R1 specs leak

Damn LOL

https://www.canonrumor​s.com …os-r1-specifications-cr0/

Does CR 0 mean that if I say the R1 will have a built in hand warmer, that is good enough to rate "CR 0"?

Sounds great, but I doubt it on many levels.

- I don't think the R1 will be an MP beast at 85mp.
- I truly doubt a global shutter will premier with such a monstrous MP sensor,.
- That DR number is proof that this is just someones fantasy

I believe they will have a high megapixel camera around what the rumor says but I have hard time believing most of the rest of the claims. If I needed that high a pixel count I would just buy a Fuji GFX100.

Based on the CR0 and their own text basically stating that they are bored, the industry is quiet, so they are just throwing out stats to get attention....

Therefore, there is nothing there but a big fat "Click Me" article.

Then there is this from the same site on another article...

I don't take it as they're "throwing it out there to get attention" but more so "Hey, here is "news" that you might have heard going around right now." Specifically saying they wouldn't publish it if it were sent to them first. I mean that's the entire point of the site, right, to publish rumors?

CR0 means no real source coupled with this comment...

Followed by the CR0 in bold and red...

This, to me, feels like just a blog to create traffic, based on the tone of the article and attention to CR0.

The same anonymous source also contacted canon price watch, interestingly. (85Mpx)

https://www.canonwatch​.com …al-shutter-quad-pixel-af/

----
--- as to the specs, here is the history, and I doubt greatly that this will have an 85Mpx sensor... The price however, wouldn't surprise me.
----

Rumors in Jan 2021 (45Mpx)

https://www.techradar.​com …cus-and-key-video-feature


The rumors in Oct 2020 (21Mpx)

https://canoncamerarum​ors.com …ed-to-have-a-21mp-sensor/

So far at this point, there just isn't much in the way of new info or rumor sources, other than the 21 vs 85Mpx sensor.

85Mpx makes sense if it is Quad-Pixel AF and you are counting each pixel in the quad separately. 85 / 4 is about 21Mpx, which is consistent with the current 1 series.

Article is mainly click-bait though.

Naturography wrote in post #19203930
Damn LOL

https://www.canonrumor​s.com …os-r1-specifications-cr0/

The DR seems high, perhaps, but that could be compared to "PDR", which tends to run much lower than DxO-like DR.

Depending on the definition of DR and what file types we are talking about, it might not be all that far-fetched.

The R5 with Dual Pixel RAW has about a stop more DR than a normal RAW in the form of pure extra highlight headroom, even though software generally does not take advantage of it. libraw.org has a downloadable tool that turns DPRAWs into 2 DNG files, and you can convert and blend them as you will, and some future tool may make them into an HDR single RAW, or converters may support the extra DR at some point in the future, (if none of them do so already).

Quad-pixel RAW, if true (hard to imagine in early 2021 in a camera with 85M full pixels, but who knows what is going on in prototypes) may provide even more headroom. Canon could choose to have quad-pixel RAWs, or they could blend the 4 images in-camera to create a single-image HDR RAW with maybe two more stops of headroom, albeit with possibly funky bokeh near the highlight clipping point, as each sub-pixel only sees through a fraction of the entrance pupil.

There is no benefit to higher pixel densities for DR at high ISOs in FF cameras, historically, but at base ISO, DR is dictated more by readout electronics than by photosite qualities, with higher pixel counts usually giving the most DR at base ISO. It's like audio tape; the faster you record at and play back at, the more the tape hiss moves into inaudible frequencies that don't interfere as much with the audio. Having more pixels forming an image is analogous to a faster tape speed, in regard to any late-stage noise that occurs after the initial gain; instead of hearing at a rate over time, we see visible hiss at a rate across the image. For the noise that occurs before the first gain stage, more pixels do not lower that type of noise; it should be about the same at the image level.

So, if we are talking about "engineering" DR (a la DxOMark's "DR"), which does not take the photon noise into account, and if we allow for some kind of HDR possible through sub-pixel images, approaching 16 stops of "engineering" DR could be feasible, as the R5 with normal RAW has a DxO DR of 14.6, and DPRAW puts even the existing R5 at about 15.6 stops. If we are talking about something like Bill Claff's PDR, which is more demanding and takes photon noise into account, and a non-HDR single-image RAW file, then over 13 stops of PDR is not likely, due to necessary photon noise, which limits how high PDR can go with a given sensor size. In fact, I think that in one of Bill's tutorial pages, he shows the theoretical limits to PDR for a given sensor size and quantum efficiency, in the absence of any read noise at all.

The definition of "DR" is going to come into scrutiny even more in the future, when photon-counting sensors without read noise start to get measured by DxO and Bill Claff. The engineering definition becomes nonsense, as there is no denominator, and PDR becomes more mathematically useful, but may still mislead to some degree, because a fixed SNR caused by only photon noise is much nicer to look at, and far less destructive to the viewing process, than when heavily weighted by read noise, which has all kinds of non-random spatial correlation that causes banding and large chromatic noise "grain" that is harder to filter out, and disappears less with reduced image display size.

Gee, John, I wish I could speak whatever language you speak in. That, or if you could occasionally attempt to be just a bit condescending and talk down to us dummies a bit!


Definition of 'dummy': n. reference to those not of post graduate Engineering or quantum Physics background

TeamSpeed wrote in post #19203993
Based on the CR0 and their own text basically stating that they are bored, the industry is quiet, so they are just throwing out stats to get attention....

Therefore, there is nothing there but a big fat "Click Me" article.

Then there is this from the same site on another article...

Not taking the bait either.

CyberDyneSystems wrote in post #19203933
Does CR 0 mean that if I say the R1 will have a built in hand warmer, that is good enough to rate "CR 0"?

You may not have realized this, Jake, but you accidentally stumbled across the fact that the R1 will have a Peltier device embedded, which pulls heat out of the body, for continuous 8k video recording, and transfers the heat to your hands to keep them warm. Conversely, in very cold weather the Pentier device is reversed, and it draws heat from your body to heat the battery and keep it functioning even in below-zero conditions. A truly professional feature found only in the R1 and not to ever be migrated downward to lesser bodies. The extra large space of the lower body permits several batteries, to power the Peltier device. And optional cooling fin attachment to the bottom of the body is used for continuous video shooting in warm conditions, so that your hands do not get too hot.

John Sheehy wrote in post #19210816
The DR seems high, perhaps, but that could be compared to "PDR", which tends to run much lower than DxO-like DR.

Depending on the definition of DR and what file types we are talking about, it might not be all that far-fetched.

The R5 with Dual Pixel RAW has about a stop more DR than a normal RAW in the form of pure extra highlight headroom, even though software generally does not take advantage of it. libraw.org has a downloadable tool that turns DPRAWs into 2 DNG files, and you can convert and blend them as you will, and some future tool may make them into an HDR single RAW, or converters may support the extra DR at some point in the future, (if none of them do so already).

Quad-pixel RAW, if true (hard to imagine in early 2021 in a camera with 85M full pixels, but who knows what is going on in prototypes) may provide even more headroom. Canon could choose to have quad-pixel RAWs, or they could blend the 4 images in-camera to create a single-image HDR RAW with maybe two more stops of headroom, albeit with possibly funky bokeh near the highlight clipping point, as each sub-pixel only sees through a fraction of the entrance pupil.

There is no benefit to higher pixel densities for DR at high ISOs in FF cameras, historically, but at base ISO, DR is dictated more by readout electronics than by photosite qualities, with higher pixel counts usually giving the most DR at base ISO. It's like audio tape; the faster you record at and play back at, the more the tape hiss moves into inaudible frequencies that don't interfere as much with the audio. Having more pixels forming an image is analogous to a faster tape speed, in regard to any late-stage noise that occurs after the initial gain; instead of hearing at a rate over time, we see visible hiss at a rate across the image. For the noise that occurs before the first gain stage, more pixels do not lower that type of noise; it should be about the same at the image level.

John- Could you please explain the highlighted text better? That doesn't make sense to me unless one is binning pixels, combining several pixels together with a loss of resolution. I don't see the correlation of higher pixel count with resolution. Rather the opposite, using the same technology to make the sensor because the smaller photo sites can't hold the same charge as larger ones, so they are limited in how much light they can collect.

John Sheehy wrote in post #19210816
So, if we are talking about "engineering" DR (a la DxOMark's "DR"), which does not take the photon noise into account, and if we allow for some kind of HDR possible through sub-pixel images, approaching 16 stops of "engineering" DR could be feasible, as the R5 with normal RAW has a DxO DR of 14.6, and DPRAW puts even the existing R5 at about 15.6 stops. If we are talking about something like Bill Claff's PDR, which is more demanding and takes photon noise into account, and a non-HDR single-image RAW file, then over 13 stops of PDR is not likely, due to necessary photon noise, which limits how high PDR can go with a given sensor size. In fact, I think that in one of Bill's tutorial pages, he shows the theoretical limits to PDR for a given sensor size and quantum efficiency, in the absence of any read noise at all.

The definition of "DR" is going to come into scrutiny even more in the future, when photon-counting sensors without read noise start to get measured by DxO and Bill Claff. The engineering definition becomes nonsense, as there is no denominator, and PDR becomes more mathematically useful, but may still mislead to some degree, because a fixed SNR caused by only photon noise is much nicer to look at, and far less destructive to the viewing process, than when heavily weighted by read noise, which has all kinds of non-random spatial correlation that causes banding and large chromatic noise "grain" that is harder to filter out, and disappears less with reduced image display size.

Which sensors are those? I'm familiar with photomultiplier tubes and avalanche photodiodes, but these are used for very low-light situations. Shot noise dwarfs other noise under these conditions.

Quad pixel autofocus?
https://www.techradar.​com …el-af-new-patent-suggests

CyberDyneSystems wrote in post #19203933
Does CR 0 mean that if I say the R1 will have a built in hand warmer, that is good enough to rate "CR 0"?

Sounds great, but I doubt it on many levels.

- I don't think the R1 will be an MP beast at 85mp.
- I truly doubt a global shutter will premier with such a monstrous MP sensor,.
- That DR number is proof that this is just someones fantasy

I think you may be on to something here - I mean, they have to dissipate the heat SOMEHOW.

For what it's worth, I posted that before we knew what the R3 was sensor wise.
Given that the R3 was built for speed, and covers pretty much everything any self respecting 1D could ever desire, I now do believe that the R1 WILL be a megapixel monster.