Canon to release major firmware update for the Canon EOS R5

[Michael Clark]

I would bet you, you can’t even blind test the differenve with a shot taken in 12 bit vs 14 bit. Go ahead try it. This concern for 14 bit is only different on paper, real world results are mute

It would all depend upon the scene. If the scene exceeds the DR range of14-bits, you'd certainly notice it even more with the 12-bit limitation.

If the scene only contains 6-8-10 stops of DR, then yes, you wouldn't be able to tell the difference.

 

[Michael Clark]

and cRaw?
What would you call an oversampled/binned image in .CR3 format?
Noise reduction in RAW images including the well known star-eater Sony issue seems to muddy all the water.
Ultimately Canon (and Nikon/Sony etc) will provide the highest level quality and manipulation flexibility in their RAW format(s)

cRAW is just a slightly lossy compression when recording the monochrome luminance values from each sensel. Based on what little I've read about it, I'm guessing almost all of that lossy compression takes place below the noise floor. No debayering/demosaicing is done to the data.

Any binning method requires debayering/demosaicing, which means certain things, such as CT, WB, BP, and WP are "baked in". That's not the case with cRAW.

 

[Michael Clark]

I would be interested to understand further... is there a source?

Compare the numbers and do the math for the resolution and frame rates for each camera. Regardless of the sensor resolution or version of Digic X each has, they all match the same analog-to-digital convertor capacity in terms of how many bits per pixel times the number of pixels times the number of frames per second they can process per second.

 

[fr34k]

Compare the numbers and do the math for the resolution and frame rates for each camera. Regardless of the sensor resolution or version of Digic X each has, they all match the same analog-to-digital convertor capacity in terms of how many bits per pixel times the number of pixels times the number of frames per second they can process per second.

AFAIK the R7 can do 14 bit in electronic shutter, no? So that one is substantially higher than the R5's which would translate to 14bit being possible in electronic shutter... But the R6 II and the R5 are very close. (and the normal series mode of the R3 too)

35 MPx * 30 fps * 14 bit = 14700 Mbit/s (R7)
24 MPx * 30 fps * 14 bit = 10080 Mbit/s (R6 II)
45 MPx * 20 fps * 12 bit = 10800 Mbit/s (R5)
24 MPx * 195 fps * 14 bit = 65520 Mbit/s (R3)

 

[koenkooi]

AFAIK the R7 can do 14 bit in electronic shutter, no?

No The R7 is also not 14-bit in ES mode. The manual doesn't specify what it actually is, though.

So that one is substantially higher than the R5's which would translate to 14bit being possible in electronic shutter... But the R6 II and the R5 are very close. (and the normal series mode of the R3 too)

35 MPx * 30 fps * 14 bit = 14700 Mbit/s (R7)
24 MPx * 30 fps * 14 bit = 10080 Mbit/s (R6 II)
45 MPx * 20 fps * 12 bit = 10800 Mbit/s (R5)
24 MPx * 195 fps * 14 bit = 65520 Mbit/s (R3)

The R3 in 195fps mode is very interesting, it doesn't do 'AF-capture-process-save' like all other modes, it will only capture till the buffer full. Only after that has happened it will start processing and saving for half a minute. And you can't use the camera when it's doing that. So the 65Gbit/s isn't sustainable for more than half a second.

So we now know that the sensor can be read at 195fps, and infer that the Digic X is bottlenecking at the processing and saving stage. Canon tends to use a capture in between for AF, so if they have a faster Digic, we could see 97fps with AF on the same sensor.

 

[fr34k]

No The R7 is also not 14-bit in ES mode.

Very interesting, very interesting indeed. I do not have one at hand atm, but I assumed because the manual only specifies the RAW files to be 14 bit that that applies to all shutter modes.

Can someone do me a favour and take two very similar images with EFCS/MS and one with ES and tell me the RAW (not cRAW) file sizes. I'd love to infer them based on the size...
On the R5 that works fairly well...

 

[Delfin]

Hope CAnon add de capability to adjust the shutter sped to less than 20 fps when using electronic shutter

 

[koenkooi]

Very interesting, very interesting indeed. I do not have one at hand atm, but I assumed because the manual only specifies the RAW files to be 14 bit that that applies to all shutter modes.

Can someone do me a favour and take two very similar images with EFCS/MS and one with ES and tell me the RAW (not cRAW) file sizes. I'd love to infer them based on the size...
On the R5 that works fairly well...

Canon pads the bits, so you’ll always get 14-bit CR3 files. People have been using rawdigger in the past to spot dips in the data histogram to see if and when a camera drops down to a lower bit depth.

 

[fr34k]

Canon pads the bits, so you’ll always get 14-bit CR3 files. People have been using rawdigger in the past to spot dips in the data histogram to see if and when a camera drops down to a lower bit depth.

Well that’s low.

AFAIK RAW is lossless compressed so even with padding files should be smaller as padding definitely has lower entropy…
Do you know if the same happens with cRAW? (files being the same size)

 

[jdp_50mm]

Please Canon add the OVF simulation mode from the R3 and R6 mark II to the R5. Removing the video recording limit would be appreciated.

 

[Bonich]

OK. So I was browsing the R5 a bit and there may be a workaround for you that works on the R6 too:
You can set a specific button to turn off the screen. Although you might have to sacrifice a button, this might work for you. Try it out.

This works great on the R with some more buttons, not so with the R5/R6
Why doesn't Canon use click functions under the rear dial as Sony does?
This would give us 4 extra buttons free to be customized.

 

[koenkooi]

This works great on the R with some more buttons, not so with the R5/R6
Why doesn't Canon use click functions under the rear dial as Sony does?
This would give us 4 extra buttons free to be customized.

Canon seems to prefer interfaces that do one thing and don't have surprises, a dial with 4 click positions could be seen as confusing. There have been a few experiments, like the touchbar in the R and the optical joystick button in the 1Dx3 and R3, with mixed results.

I not only agree with you that having a d-pad in the dial would give us 4 extra buttons, but those buttons are in a place that makes them easy to find and hit. I have trouble hitting the 'info' button with 100% accuracy on my R5 because it's in an awkward place, right next to another button.

I also wish that the optical joystick thingy finds its way to more (read: cheaper) cameras. The touchscreen based AF selection on the RP/R8 works great for me, being right eye dominant, till the weather turns into glove season

 

[fr34k]

Please Canon add the OVF simulation mode from the R3 and R6 mark II to the R5.

Can you tell me why you'd like the OVF sim? I used to use real OVFs and since moving to EVFs I really appreciated to see the final image in advance (with WB, brightness etc...)

I don't mind them adding that (actually, the contrary: every additional feature added is a good one), but I'd like to know why you'd like OVF sim.

 

[koenkooi]

[...]Can someone do me a favour and take two very similar images with EFCS/MS and one with ES and tell me the RAW (not cRAW) file sizes. [...]

I've uploaded 2 files to https://drive.google.com/drive/folders/1hWvzqinsSFHEd51nu5WvZW15WrXUew5G?usp=share_link

MBP:Bits koen$ du -ks *
48628    R5 EFCS 13-bit.CR3
49692    R5 ES 12-bit.CR3

MBP:Bits koen$ exiftool -Shuttermode *
======== R5 EFCS 13-bit.CR3
Shutter Mode                    : Electronic First Curtain
======== R5 ES 12-bit.CR3
Shutter Mode                    : Electronic

The EFCS image is smaller, but there was a small change in ambient light (stupid clouds!).

 

[fr34k]

I've uploaded 2 files to https://drive.google.com/drive/folders/1hWvzqinsSFHEd51nu5WvZW15WrXUew5G?usp=share_link

MBP:Bits koen$ du -ks *
48628    R5 EFCS 13-bit.CR3
49692    R5 ES 12-bit.CR3

MBP:Bits koen$ exiftool -Shuttermode *
======== R5 EFCS 13-bit.CR3
Shutter Mode                    : Electronic First Curtain
======== R5 ES 12-bit.CR3
Shutter Mode                    : Electronic

The EFCS image is smaller, but there was a small change in ambient light (stupid clouds!).

This is curious. I seemed to remember when I was on a bird photoshoot the 13bit files to be noticeably smaller than the 14bit, but I also took two photos myself just now and they're basically the same.
I was also originally talking about the R7 as I own an R5 myself but I guess that would then also yield practically no difference.

 

[SHAMwow]

It isn't possible to have *real time* blackout-free bursts with a mechanical shutter that physically closes between exposures. The blackout *can* be eliminated by keeping the first image active until the subsequent image takes its place, but that would result in a juddery display. If you want a real time blackout-free EVF without any juddering, electronic shutter is the only way.

Then I don't really get the question. They're asking for blackout free, but we already have that in electronic. If its only possible in electronic, and we have it, why are we asking for it?

 

[neuroanatomist]

Then I don't really get the question. They're asking for blackout free, but we already have that in electronic. If it’s only possible in electronic, and we have it, why are we asking for it?

Possibly because the only ‘we’ that have it are those of us who own an R3.

 

[entoman]

Possibly because the only ‘we’ that have it are those of us who own an R3.

... and certain Sony models ..... is it OK to use the "S" word here?

 

[AlanF]

... and certain Sony models ..... is it OK to use the "S" word here?

And the Nikon Z9. But, the R5 doesn't really have perceptible blackout.

 

[jdp_50mm]

Can you tell me why you'd like the OVF sim? I used to use real OVFs and since moving to EVFs I really appreciated to see the final image in advance (with WB, brightness etc...)

I don't mind them adding that (actually, the contrary: every additional feature added is a good one), but I'd like to know why you'd like OVF sim.

When I read this article by DP Review it looked appealing. https://www.dpreview.com/articles/2...-ovf-simulation-what-it-is-and-why-it-matters