This is likely Canon’s lens roadmap for 2020

[Bonich]

I totally agree with that. It's like someone who doesn't want to lug around a 3kg 1.2 prime does not want weather sealing or better build quality either.

I like the EF 35 IS f/2 very much, it is well built and "small".
The RF 1.8 with the changing length (by focus) seems to be more vulnerable and looks poor.

But yes, a RF 35 / 85 with 2 IS would be nice. But because there is a RF35 1.8, there will be many other new lenses before they update this one with a similar (but well-built) pendant....
I don't need a RF 35 1.2 without IS. This is may be superb, but not for my usage.

May be it looks poor in your eyes, the 35 1.8 produces stunning results and is very versatile.
Give it a try, it even changed my shooting style enabling perspectives never been able before.
Yes, it extends focussing: Wide aperture, very compact, 1:2 magnification, limited focus breathing, how to be done without?

 

[Aussie shooter]

Why no AF @F11 with the R series???
Why only very still objects?
Yes, perfect framing without monopod/beanbag, tripod will be some challenge especially with the 800mm.
To give you an Idea about F11 here a shot with 16 years old technology F10, 500mm x1.4, ISO200, 1/800sec (EOS 1D MII):
Sorry canonrumors had stolen the sun reflection point on the eye :-(
View attachment 190774
The very first important issue is to be at the right moment at the right location with the right light and the right subject.

Not too bad having equipment with you. You never take around F4.0 just accidentally. (I know, I use F4.0 600mm or 500mm for nearly 3 decades now).

I disnt say it cant focus with a max apature of f11. I said it is much harder than f4. I know the mirrorless do far better than a dslr at that apature but it still slows everything down. But your example shot(nice shot by the way) is an example of good light and a relatively still subject(a hovering tern is not easy as they dont often hover for long but there is always that period of pause before the dive). I have shot them quite often as they do make such awesome subjects. I have no doubt these will be very good lenses but they will have to be cheap( in comparison to a big white not compared to a nifty fifty) to be worthwhile. Of they are then they will sell like hotcakes despite the limitations they will have

 

[SecureGSM]

You sell a crap-ton of RP and maybe R6 bodies with them. That's what you do.

How you can sell a crap-ton of such a specialty focal length lens in such a stagnant, downward spiral market? 600/11, 800/11..
out of all non pro camera owners left Out there, what percentage of those will ever consider longer than 200mm PRIME lens?
Yeah, 100-400 zoom. May be. Only if compact and cheap. Take a look at Sigma and Tamron 100-400 lenses. F6.3 at the long end. Size of your stubby.
Yup, makes sense.
there is no market to sell tons of these cheap telephoto primes in volume.
Sell tons of RP / R6 bodies? Yup, with 24-105 STM as kit. 100-400 / xx -6.3 or 7?? Yup.

 

[Michael Clark]

Where is the RF 135? The EF 135 f/2 is the oldest lens in Canons lineup, 24 years old and never been updated. Ridiculous.

And still takes great photos.

 

[Michael Clark]

Now do that with an 800mm lens of a little bird sitting on one of the branches in the shaded area under the tree. Because that is what an 800mm lens is for. Not for taking landscape shots of a back garden.

That's one of the things an 800mm lens may be for. It's far from the only thing.

 

[Michael Clark]

First, I believe he was joking about your picture. I personally believe it is VERY good (super sharp, open wings, nice lighting). I do not believe that someone could dislike this specific picture. Anyway this is also subjective.

But allow me to express some concern about your conversions. I am not 100% familiar I admit but I believe that you take into account that you would fill the frame the same in both cases (hence the different distance, the different total light - so different probably signal to noise ratio -on the sensor, the different DOF, etc).

In my case I can select between a crop (20mpixel) and a FF (50mpixel) Add to that the fact that in birding at 90% of the cases we are FL limited. So the distance from the bird would be the same crop or FF. At the same time both cameras I mentioned have approximately the same pixel density 20 * (1.6^2) = 20 * 2.56 ~ close to 50

In other words the same number of same size pixels would exist for the bird. We could either crop a lot in FF case or less at APS-C case. The final pixels would be the same.

Of course we are talking about using the exact same lens in both cases.

Taking all the above into account and assuming similar sensor technologies

1. How can you claim that f/7.1 on aps-c is the same as ff at f/11?
2. How about wanting to use these f/11 lenses on crop cameras (a mirrorless crop could happen just like z50). Assuming your conversions the equivalent ff stops would be a disaster!

P.S Aside all of the above, I believe that 800 f/11 will be a huge success!

The difference for budget limited folks is that a 20 MP EF crop body currently goes for about $1,400. A 50MP EF FF body is still $3,400 from authorized dealers.

In the future it appears the 20 MP crop body will not be part of the RF system. A 20 MP FF body, however, is on the near horizon.

If you can afford a $3,500 camera and $5,000 lenses, these f/11 lenses are not for you.

 

[Michael Clark]

tricky... even 1/180 with 600mm + 2x moves 2-3 pixels and blurs detail. You need 1/500th if shooting around 1000mm to get a sharp image.

The moon is in direct sunlight. Use the "lunar 11" rule of thumb. It's only one stop dimmer than "sunny 16."

Due to the Moon's motion relative to the surface of the Earth, it moves its own diameter every two minutes.

At f/11 one would need to use ISO 800 to use 1/800 second, or ISO 1600 to use 1/1,600.

Now consider that an R mount camera to use with this lens will have IBIS in addition to the lens having IS. The only motion to be concerned with will be that of the subject.

 

[Michael Clark]

I find that overexposure is more of an issue with a full moon than camera shake on a clear night.

Just use the "lunar 11" rule of thumb. It's only one stop slower than "sunny 16." The moon is illuminated by direct sunlight. But if you properly expose the moon, you won't have very many stars, if any, break through the noise floor. And nothing on the ground will be anything other than totally dark unless you have another source of illumination.

 

[Michael Clark]

I'm not seeing much discussion on the 85mm f/2 macro IS STM. That would be an interesting lens. If it can do 1:1, then it could replace both the 85 f/1.8 portrait lens and a 100mm macro for most people, so it seems like a value play. I'm guessing that the 100mm focal length will only have the "L" macro unlike the current offerings in the EF system.

I'm not a fan of using macro lenses for portraits. The flat field correction a lens needs to do macro well tends to kill the smooth bokeh we desire in portrait lenses.

 

[Michael Clark]

And it will be useless shooting football inside a stadium at night too.

It will be a budget lens. Not for every occasion or lighting condition but plenty. I did shot very good images in the dark forest at ISO 800-1600 at F5.6 and 640mm on old Canon 20D/7D so 1 or 2 stop ISO increment should be still usable on a modern full frame.

Of course, you can always buy the $15000 800mm 5.6 and carry all the 4.5 kg weight up to a mountain. I am happy that there will be choice.
Brighter lenses will be there too.

I have a feeling these lenses will be more unusual than the specs show. They will have some weird optics/design.

You can't shoot football in most stadiums at night with an 800/5.6, a 600/4, or a 500/4, either. And those lenses cost a lot more than the 800/11 and 600/11 will. Night stadiums are why 300/2.8 and 400/2.8 lenses are so popular among sports shooters even as expensive as they are. If you're in a larger pro or major college stadium, then f/4 can be useful (which is why 1.4X TC/extenders are also in the bags of most sports pros). But for high schools and small or mid-size colleges, forget it . You need f/2.8. Period.

Not every lens is made to do everything

 

[Michael Clark]

I guess I am treating exposure as equivalent to the concept of "flux" (amount of something flowing through per unit area, used for radiation and electromagnetic fields), and the exposure triangle is just a quickie way of comparing two different sets of settings to see they have the same flux (or how different they are). But photographers never use the word flux (my engineering background betrays itself).

There is, however, one more thing to consider, and that's how much of the light is simply absorbed in the glass elements of the lens itself. If you're using the same lens in two different settings, you can ignore that (since you're making comparisons), but with two different lenses that can become a factor as they might differ.

Once accounting for that I think I can safely think of the two things as equivalent.

It sounds like for noise comparisons you will want to compare the amount of light hitting each pixel, and of course the pixels will have different areas AND might be engineered differently to boot (imagine a pixel off a fifteen year old sensor versus one today). You then have total light hitting an object (flux times its size). (And thus it seems to me the total sensor size is relevant ONLY IF paired with the number of pixels so you can then effectively be using its the pixel size.)

I wonder if all the arguments about "full frame being less noisy than APS-C" or alternatively "Resolution being proportional to noise" are just because people are leaving out half of what determines pixel size (either the count or the sensor size).

In the context of what you're talking about, most noise is due to the Poisson distribution of photons in light. Light's "flux density" is not uniform. The more total photons you gather, the less the randomness of the distribution of photons will affect your result.

 

[Michael Clark]

Here's a thought that a couple of posts have alluded to it, but not clarified.

Surely this is all about luring/tempting customers away from APS-C SLR to full frame?

Not only APS-C , but also Micro Four-Thirds.

If Canon's entry level FF RF body with IBIS and the lenses it can use are as light and cheap as comparable µ4/3 bodies and lenses with half the focal length and max aperture, a lot of folks will move.

A 400mm f/5.6 lens on a µ4/3 body should perform about the same as an 800mm f/11 lens on FF with regard to field of view and noise performance, all else being equally.

 

[SecureGSM]

I'm not a fan of using macro lenses for portraits. The flat field correction a lens needs to do macro well tends to kill the smooth bokeh we desire in portrait lenses.

Very well said. Thank you.

 

[Michael Clark]

Advancements certainly haven't stop, but with regards to most aspects, they have slowed down a lot.

For noise and low light performance for example, there are two aspects that are subject to improvements. There is the efficiency of how much of the light hitting a sensor is converted to electricity and therefore detected. This is currently in the high 80% range for consumer cameras AFAIK. So even if 100% was achievable, that would make things not much better.

And the other aspect is the noise generated inside the camera electronics due to heat (dark current noise) and the read and amplifier circuitry (read noise). The latter is what has improved over the last years a lot. But it also has a lower limit. I don't know where exactly that would be, but obviously the read noise is very small in modern sensors and as it can't go negative, it seems we are approaching the practical minimum. Dark current can be reduced by cooling the sensor, which is actually a big part of dedicated astro cameras. But for general purpose photography, not much heat is generated in the sensor. And it should also be less now that the newer cameras are using more energy efficient parts.

As for megapixels, this is where we have seen most improvements in the recent past. You now can get as far as 32 MP APS-C sensors! Scaled up to FF that would be just over 80 MP. Quite a lot. But there's a limit as well: the diffraction limit. One your resolution has climbed so high that all apertures you frequently use are beyond the diffraction limit, you're getting the most detail for your money. But adding more MP won't help anymore. We are not at that point yet, a 32 MP sensor can resolve all the detail from f/6.3 and upward.

But getting past the diffraction limit involves some pretty specific techniques, so once we are there, that should be Pretty much it for the megapixel race. At the current rate of MP growth, it while take over a decade to get there though.

There are improvements to be had in DR. Canon has demonstrated that with its new Dual Gain Output sensor, and Sony has a different approach with dual amplifier stages that does also improve DR, although far less.

Don't mistake diffraction limited aperture with the much narrower diffraction cutoff frequency. DLA is when diffraction begins to be significant enough to be detected at the single pixel level. You can go several stops beyond the DLA before the increased effect of diffraction becomes noticeable at typical display sizes and viewing distances.

Current circles of confusion used for typical display sizes are about 6-8 pixels wide with the densest current FF and APS-C sensors. So unless you are blowing an image up to 60x40 inches and viewing it from one foot (or viewing a 24 MP image on a 96 ppi monitor at 100%), the DLA doesn't really come into play like most folks think.

 

[Michael Clark]

Consistent with what? With Moore's law? I don't think so.

There's less than 2 stops of increase in high ISO DR of FF cameras in the last 15 years. And now we are already pretty close (within 1 stop) to its theoretical limit.

And most of the improvement made in the last fifteen years was already available by around 2012. From 2005 to 2012 the best sensors on the market of the same size gained almost two stops in terms of DR and base ISO. From about 2012 until 2020 we've gained maybe an additional half stop. It is true that it took Canon a little longer to get here, but others were within about a half stop of the current state of the art seven or eight years ago.

 

[Michael Clark]

F11 it's still too dark for 600mm. For 800mm it might be acceptable. But for 600mm you can add a TC to the 100-400 and still have 1 stop advantage.

I'm guessing these non-L STM lenses will be noticeably cheaper than the 100-400 II before one adds the cost of a TC.

These lenses are not for folks who currently use 100-400 II + 1.4X III combos.

 

[Dexter75]

And still takes great photos.

so does the 15 year old 5D but that’s not the point. We shouldn’t have to be using 24 year old lenses adapted on our brand new mirrorless cameras....

 

[Michael Clark]

I don't think there was any limit on how far apart the two amplifications needed to be. The feature is called dual ISO. I'm pretty sure I used ISO 100 and ISO 1600 at one point, just to try itA. As this was on a T3i (strong, strong fixed pattern noise when pushing shadows too much), you did actually gain something there since the sensor was far from ISO invariant.

Here is the thread the developers assigned for posting images made using the feature:

https://www.magiclantern.fm/forum/index.php?topic=7402

(Obviously, the HDR look is pretty over the top in many there. But that's basically the point)

But thinking about it, for the current generation sensors that are so close to being ISO invariant, using a high and a lot magnification should indeed yield little advantage over just using the lower one and creating the HDR in post... Or maybe I'm too tired and missing something currently.

Edit: then again, Canon is writing 16 stops Dr in their material. So clearly they see an advantage in terms of readnoise and DR. So I guess I'll get some needed sleep

My point is: If you have to reduce exposure to protect highlights for the parts of the sensor amplified at ISO 1600, then at the same time you're also forced to reduce exposure of the parts of the sensor amplifying at ISO 100 by four stops compared to what you could expose if you only needed to protect the highlights at ISO 100. Only if the technology can overcome the blown highlights in the ISO 1600 portions without precipitating weird artifacts will it be very useful. Magic Lantern did not achieve that. One can stretch a single frame raw image using "HDR" software and get similar results to those examples that used ML.

 

[Michael Clark]

You can see the brightest parts but only as a mid-tone. If you want an exposure where the brightest parts of the moon are only value 90 or something, sure, shoot at 1/500 with moderate aperture. But to get the full contrast range of the moon you'll need much higher ISO or longer shutter.

The more you raise ISO, the less DR you have, because the highlights blow out much sooner.

 

[Michael Clark]

May be it looks poor in your eyes, the 35 1.8 produces stunning results and is very versatile.
Give it a try, it even changed my shooting style enabling perspectives never been able before.
Yes, it extends focussing: Wide aperture, very compact, 1:2 magnification, limited focus breathing, how to be done without?

Some folks care more about how a lens looks mounted on their camera and what others will think about how it looks mounted on their camera than they care about how the images they get with that lens look.