a 70-200mm 2.8 that is fatter and probably heavier than the EF one? And it'll probably be twice the price as well, to keep the RF tradition alive. Oh how about a 24-105 that's pretty much identical in specs and IQ to the EF II? How's that an advantage?

Canon's RF designers seem to have Sigma Envy and want to show that they too have big manly bits. Meanwhile I'm looking at the EF 35mm IS because it's small and light.
I own three Canon 24-105mm zooms, two in EF and the RF version. Your so wrong about the RF version its technically superior to the EF versions. We have an EF adaptor for our optical lens projector and have made an RF adaptor so please quantify how you have made this technical observation?
 
  • Like
Reactions: 1 user
Upvote 0
Feb 28, 2013
1,612
272
70
The disadvantage of the RF mount is you cannot simply remount existing optics to work with a shorter back-focus (it may have worked to an extent the other way). That means brand new designs and the limitation is all opto-mechanical engineering time / personnel, as well as manufacturing capacity. Add to this Canon is still designing both motion picture and broadcast lenses as well as manufacturing these. Its a huge commitment but one they are facing up too with determination.

Having met with a Canon lens engineer today, plus a sales colleague, they are more than up to the task. They are getting a wide industry perspective not just on direct replacements but "like to haves", "unusual & challenging optics" as well as pushing boundaries not possible with the EF mount whilst also looking beyond traditional applications.

Have faith, Canon know what they are doing.
 
  • Like
Reactions: 1 user
Upvote 0

mk0x55

[5DsR]
Nov 16, 2018
56
52
Every single picture, any part that has a sharp contrast between light subject and dark background (and vice versa). This, for example:
View attachment 184182

Or this:
View attachment 184184

Which is at least a millennium old technique of "3D pop" image enhancement, but has nothing to do with lenses.
Thanks for your examples; I had a closer look and saw it, too on the swan. By the way, the yellow machine is downsized by almost nine times... that hardly retains the sharpening halos. By the way, do you know how human visual perception works? Our brains do exactly the same - they lay a mask over bright objects in a contrasting scene (and not only luminosity, but color, too) and if you slightly move your eyeball then you will see it. Since the dynamic range of a daytime scene is much greater than whatever the monitor can shine on you, the effect will also be stronger.

Below I provide these images with the sharpening completely turned off in Capture One. Now guess what - no difference to the 3D pop of the image, really. Att 100% it may even look better now unsharpened. The 3D pop dwells in something else - most likely a good degree of preservation of the tonal gradation across the image, not just sharp transitions around contrasty edges.

This is the swan picture with all sharpening turned off (it is now att 100% crop):
184187

A 100% crop from the yellow machine with all sharpening completely turned off - the fine tonal gradations and the 3D pop are still there, especially if you back out of the screen a little:
184188

Now the whole image downsized to 1024 pixels on the long edge by Photoshop's bicubic method optimized for smooth gradations - still "pops" nicely out, at least as far as I can see. Still with all sharpening completely turned off:
184189

I think this rules out that it's sharpening in post that adds the 3D pop. Whatever the pop dwells in; it contributes to the object being recognized by my brain as floating in a pleasant 3D space rather than a flat sterile and boring surface (slightly exaggerated words here).

Do you see the subtle shifts of tonality across the swan or the yellow machine as their different surfaces curve? This is more what I concentrated on rather than pixelpeeping for sharpness, sharpening and halos (all that is about high-frequency intertonal detail; I'm studying the low-frequency one here). That is undoubtedly what promotes my perception of that 2D image as a 3D scene.

Don't forget what the ultimate goal of photography as an art is: Mediate the strongest way of seeing a scene.

If this 3D pop rendering character of a lens (and even if it was 100% postprocessing which it is not, actually to about zero degree) helps me mediate a stronger way of seeing a scene than otherwise, then it is a heck of an asset to my photography. And in such case I can't care less what some dogmatically limited a-priori judging mind might say about it. It simply works and it works beautifully. Above is some evidence for it.

EDIT:
No, it's physically impossible.
It might be for the sharpening halos...
On the other hand, the lens gets in the light through a pretty wide front element - light that at different positions as entering the lens througout the time of the light travel and capture has different spectrum of intensity, wavelength, polarization... you name it. The information input to the lens is far far richer than whatever the sensor can capture, irrespectively of resolution. Why do you think that it is physically imposible to make a lens that would convert part of that information into a part of the information that is captured by the sensor, so as to help our brains reconstruct the captured 2D image into a 3D scene in our perception?
Maybe this is what every lens is supposed to do, but the more optical imperfections of e.g., lesser quality glass; and imperfections introduced by optical corrections from a lens, the more this reproduction suffers - in the same direction as Yannick etc. claim. I say maybe, as I don't have any proof of that nor any clear evidence for that statement.

I actually intend to follow up on these examples with a comparison of the same scene captured by two or three different lenses. Let's see what that yields.
 
Last edited:
Upvote 0
Jul 21, 2010
31,091
12,855
Focus confirmation and (possibly) AF become more likely with DPAF that can focus down to f/11.
Given that the aperture is electronically controlled, and focusing can be performed wide open (which is how it’s done in a DSLR), I’m not sure that DPAF has any relevance to the issue.
 
Upvote 0
Jul 21, 2010
31,091
12,855
Every single picture, any part that has a sharp contrast between light subject and dark background (and vice versa).
Those are the places where aliasing often shows up in an image. Given the mention of 50 MP and previous posts indicating @mk0x55 owns a 5DsR, I suspect that’s what you’re seeing.

I think this rules out that it's sharpening in post that adds the 3D pop. Whatever the pop dwells in; it contributes to the object being recognized by my brain as floating in a pleasant 3D space rather than a flat sterile and boring surface (slightly exaggerated words here).
I think your ‘3D pop’ is a depth of field effect, with a large leavening of wishful seeing and perhaps a small leavening of aliasing.
 
Upvote 0

mk0x55

[5DsR]
Nov 16, 2018
56
52
I think your ‘3D pop’ is a depth of field effect, with a large leavening of wishful seeing and perhaps a small leavening of aliasing.
Well, it can happen that I fail to see the visual evidence for what I claim (and I yet have to demonstrate that point here).

However, what I've been talking about relates also to what Lloyd described in his article under the chapter "Penetrating power: distinguishing dark tones" in here:
https://lenspire.zeiss.com/photo/en/article/micro-contrast-and-the-zeiss-pop-by-lloyd-chambers

And that actually is how I also understood it and what I understood even Yannick describes in his blog. (The difference between the two's emphasis being that while Lloyd dominantly discusses the crispness of the lens's rendition on a very local scale; Yannick seems to only mention this Lloyd's concept of penetrating power, or tonal fidelity on a more global scale of the lens; plus blaming the current resolution-centric trends and creating controversy by describing a strong competition between the lens' optimization on resolution and corrections, wide-aperture optical performance, and this tonal fidelity that he terms as the 3D look.)

The sharper (or otherwise more distorted) the tonal transitions are rendered by a lens from a completely neutral transition (the ideal of linearly going from fully bright in the scene to completely black); the more the 3D-pop evoking rendering shall to my understanding be impaired.
This I also relate to the fact that our digital cameras capture most tonal detail close to the luminosity values of sensor saturation, i.e., closer to the highlights (see even https://www.cambridgeincolour.com/tutorials/gamma-correction.htm), where they also are least affected by read noise etc. Shadows as we humans perceive them contain non-linearly less tonal detail than highlights when captured by our digital sensors. If the already inferior tonal detail of the shadows is further compressed by the lens' rendering to even more shadowy areas of the captured tonal spectrum, then we're really losing a lot of what helps the brain reconstruct a 3D scene.

Anyways, let's see if I manage to find any discernible differences when I compare the lenses on a suitable scene in some reasonably stable weather & lighting conditions.
 
Last edited:
Upvote 0
I did not say it was impossible for EF. It just makes more sense to do it for RF, because of more space in the lenses, faster communication protocol and more fine-grained autofocus control by default.

I think the mechanics of tilt-shift lenses are the far bigger obstacle with regards to autofocus than the focusing system of the camera. But, I am not an engineer so what do I know... I don't mean to be hard on you, I just value the image quality potential of RF far more than features like AF, and when I'm using my tilt-shift, MF is not really ever a problem.
 
Upvote 0
Jul 21, 2010
31,091
12,855
Th AF system of most dslr's work with f/5.6 or wider apertures. Sensor-based focusing is definitely a game changer.
How? DSLRs AF with the lens wide open, so on-sensor AF is only ‘game changing’ with a lens having a max aperture of f/8 or narrower. (Are you simply unaware that’s how it works? The selected aperture is irrelevant, only the max aperture of the lens matters.) That’s mainly f/4 lenses with a 2x TC or f/5.6 with a 1.4x or 2x TC (and many recent higher-end DSLRs have PDAF at f/8). We were discussing TS lenses. In that context, those who use a 2x TC with the wide angle TS lenses would have their game changed. All six of those people.

On sensor AF is game-changing from the standpoint that Canon can legitimately come out with f/6.3 lenses for MILCs that don’t have to ‘spoof’ the AF system like the less expensive 3rd party telephoto lenses for DSLRs.
 
Upvote 0
How? DSLRs AF with the lens wide open, so on-sensor AF is only ‘game changing’ with a lens having a max aperture of f/8 or narrower. (Are you simply unaware that’s how it works? The selected aperture is irrelevant, only the max aperture of the lens matters.) That’s mainly f/4 lenses with a 2x TC or f/5.6 with a 1.4x or 2x TC (and many recent higher-end DSLRs have PDAF at f/8). We were discussing TS lenses. In that context, those who use a 2x TC with the wide angle TS lenses would have their game changed. All six of those people.

On sensor AF is game-changing from the standpoint that Canon can legitimately come out with f/6.3 lenses for MILCs that don’t have to ‘spoof’ the AF system like the less expensive 3rd party telephoto lenses for DSLRs.


I'm quite aware of how it works. T/S lenses have trouble with AF because of the mechanical vignetting caused by moving the exit pupil around. The combination of having phase detection in 80% of the sensor and sensitivity down to f/11 means that there should be a lot more flexibility for the angle of the rays hitting the AF sensor.

In simpler terms, on a dSLR, a lens at maximum shift or tilt will be "seen" as having a smaller aperture because half of the rays will be blocked by the narrower angle of view of an AF sensor that needs to see the entire f/5.6 aperture to function.

This is also why some AF sensors will AF at f/2.8 only with the center AF point.
 
Upvote 0
Jul 21, 2010
31,091
12,855
I'm quite aware of how it works. T/S lenses have trouble with AF because of the mechanical vignetting caused by moving the exit pupil around. The combination of having phase detection in 80% of the sensor and sensitivity down to f/11 means that there should be a lot more flexibility for the angle of the rays hitting the AF sensor.

In simpler terms, on a dSLR, a lens at maximum shift or tilt will be "seen" as having a smaller aperture because half of the rays will be blocked by the narrower angle of view of an AF sensor that needs to see the entire f/5.6 aperture to function.

This is also why some AF sensors will AF at f/2.8 only with the center AF point.
Thanks, that makes sense.

However, it’s still not a capability conferred by the RF mount. A DSLR in live view could have the same functionality (and live view is a logical choice for use of a TS lens, since it obviates metering problems with lens movements as well).
 
  • Like
Reactions: 1 user
Upvote 0
Thanks, that makes sense.

However, it’s still not a capability conferred by the RF mount. A DSLR in live view could have the same functionality (and live view is a logical choice for use of a TS lens, since it obviates metering problems with lens movements as well).

You're welcome. I agree a dSLR in LV could do that now, though new optical designs could absolutely smoke existing ones were they to use the new lens mount.
 
Upvote 0
Jul 21, 2010
31,091
12,855
...though new optical designs could absolutely smoke existing ones were they to use the new lens mount.
I doubt that, except perhaps for the TS-E 17. Certainly the RF lenses we’ve seen so far don’t support that viewpoint. The RF 50/1.2 is sharp, but so is the Otus for EF. Canon has patents for EF f/2 standard zooms (and they couldn’t manage to get the RF version out to 24 mm, personally I’d take the wider FOV over the stop of light in a standard zoom), and the RF 24-105 is essentially identical to the EF version. Maybe Canon’s market-speak will turn out to be true in some cases (e.g. a UWA zoom), but we haven’t seen that yet.
 
Upvote 0
Feb 28, 2013
1,612
272
70
This thread is about the new RF lenses but playing into that is the cameras. A Japanese organisation called BCN+R has analysed sales data for the EOS RP and despite strong sales in the first month the second month sales have "collapsed" in Japan. The article puts this down to lens choice which I would tend to agree with an entry level full-frame mirrorless camera without entry level RF lenses seems like a rare "own goal" by Canon. With no immediate lens like an RF version of the EF 24-105mm f3.5-5.6 IS STM (Which in the UK is listed at £ 414.99 as opposed to the EF 24-105mm f4L IS USM II at £ 1,029.99 or the RF 24-105mm f4L IS USM at £ 1,119.99) its a strange way to promote entry into a camera system. The RF body is £ 1,399.99 with the mount adaptor and the EOS R £ 2,349.99 with mount adaptor.
With nothing yet mentioned on the road map aside from the RF 24-240mm which while compact for a 10-1 zoom is still not your average walk-around lens its a strange situation.
 
  • Like
Reactions: 1 user
Upvote 0
The lenses shown in the photo are the Heliar and the EF 17-40mm, not the 11-24mm.

Yes, but only because I don't own the 11-24. I used the 17-40 for scale, but I did use the words "1/3 the size and weight of the nearest Canon lens." Considering that the 17-40 is an inexpensive featherweight, I figured it would be understood that the nearest Canon lens would be the one with the 11mm focal length.
 
Upvote 0
I doubt that, except perhaps for the TS-E 17. Certainly the RF lenses we’ve seen so far don’t support that viewpoint. The RF 50/1.2 is sharp, but so is the Otus for EF. Canon has patents for EF f/2 standard zooms (and they couldn’t manage to get the RF version out to 24 mm, personally I’d take the wider FOV over the stop of light in a standard zoom), and the RF 24-105 is essentially identical to the EF version. Maybe Canon’s market-speak will turn out to be true in some cases (e.g. a UWA zoom), but we haven’t seen that yet.

Every T/S lens in the lineup except the 90 is a retrofocus design. Getting rid of that can lead to less distortion, greater light transmission, more compact size,and better corner sharpness. The latest Canon lenses do a better job than the older models, but there's always room for improvement. ;)
 
  • Like
Reactions: 1 user
Upvote 0
Jul 21, 2010
31,091
12,855
..new optical designs could absolutely smoke existing ones were they to use the new lens mount.
...less distortion, greater light transmission, more compact size,and better corner sharpness. The latest Canon lenses do a better job than the older models, but there's always room for improvement.
Sorry, but:
  • less distortion for lenses with only ~1% distortion
  • greater light transmission (unlikely to be more than a stop) for lenses which are commonly used stopped down
  • more compact size for lenses the size of a standard zoom that are usually used on a tripod, and
  • better corner sharpness for lenses which are already quite sharp in the corners
...sounds nothing like 'absolutely smoking' the existing lenses. Perhaps if a TS-E 17 for the RF mount could be made without the bulbous front element and thus take front filters, that would 'absolutely smoke' the current TS-E 17. But given the need to support an effective FoV of an 11mm rectilinear lens, I don't see that happening (and short of that, the ability to use the current TS-E 17 on an EOS R with an ND or CPL behind it far outweighs any minor improvements in distortion or sharpness in an RF version, at least for me).

The RF mount is nice and offers some interesting possibilities, but I think you're 'absolutely smoking' something if you believe it's some sort of miracle mount. ;)
 
Upvote 0