Canon announces development of the EOS R5 full-frame mirrorless camera

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
I have never said nor implied this: "... T-stops going up in proportion to the magnification ..."
A zoom with a constant f-stop number: What does that mean for the aperture diameter? (example of 70-200 f/2.8)
- @ 200mm --> aperture diameter of 71mm
- @ 70mm --> aperture diameter of 25mm
So, while the front lens element is always >= aperture diameter, there is no decrease of t-stop value to expect!
Your graph proofs what I was stating.

"However, if you do magnify in front of the aperture then the t-stop is quite different from the f-stop value. While that can be the case for a zoom at some focal length, it makes no sense for a prime."

These three popular primes aren't any different than the zooms cited earlier. They're all about 1/3 to 1/2 stop slower T-stop than f-number.

20200216ss4.png
 
Last edited:
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
I have owned three over the years and copy to copy variance is not really a thing for this lens like an original 24-105L or an original 100-400, ugh those were all over the place.

Primes will almost always show less variation out of the factory than zooms. Anyone who reads Uncle Roger's blog knows that.

But primes can be dropped and knocked out of proper alignment, too. Anyone who reads Uncle Roger's blog knows that, too.
 
  • Like
Reactions: 1 user
Upvote 0

koenkooi

CR Pro
Feb 25, 2015
3,575
4,110
The Netherlands
RAW video will almost certainly be 12-bit, but so will the 20fps stills shooting mode. Even on the 1DX3 20fps e-shutter mode is 12-bit (as confirmed by fine print in the manual), and that's at half the number of megapixels.
I read the manual yesterday and couldn’t find it, which section is it in? It took me 3 readings to locate it in the M50 manual, Canon is pretty good at hiding it :)
 
  • Haha
Reactions: 1 user
Upvote 0
What is needed for an R1 EVF?

Having done action photography since I was a child, I am very curious what is needed to make an EVF work for it (broadcast cameras do it)? I have used an EVF on video cameras and Powershots years ago, but not for action. I got hands on my first MILC yesterday (the R) and was pleased with the resolution and general quality of the EVF, until moving the camera with my eye up to it. It was horrific, and I don't think I could shoot playing puppies, let alone BIF or motorsports. I think we may have to evolve to use the LCD only, or some other solution like tethered Google glasses. Thoughts?
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
if it's a July release, i doubt WPPI will be able to tell us much other than just have units under display.

There's CP+ , PhotographyShow, NAB, PhotoKina all lined up before this baby gets released. If this was imminent release, then sure, they would have live models that you could play with that show.


I can't see that happening this far away from release.

yes the hardware is usually finished well before now, but usually the firmware isn't done yet.

If I was looking at a crystal ball. this is what I think we'll see at the shows;

- WPPI - units under display. Maybe touch / hold non functioning units.
- NAB - 8K on display.
- PhotoKina - fully operational models, CFE cards locked up though

on the 20fps .. it's REALLY a lot of data to move.

I wouldn't be surprised to see AF/AE locked for 20 fps. it's not a question of crippling, it's a case of it being 2.25x the data that the 1DX Mark II is processing at 20 fps.

we're talking somewhere in the range of 1.8 Gigapixels / second - just to put this into perspective, and far more than that needed to be processed.

however - conversely - it sounds from the 1DX Mark III whitepaper that DIGIC X has ASICS in it specifically for DPAF AF. Which is a first.

So while it MAY happen, i'm leaning to AF/AE locked on all the extreme performance options in the R5;
- 20 fps
- 8K30p
- 4K120p
That way i'll be pleasantly surprised if they do have DPAF / AF/AE.

Sony, et all fanboys will call this "crippling" i'm sure. however, it's doing things that no other ILC can do at the moment - not sure how you'd call that crippling.

I wonder how many of theses shows will also be impacted by the coronavirus event, since the camera industry is centered on the Pacific rim?
 
Upvote 0
My question was:
when comparing R5's maximum FPS to the 1DX3, is it being overlooked that the real feat is that the 1DX3 has a virtually unlimited buffer? couldn't the R5 also be capable of 20 FPS RAW, but just have a much more limited buffer, and simply have to take a longer break between shutter presses while writing to disk?

8k 30fps is continuous, because that is the nature of video.
Compression aside, that transfers more pixels to disk than
45 MP @20fps do.

Compression is done in DSPs that are specifically designed
for this - likely no bottleneck.

Provided that the R5 employs double CFexpress cards as well,
it will very likely have unlimited buffer just the same.
 
Upvote 0
What is needed for an R1 EVF?

Having done action photography since I was a child, I am very curious what is needed to make an EVF work for it (broadcast cameras do it)? I have used an EVF on video cameras and Powershots years ago, but not for action. I got hands on my first MILC yesterday (the R) and was pleased with the resolution and general quality of the EVF, until moving the camera with my eye up to it. It was horrific, and I don't think I could shoot playing puppies, let alone BIF or motorsports. I think we may have to evolve to use the LCD only, or some other solution like tethered Google glasses. Thoughts?
At least 120fps progressive with 5,7M dots.
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
What is needed for an R1 EVF?

Having done action photography since I was a child, I am very curious what is needed to make an EVF work for it (broadcast cameras do it)? I have used an EVF on video cameras and Powershots years ago, but not for action. I got hands on my first MILC yesterday (the R) and was pleased with the resolution and general quality of the EVF, until moving the camera with my eye up to it. It was horrific, and I don't think I could shoot playing puppies, let alone BIF or motorsports. I think we may have to evolve to use the LCD only, or some other solution like tethered Google glasses. Thoughts?

Broadcast cameras have lag also, though not to the same extent as some MILCs do, but there's no need to precisely time "peak action" in one frame. So the camera operator can just look at the world through the monitor and it works well enough for video.
 
Last edited:
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
For some reason, Canon didn't include 24p in the RP and 90D (but there was 30p), which caused a lot of stirring and boiling, they fixed it in subsequent firmware updates. That's one of the reasons (but not the only one) for the 'Canon's cripple hammer' term.
Generally, specs at this stage are the "maximum" capability of the hardware, not all available modes. An f/2.8 lens is also capable of other apertures up to f/22 or f/32, but the marketing is going to concentrate on the ability to do f/2.8, not f/5.6 too.
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
MPB sell them quite often, and at "acceptable" prices.
But, as you wrote, the risk is getting a poor copy, no risk with serious dealers offering an exchange guaranty.

I think MBP gives a 15 day return guarantee. I've seen the same Sigma 120-300mm f/2.8 Sports listed three or four times in the past few months with 1-2 week gaps in between when the lens is listed as "out of stock." If it is not the same copy of the lens, they're all rated the same condition.

Wow, they have nine currently for sale. That's a big change from a few years ago when Iwas looking for one.

I guess the Sigma 135 ART, with its 2D flat test chart performance, made everyone want to dump their 135 L! I still think the 135 L has better out of focus backgrounds than the Sigma.
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
And that's why I never base my buying decision on Lens Rentals full aperture tests.
They simply don't do a lens justice. (coma, color rendition, bokeh, etc...)
What really matters, is how a lens performs in a situation it was designed for, and that's never flat-brick-wall -photography (exception :TSE lenses).

TS-E as well as Macro and 2D document reproduction systems, which tend to use highly corrected LF lenses and digital scan backs these days.
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
In my EOS R there is no fast way to change fps. Either multifunction button which behaves like a menu so not immediate or using the touch screen after activating the buttons. Not cool! Not fast. YMMV.

Where is the fast way to change fps on a 5D Mark IV/III? Or are you talking about changing Drive Mode from 'Single' to 'Continuous'?
 
Upvote 0

koenkooi

CR Pro
Feb 25, 2015
3,575
4,110
The Netherlands
Page 945, or search for "12 bit". :p

I see it now, thanks! So both raw video and e-shutter are 12 bit. I shot in 12 bit mode on my M50 for months, where I kept thinking "This is supposed to have more DR than my RP, but there's almost no highlight recovery". After setting the drive mode to a slower FPS to get 14 bit RAWs again I finally noticed a positive difference between M50 and RP files.
 
  • Like
Reactions: 1 users
Upvote 0

joestopper

Rrr...
Feb 4, 2020
233
212
"However, if you do magnify in front of the aperture then the t-stop is quite different from the f-stop value. While that can be the case for a zoom at some focal length, it makes no sense for a prime."

These three popular primes aren't any different than the zooms cited earlier. They're all about 1/3 to 1/2 stop slower T-stop than f-number.

View attachment 188755

Thanks for sharing. Confirms what I was saying.
 
Upvote 0

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
Thanks for replying, but your answers are taking me farther and farther away from my original question.

I didn't bring up video in my original question, but I still don't quite get how you arrived at the data size of 20 FPS still images being comparable to 8k30 video (in an as yet unknown format). I don't see how we'd even begin to compare data throughput without certain critical missing variable. But I admit, I'm ignorant about a lot of this.

At any rate, my question was more in response to this post:



My question was:

when comparing R5's maximum FPS to the 1DX3, is it being overlooked that the real feat is that the 1DX3 has a virtually unlimited buffer? couldn't the R5 also be capable of 20 FPS RAW, but just have a much more limited buffer, and simply have to take a longer break between shutter presses while writing to disk?

The general assumption is that at 8K30 the camera's processor would not be able to compress anything, so the output would be uncompressed raw. With video, buffer size is more or less irrelevant - the camera must be able to write to the card at least as fast as the data rate of the video.

But beyond that, max frame rates are measured before the buffer fills. The 5D Mark IV can do 7 fps, but only for about 21 to 30 raw images (depending on card speed), then it slows to anywhere from about 4.4 fps with a fast UDMA 7 CF card to 1.6fps with a fast UHS-I SD card.
 
Last edited:
  • Like
Reactions: 1 user
Upvote 0