The Canon EOS R5 Mark II coming in Q2, 2023? [CR2]

[StefanoEsse]

Read the first note under the graph:
"Notes:
These raw values are not appropriate for comparing camera models because they are not adjusted for gain or area."

So yes, you are wrong.

Thank you for your comment. You have to compare similar sensor area and density. It is clear and logic. The point has been clarified by Bryan years ago (https://www.the-digital-picture.com...d-of-the-R5-for-the-Lowest-Image-Noise-Levels):

" Comparing same-size imaging sensors, the lower the resolution, the larger the photosites. Larger pixel wells can collect photons at a higher rate than smaller ones, generating a higher SNR (Signal to Noise Ratio) that results in lower noise levels. Therefore, do not expect pixel-level noise performance from an ultra-high-resolution imaging sensor to match that from a similar generation low-resolution imaging sensor.

That said, the final output size is what matters in the real world. To make the Canon EOS R5 vs. Canon EOS R6 comparison relevant, the R5 image (oversampled in this case) must be reduced to 20 MP. An R5 image can be very simply downsized to R6 image dimensions, and then the R5 noise levels appear at least as good the R6 noise levels.".

But if you do not dowsize an R5 image, the R5 one could have more noise at the pixel level, it seems to me.
Again, maybe I'm wrong (and ever happy to learn something new).

 

[AlanF]

Thank you for your comment. You have to compare similar sensor area and density. It is clear and logic. The point has been clarified by Bryan years ago (https://www.the-digital-picture.com...d-of-the-R5-for-the-Lowest-Image-Noise-Levels):

" Comparing same-size imaging sensors, the lower the resolution, the larger the photosites. Larger pixel wells can collect photons at a higher rate than smaller ones, generating a higher SNR (Signal to Noise Ratio) that results in lower noise levels. Therefore, do not expect pixel-level noise performance from an ultra-high-resolution imaging sensor to match that from a similar generation low-resolution imaging sensor.

That said, the final output size is what matters in the real world. To make the Canon EOS R5 vs. Canon EOS R6 comparison relevant, the R5 image (oversampled in this case) must be reduced to 20 MP. An R5 image can be very simply downsized to R6 image dimensions, and then the R5 noise levels appear at least as good the R6 noise levels.".

But if you do not dowsize an R5 image, the R5 one could have more noise at the pixel level, it seems to me.
Again, maybe I'm wrong (and ever happy to learn something new).

That is a very good way of looking at it. I think the basic way of looking at it is that DR, S/N etc are all measured by viewing the image at the same physical size, for example on the photonstophotos site it's equivalent approximately to an 8" wide print viewed at arms length. People get confused by this fact as they read that smaller pixels have more noise and lower well depth and think of everything at the the pixel level. The physics is quite simple. The S/N of the image at high iso is just due to the fluctuations in the number of photons ( = sqrt number) hitting the sensor, and this depends on the area of the sensor and not the number of pixels, as long as the gaps between them are effectively small (and/or the light collected by microlenses) and all else being similar. If you compare, say, a 1000px x 1000px crop on an R6 with a 1000px x 1000px crop on the R5 the R5 will, of course be noisier because it has a small area.

 

[neuroanatomist]

But if you do not dowsize an R5 image, the R5 one could have more noise at the pixel level, it seems to me.
Again, maybe I'm wrong (and ever happy to learn something new).

Yes, the R5 will have more noise at the pixel level. Why does that matter? Would you rather compare pixels, or pictures? Say you need to print at a size large enough that the R5's image is at 1:1. In that case, the noise would be more evident than in the same picture takes with the R6. However, the R6 will have lower resolution and thus will suffer from a loss of sharpness. Of course, you could apply the heavier NR to the R5 image, which would remove more of the noise...and bring the sharpness down so it's more comparable with the lower resolution R6.

The bottom line is that for all practical purposes, noise at the pixel level is irrelevant. Only noise at the image level matters, and that noise is proportional to the area of the sensor, not to the size of the pixels the compose the sensor.

 

[Kit.]

Arguably, when using an AI denoiser, the higher pixel count may result in both cleaner and sharper images.

 

[David - Sydney]

What cameras with .CR3 output have mRAW/sRAW options? As far as I'm aware, that stayed with .CR2 and cRAW replaced it with .CR3.

As far as I know, only the 5Ds/r released in 2015 had mRaw/sRaw options due to the high mp sensor. The .CR3 format came out with the M50 and R bodies in 2018.

 

[AlanF]

Arguably, when using an AI denoiser, the higher pixel count may result in both cleaner and sharper images.

You can see this in practice with DxO PL6 and Topaz AI noise reduction with the R7.

 

[Michael Clark]

As far as I know, only the 5Ds/r released in 2015 had mRaw/sRaw options due to the high mp sensor. The .CR3 format came out with the M50 and R bodies in 2018.

My 15MP 50D, 18MP 7D, 20MP 7D Mark II, and 30MP 5D Mark IV all have mRAW (≈75% linear resolution) and sRAW (≈50% linear). My 21MP 5D Mark II and 22MP 5D Mark III have sRAW 1 (≈66%) and sRAW 2 (50%). The 1Ds Mark III and 1D Mark III had a single sRAW (50%), as did the 40D. The 1D Mark IV, 1D X, and 1D X Mark II had mRAW (75%) and sRAW (50%).

sRAW, and then shortly thereafter mRAW as well, were included among all of the higher tier bodies from around 2007 on until .cr3 replaced .cr2 with the introduction of DiG!C 8 in 2017 (M50, EOS R).

Every DiG!C III xD and X0D introduced in 2007 had a 50% linear sized sRAW, and from the DiG!C 4 introduction in 2008 through the end of DiG!C 7 every xD or x0D had either two sizes of sRAW (≈66%, 50%) or both mRAW (75%) and sRAW (50%).

 

[Michael Clark]

The R came out 4 years ago.

It's been five years since the EOS R was introduced in late 2017. The RP was the 2018 model.

 

[Michael Clark]

This side of the Atlantic, poo-poo is baby talk for excrement while pooh-pooh means dismissing an argument as fallacious.

Over here both poo-poo and pooh-pooh are used interchangeably and can mean either one.

 

[Michael Clark]

Waiting. Waiting. And more waiting.

The general vibe in the marketplace (in my opinion) just seems to be that Canon is following along in the advancement of mirrorless. We were all on the edge of our seats when the EOS M came out, thinking it was going to be some grand foray into Canon's mirrorless play... but the R5 didn't come until 8 years later.

EIGHT. YEARS.

WTF?

The EOS M was introduced in 2012.
The EOS R was introduced in 2017.

Since when is from 2012 to 2017 EIGHT. YEARS?

WTF? Indeed!

 

[neuroanatomist]

It's been five years since the EOS R was introduced in late 2017. The RP was the 2018 model.

Fact check #1: The EOS R was announced on September 5, 2018.

The All-New EOS R System | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Evolution Meets Revolution as Canon Announces The All-New EOS R System. MELVILLE, N.Y., September 5, 2018.

www.usa.canon.com

Fact check #2: The EOS RP was announced on February 13, 2019.

EOS RP Full-Frame Mirrorless Camera | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Full Frame For The Masses! Canon Introduces Its Second Full-Frame Mirrorless Camera — The EOS RP. February 13, 2019.

www.usa.canon.com

Fact check #3: It is 2022, therefore it has been four years since the EOS R was introduced (use a calculator if you need to).

The EOS M was introduced in 2012.
The EOS R was introduced in 2017.

Since when is from 2012 to 2017 EIGHT. YEARS?

WTF? Indeed!

Repeat of Fact check #1: The EOS R was announced on September 5, 2018.

The All-New EOS R System | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Evolution Meets Revolution as Canon Announces The All-New EOS R System. MELVILLE, N.Y., September 5, 2018.

www.usa.canon.com

The period from 2012 to 2018 comprises SIX. YEARS. (not eight, silly but real mistake on my part, and what I get for multitasking – thanks, @SteveC ). Still, either way it's more than the FIVE years between 2012 and 2017.

Score for @Michael Clark: 0 of 4 correct statements. Record fail? Indeed!

We all know you have a nearly insurmountable difficulty admitting you're wrong, but that's a personal best for number of completely false statements in two posts. Ermmm...congratulations?

 

[SteveC]

The period from 2012 to 2018 comprises EIGHT. YEARS. (Again, use a calculator if you need to.)

Score for @Michael Clark: 0 of 4 correct statements. Record fail? Indeed!

We all know you have a nearly insurmountable difficulty admitting you're wrong, but that's a personal best for number of completely false statements in two posts. Ermmm...congratulations?

Even being inclusive on both ends, January 1 2012 through December 31 2018, that's seven years (minus one day). Exclusive on both ends (Dec 31 2012 through 1 Jan 2018), it'd be five years (plus one day). Realistically with random dates at both ends, it's about 6 years; but there's just no way to torture it into being 8.

 

[entoman]

Canon will easily offer those options: R1 for 90MP, R5 for 45MP, and R6 Mark II for 24MP.

Yes they could. But my point is that I (and probably others) would like all those RAW resolution options in a single body, i.e. the R1, with the lower resolutions uncropped, which is perfectly possible via pixel binning, and would retain the full colour gamut and DR (unlike JPEG).

Having said that, it's highly unlikely that I'd get a R1 as I couldn't justify the cost, so I'll stick with my R5 for at least one more year, and probably longer, unless the R5 Mkii provides really worthwhile additional features, which is rather unlikely.

 

[Michael Clark]

Fact check #1: The EOS R was announced on September 5, 2018.

The All-New EOS R System | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Evolution Meets Revolution as Canon Announces The All-New EOS R System. MELVILLE, N.Y., September 5, 2018.

www.usa.canon.com

Fact check #2: The EOS RP was announced on February 13, 2019.

EOS RP Full-Frame Mirrorless Camera | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Full Frame For The Masses! Canon Introduces Its Second Full-Frame Mirrorless Camera — The EOS RP. February 13, 2019.

www.usa.canon.com

Fact check #3: It is 2022, therefore it has been four years since the EOS R was introduced (use a calculator if you need to).



Repeat of Fact check #1: The EOS R was announced on September 5, 2018.

The All-New EOS R System | Press Release | Canon U.S.A., Inc.

Canon Press Release Details: Evolution Meets Revolution as Canon Announces The All-New EOS R System. MELVILLE, N.Y., September 5, 2018.

www.usa.canon.com

The period from 2012 to 2018 comprises SIX. YEARS. (not eight, silly but real mistake on my part, and what I get for multitasking – thanks, @SteveC ). Still, either way it's more than the FIVE years between 2012 and 2017.

Score for @Michael Clark: 0 of 4 correct statements. Record fail? Indeed!

We all know you have a nearly insurmountable difficulty admitting you're wrong, but that's a personal best for number of completely false statements in two posts. Ermmm...congratulations?

Concerning #1, #2, and repeat #1:

I was, as Uncle Roger likes to say, "... um, less correct than I originally thought." In other words, I was wrong. I was looking at the chart in this Wikipedia template and got lost in the columns. Maybe the dyslexia that has been getting worse as I am evidently developing early onset dementia had some effect. Maybe not.

Merry Christmas!

Concerning #3: Context is everything. The comment to which I was replying was the one which made the claim that it had been (and I quote to make it more obvious for the more obtuse among us):

"EIGHT. YEARS."

"WTF?"

My apologies for replying by using some of the exact same words the comment to which I was replying used verbatim without making it so obvious that the words I used were the exact same words as the end of the comment to which I was replying. I can see how that might confuse some simpleminded readers.

 

[Michael Clark]

Yes they could. But my point is that I (and probably others) would like all those RAW resolution options in a single body, i.e. the R1, with the lower resolutions uncropped, which is perfectly possible via pixel binning, and would retain the full colour gamut and DR (unlike JPEG).

Having said that, it's highly unlikely that I'd get a R1 as I couldn't justify the cost, so I'll stick with my R5 for at least one more year, and probably longer, unless the R5 Mkii provides really worthwhile additional features, which is rather unlikely.

And my point is that Canon, who gets to decide what each of their products offer and do not offer, would prefer that you buy all three of those cameras instead of just one of them.

 

[neuroanatomist]

Concerning #1, #2, and repeat #1:

I was, as Uncle Roger likes to say, "... um, less correct than I originally thought." In other words, I was wrong.

It’s a Saturnalia miracle!

 

[Del Paso]

You can see this in practice with DxO PL6 and Topaz AI noise reduction with the R7.

OK...
And once again you "sold" me another Topaz product.

 

[Ozarker]

9,44 M-dot EVF sounds almost too good!
Give it an eye-control AF, and it will be the perfect camera for me.
Or is it just a wet dream???

Every camera I've had was the perfect camera.

 

[Ozarker]

*This is just me*, but...

I'm happy with 45MP and don't want or need larger file sizes or more MP *unless* the camera also has an option to produce uncropped, uncompressed pseudo-RAW images.

I don't want or need 30fps. More important to me is having slow/medium/fast burst options with electronic shutter, and ability to bracket exposures with electronic shutter.

Pixel-shift isn't much use to me unless the whole sequence can be carried out quickly enough to permit hand-held shots and shots with a small degree of subject movement.

Video doesn't interest me, I wish they'd just produce an additional video-orientated model, and put a tilting-flippy screen (Panasonic-style) on the stills-orientated version.

9.44 EVF is very welcome as more resolution makes it easier to visually judge sharpness and depth of field. I don't care if the res drops during bursts. I sure as hell hope they've reduced the EVF lag.

Not sure about dual CFE. I currently have 4 CFE cards and shoot backups to the second (SD) card on my R5. With 2 CFE slots it would mean I'd need another 4 CFE cards, and they ain't cheap.

Sounds like you need to buy my original R.

 

[Del Paso]

Every camera I've had was the perfect camera.

So, I guess you didn't own a Sony...