Canon EOS R1 Spotted in the wild at the Monaco GP

[koenkooi]

I don't think so

GitHub - lclevy/canon_cr3: Describing the Canon CR3 fileformat from Canon M50 / EOS R / SX70 HS / EOS RP...

Describing the Canon CR3 fileformat from Canon M50 / EOS R / SX70 HS / EOS RP... - lclevy/canon_cr3

github.com

The parsing tool returns '14' for depth in CMP1 for an R8 ES shot:

Mac-Studio:canon_cr3 koen$ ./parse_cr3.py 20240526\ 1345\ IMG_4857\ Canon\ EOS\ R8\ -\ Canon\ RF\ 100mm\ F2.8L\ MACRO\ IS\ USM\ -\ 0.31\ m\ -\ Electronic\ Shutter\ -\ CRAW\ -\ IMG_4857.CR3  -v 3 | grep cmp1
trak2 cmp1(iw=1624, ih=1080, tw=1624, th=1080, d=14, p=4, cfa=0, extra=0, wl=3, b35=0, hsize=872)
trak3 cmp1(iw=6188, ih=4120, tw=6188, th=4120, d=14, p=4, cfa=0, extra=0, wl=3, b35=0, hsize=872)

 

[Quarkcharmed]

The parsing tool returns '14' for depth in CMP1 for an R8 ES shot:

Mac-Studio:canon_cr3 koen$ ./parse_cr3.py 20240526\ 1345\ IMG_4857\ Canon\ EOS\ R8\ -\ Canon\ RF\ 100mm\ F2.8L\ MACRO\ IS\ USM\ -\ 0.31\ m\ -\ Electronic\ Shutter\ -\ CRAW\ -\ IMG_4857.CR3  -v 3 | grep cmp1
trak2 cmp1(iw=1624, ih=1080, tw=1624, th=1080, d=14, p=4, cfa=0, extra=0, wl=3, b35=0, hsize=872)
trak3 cmp1(iw=6188, ih=4120, tw=6188, th=4120, d=14, p=4, cfa=0, extra=0, wl=3, b35=0, hsize=872)

Hmm interesting. They probably do 12-bit readout and then scale it up for simplicity.
However the number of bits per sample parameter is a variable, it's not a constant in the format specification. If they change the readout depth, they don't necessarily have to change the format. But decoders should correctly support the but depth as a variable that can be > 14.

 

[EOS 4 Life]

Does it? The problem is, what are actually the rumored specs of the R5 II espcially when it comes to readout speed?

Faster than the a1, Z 9, and Z 8.

 

[EOS 4 Life]

If it also shares the same mechanical shutter as the 1dxIII it will have 16fps

The 1DX III has 20 FPS MS in Live View which would be equivalent to a mirrorless camera.

 

[EOS 4 Life]

16 bit readout in mechanical sounds interesting, I don't think Sony offers anything like that except in A7S, and that's video.

16-bit video from Sony cameras is a complete fabrication made by Sony themselves.
Please don't help spread their lies.
It is 12-bit external RAW video.

 

[EOS 4 Life]

The R1 looks like a smaller camera, maybe it's the photo. I've been using the R5. 24 MP? That's a small high quality print, 11 x 14 inches compared to the R5's 18 x 27 inches.

It is a fatter camera which makes it look smaller.
It is actually larger.

 

[EOS 4 Life]

Upload directly from courtside without having to pipe through the CameraConnect app?

Heck, upload instantly.
That would save a lot of time.

 

[EOS 4 Life]

Would it have hurt Canon to use a 45 MP sensor instead of 24?

Yes, and no.
45 MP would hurt the target market of sports and photojournalism.
What confuses me is why they do not target that market with the R3 II and the high-end general market with the R1.
That is what Sony is doing with the a9 III and the a1.

 

[EOS 4 Life]

I totally butchered the process of moving the posts about the cryptography thing. So the missing messages are my mistake.

Wouldn't having an eSIM help for the cryptography?
The way Canon is implementing it would require outside communication.
Of course, it could always keep a ledger on your phone.

 

[Canon Rumors]

Wouldn't having an eSIM help for the cryptography?
The way Canon is implementing it would require outside communication.
Of course, it could always keep a ledger on your phone.

I am forcing the issue a bit based off a bad google translation and a press release from 6 months ago.

I'm going to chill until someone that might actually know something, says something.

 

[koenkooi]

Hmm interesting. They probably do 12-bit readout and then scale it up for simplicity.
However the number of bits per sample parameter is a variable, it's not a constant in the format specification. If they change the readout depth, they don't necessarily have to change the format. But decoders should correctly support the but depth as a variable that can be > 14.

Since the R1 is a flagship for Canon, I expect Adobe to publish support for it a few days after it starts shipping. For DxO it will likely involve buying a whole new version of their tools and most likely only in 2025.

 

[uPhotic]

The 1DX III has 20 FPS MS in Live View which would be equivalent to a mirrorless camera.

I thought live view was it acting as a e shutter but I'll stand corrected

 

[koenkooi]

I thought live view was it acting as a e shutter but I'll stand corrected

Going into liveview mode unlocks the e-shutter menu option, but you still get to pick between MS, EFCS and ES. The mirror being locked up is the most likely explanation for the increased FPS.

 

[uPhotic]

Going into liveview mode unlocks the e-shutter menu option, but you still get to pick between MS, EFCS and ES. The mirror being locked up is the most likely explanation for the increased FPS.

Thank you for explaining, makes sense. So that would reflect at least 20fps shutter in the R1, if they don't give the cripple hammer.

 

[koenkooi]

For me, it’s a pain to use! Enabling it is a tedious as usual, you might be able to assign it to a button or put it in ‘My Menu’. The capturing part is OK, not great. But then it locks up the camera till it writes out the buffer to card completely. That can take a full minute! So you can’t take any photo’s during that time.
And afterwards you have a single, huge file that only DPP4 can read. And DPP4 doesn’t allow you to extract all the frames at once, you have to manually export them, one at a time.

Contrast that to other vendors where the pre-capture mode doesn’t lock you out of the camera and all the frames are regular, separate files on the card.

I have to amend that a bit While looking at parse_cr3.py to run the 14-bit experiment for @Quarkcharmed I came across https://github.com/dnglab/dnglab. A quick build of that and going outside to make a roll:

Mac-Studio:Rolls koen$ dnglab convert --image-index all 20240527\ 1252\ IMG_5014\ Canon\ EOS\ R8\ -\ Canon\ RF\ 100mm\ F2.8L\ MACRO\ IS\ USM\ -\ 0.29\ m\ -\ Mechanical\ Shutter\ -\ RAW\ -\ CSI_5014.CR3 CSI_5014.dng

Writing DNG output file: CSI_5014_0003.dng
Writing DNG output file: CSI_5014_0005.dng
Writing DNG output file: CSI_5014_0007.dng
Writing DNG output file: CSI_5014_0000.dng
Writing DNG output file: CSI_5014_0004.dng
Writing DNG output file: CSI_5014_0002.dng
Writing DNG output file: CSI_5014_0006.dng
Writing DNG output file: CSI_5014_0001.dng
Writing DNG output file: CSI_5014_0015.dng
Writing DNG output file: CSI_5014_0009.dng
Writing DNG output file: CSI_5014_0012.dng
Writing DNG output file: CSI_5014_0013.dng
Writing DNG output file: CSI_5014_0011.dng
Writing DNG output file: CSI_5014_0014.dng
Writing DNG output file: CSI_5014_0010.dng
Writing DNG output file: CSI_5014_0008.dng
Writing DNG output file: CSI_5014_0020.dng
Writing DNG output file: CSI_5014_0022.dng
Writing DNG output file: CSI_5014_0016.dng
Writing DNG output file: CSI_5014_0023.dng
Writing DNG output file: CSI_5014_0019.dng
Writing DNG output file: CSI_5014_0021.dng
Writing DNG output file: CSI_5014_0017.dng
Writing DNG output file: CSI_5014_0018.dng
Converted 24/24 files
Total time: 283.12s

DxO PR4 accepts the resulting DNGs, but it doesn't crop out the sides of the image that are used for e.g. blackpoint calibration. This might be intentional behaviour, you can do all kinds of neat things if you have access to the whole sensor and the software is called dnglab

This makes pre-capture slightly less useless for me, but only slightly.

 

[TwoWheeler]

No gigapixel sensor, galactic shutter and no flippy screen that can do all the positions of the Kama Sutra?

They’re doomed! Doomed, I say!

 

[Ozarker]

24mp 6000 x 4000

30mp 6709 x 4473

Bitrate and readout are way more important than a meaningless jump in pixel count.

Meaning is in the heart of the beholder, I think. Going from a 5D Mark III to the 30mp R was huge for what I do. The R5 is probably better for me. High bitrate and no problem attaining a meaningless MP.

 

[amorse]

16 bit readout in mechanical sounds interesting, I don't think Sony offers anything like that except in A7S, and that's video.

Will the R5II also have 16-bit readout? That'd be tempting for landscape. But then they'd need to significantly reduce the read noise too (for either R1 or R5II).
16 bits is actually almost groundbreaking, unseen (as far as I'm aware) in FF cameras.

Yeah the 16 bit piece is what has my attention more than anything else in the rumour. If that's legit and makes its way to the R5II, that could be pretty compelling!

 

[Quack]

I have used topaz and DXO for a while. It’s great what you can achieve with software these days but it doesn’t beat real detail, and I’m speaking from an R5 and GFX 100 perspective. Horses for courses but I don’t get how or why the R1 wouldn’t be able to match an older Z9 in the resolution department.
Now of course if it’s sports and sports only and all one is interested in is a football player filling the frame or a car, why not, but the wildlife photographer in me feels left out!

Maybe you - and other wildlife photographers are not the target market. Maybe Canon thinks you - and all those others who want 45 MP - are smart enough to understand that THEY OFFER A PRO LEVEL 45 MP CAMERA. So why should the R1 match the Z9 in MPs if the R5 and rumored R5 II can match the Z9? The idea that all the top level cameras should be the same MPs is stupid, in my opinion. They serve different target markets, they excel in different things. Therefore different MP counts. Is that really so hard to understand?

 

[Canon Rumors]

Maybe you - and other wildlife photographers are not the target market. Maybe Canon thinks you - and all those others who want 45 MP - are smart enough to understand that THEY OFFER A PRO LEVEL 45 MP CAMERA. So why should the R1 match the Z9 in MPs if the R5 and rumored R5 II can match the Z9? The idea that all the top level cameras should be the same MPs is stupid, in my opinion. They serve different target markets, they excel in different things. Therefore different MP counts. Is that really so hard to understand?

That missing mechanical shutter in the Z9 kills that thing for a lot of people. Sometimes it's simply about the trade-offs for a consumer.

For now, I'm trying to spread some vibes for the R3 to become the R3 S. Same body size, big megapickles with a standard set of video features. (I personally don't like battery grips).