Is a 150mp Canon EOS R camera on the way? [CR1]

[scyrene]

Canon have been pushing the high MP envelope with prototype sensors for a long time (though nothing much recently). However, my personal opinion is a ~83MP sensor with the same pixel density as the 90D is still the most likely for a 5Ds(R) successor.

 

[tron]

Canon have been pushing the high MP envelope with prototype sensors for a long time (though nothing much recently). However, my personal opinion is a ~83MP sensor with the same pixel density as the 90D is still the most likely for a 5Ds(R) successor.

And since a mirrored implementation of this pixel density exists (90D) one can only hope for a 5DsRII
OK OK this is just a personal wish...

 

[Sharlin]

Such a sensor would dramatically reduce the occurrence of false color and other nasty artifacts arising from deBayering.

That's a quite important point. One could simply think of a 150MP-sensor camera as a ~38MP-sensor camera that actually has full-resolution rather than half- or quarter-res chroma channels (although the red and blue channels would still have a stop less DR than the green channel).

 

[Quarkcharmed]

to shoot e.g. landscapes the classic way (f/16 or 22)

On full frame, f16-22 isn't a classic way though, f8-11 would be a more typical range, depending on the focal length.

 

[Quarkcharmed]

Put another way, a 50mp sensor at f/16 is not as sharp or detailed as the same sensor at f/5.6. But it's still sharper than a 22mp sensor at f/16. Both cameras are "limited" by the f/16 aperture, but the higher resolution sensor still pushes the final result closer to the theoretical best it can be for f/16.

It probably does, but I wonder if there will be a practical difference between 150Mp and say 80Mp. I suspect it'll win so little at a very high cost. For example, the pixel area will be >4x smaller than on a 30Mp sensor and such a sensor will have issues with well capacity and dynamic range.

 

[Treyarnon]

On full frame, f16-22 isn't a classic way though, f8-11 would be a more typical range, depending on the focal length.

Back in the film days, we used to use F16-22. Its only since digital came along that we opened up to the F8-11 range.
Ansel Adams used to shoot at F64 didn't he?

 

[Mr Majestyk]

More likely pixel shift mode for R5.

 

[deleteme]

Make it a 3 layer 'foveron style' sensor, with 50MP on each layer, and take my money Canon!
(for the record, I have about the same amount of money as the odds of this coming true - 0 )

A billion to one? I will be your friend.

 

[Quarkcharmed]

Back in the film days, we used to use F16-22. Its only since digital came along that we opened up to the F8-11 range.
Ansel Adams used to shoot at F64 didn't he?

Ansel Adams used to shoot large format. 8"x10", 4"x5". Are you aware how that affects the depth of filed compared to a full frame camera?
It's not about digital vs. film at all, it's about depth of field and circle of confusion vs diffraction limit.

 

[deleteme]

More likely pixel shift mode for R5.

I have been waiting for this info as everyone else with IBIS seems to add the feature.

 

[TracerHD]

I recommend a 250MP sensor, with "pixel/sensor shift" we could get 1GigaPixel with only 4 shots ....

 

[Sator]

Yes: More Ultra High-Resolution MTF Experiments

Thank goodness someone posted this link to Roger Cicala's post.

It strongly questions the dogma that "sensors are out-resolving lenses". The resolution limits of each format remain uncertain at this stage. It is possible to determine what the reasonable resolution limits of each format are, but that would be a very costly experiment to run.

As for the business of diffraction degrading optics down to a certain f stop these figures get quoted more like dogma rather than science. Never are the mathematical calculations (along with the justification of implicit assumptions used in the calculations) shown so that others can double-check how the f stop limit of usability was derived. Instead, all you get is an f stop number quoted at you like the number "42" given as the meaning of life in The Hitchhiker's Guide to the Galaxy. Even if the theoretical calculations were to be shown (so that they can be independently verified), they are still hypothetical and speculative in nature. You still need to run the empirical experiment proving that the degradation from diffraction is significant.

I also agree with Roger Cicala about the DxO Mark stuff about perceptual megapixels being nonsense ("those of you who believe in perceptual megapixels or that the earth is flat" Roger and out). DxO Mark has never published an externally verifiable methodology for how they measure this, once again making the numbers they post no better than the ex machina assertion that the meaning of life is "42".

 

[Quarkcharmed]

You still need to run the empirical experiment proving that the degradation from diffraction is significant.

With my EF 16-35 f4L, say at 16mm, f8 is sharp, f11 is acceptable in most cases (although I prefer f9 or f10 when possible), f16+ clearly shows loss of sharpness compared to the same scene shot at f11.
Sometimes I'm forced to use f16+ for misc. reasons, but prefer to use f11 and wider. Also I prefer to focus stack rather than increase the f-number, if possible.

I also agree with Roger Cicala about the DxO Mark stuff about perceptual megapixels being nonsense ("those of you who believe in perceptual megapixels or that the earth is flat" Roger and out). DxO Mark has never published an externally verifiable methodology for how they measure this, once again making the numbers they post no better than the ex machina assertion that the meaning of life is "42".

In terms of resolution/sharpness measurements, DxO is somewhat useful if you want to figure the 'sweet aperture spot' of your lens. But I agree that in general the perceptual megapixels thing is an obscure metric and scientifically speaking, a random number because the calculations aren't published, aren't peer-reviewed and aren't falsifiable.

 

[GoldWing]

Not such a far fetched rumor at 100mb.

Fuji, Hasselblad seem to enjoy the market why not Canon.

 

[kten]

Not such a far fetched rumor at 100mb.

Fuji, Hasselblad seem to enjoy the market why not Canon.

they are medium format though which are completely different market

 

[privatebydesign]

So did anybody here ever see a decent/sharp/detailed image from an iPhone? Good enough to print to 8” x 10”? I think that is a pretty reasonable print size and I know from that 12mp iPhone camera you can get very detailed high quality prints.

Now ignore everything you think you know about DLA, pixel density, airy discs and all the other garbage people talk, scale that iPhone sensor up to full frame and you have a 420mp sensor @ f11 and the same relative size print enlargement is 50" x44".

Yes print your iPhone image out to 8"x10" and you are holding a crop of a 420mp f11 ff sensor print to 50"x44".

Those numbers should end this regularly repeated thread of dismay, angst, disinformation and ignorance (but it won’t). Now I doubt there are many people that need to print or crop to those extremes but to my mind 150mp doesn’t seem any more extreme than 50mp sounded a few years ago, mind you Canon we’re playing with the idea of extreme pixel density and ways of reconfiguring detail way beyond the airy disc limitations many years ago,.

 

[Quarkcharmed]

Yes print your iPhone image out to 8"x10" and you are holding a crop of a 420mp f11 ff sensor print to 50"x44".

a phone sensor isn't that easy to scale up (or should I say, enlarge with the same pixel density). I suspect electronics will be getting more and more tricky to do, also the production cost will grow exponentially. But most important thing is, even in case we enlarge them, the phone sensors are only 10 or 12 bit ones. Tiny photosites just don't have enough well capacity. That limits the dynamic range and overall IQ.

Have you ever worked with raw files from phones, before they're processed into jpegs? I have, and they're tear-shedding.

 

[privatebydesign]

a phone sensor isn't that easy to scale up (or should I say, enlarge with the same pixel density). I suspect electronics will be getting more and more tricky to do, also the production cost will grow exponentially. But most important thing is, even in case we enlarge them, the phone sensors are only 10 or 12 bit ones. Tiny photosites just don't have enough well capacity. That limits the dynamic range and overall IQ.

Have you ever worked with raw files from phones, before they're processed into jpegs? I have, and they're tear-shedding.

And none of that was my point. My point was we can already see what shutter speeds we need, what apertures work, what even much greater pixel densities actually look like, that even moulded plastic lenses are plenty good enough to resolve much higher pixel densities than 150 mp on a FF sensor etc etc. We have that information in the palm of our hands (literally), we don't need people telling us it can't be done when it can, it is, and the IQ boundaries are very much further away than even the latest rumor.

So people stop wringing your hands about a rumored 150mp 135 sized sensor it will happen (Canon have been showing off their fully working 120 mp APS-C sensor for so long they don't do it anymore and they had a 250 mp APS-H sensor announcement several years ago), the sky won't fall, all the images won't be mush, the shutter speed doesn't need to be 1/16,000 sec, apertures we now use will work fine, the lenses you own won't be 'out resolved', DLA is nothing more meaningful than the Defense Logistics Agency and is yet another made up bullshit term for you all to obsess over.

 

[Sporgon]

Back in the film days, we used to use F16-22. Its only since digital came along that we opened up to the F8-11 range.
Ansel Adams used to shoot at F64 didn't he?

Yes but that was on large format ! Assuming 10x8 as large format then to achieve a FF equivalence in dof, f/64 is equivalent to only around f/5.6. To achieve a decent dof through the image you have to use tilt and change the plane of focus.

 

[privatebydesign]

To be accurate 8" x 10" has a 0.143 crop factor, so f64 on 8" x10" is equivalent to f9 on a ff camera. It also means a 12", or 300mm, lens is a slightly wide standard lens giving the fov of a 43mm lens on a FF camera.