Report: The next EOS R camera is reported to be undergoing field testing [CR1]

[3kramd5]

Has Sony recovered the front end investments they put into developing their full frame cameras? Maybe so, maybe not.

Full frame cameras? I don’t know, but given how quickly they spun out 9 of them, I’d imagine they have minimized development cost.

The post I quoted referred to sensors, though. I can’t imagine Sony Semiconductor Solutions would keep that product line alive if it lost money.

 

[SwissFrank]

Canon have admitted in interview that they cannot do what Sony are doing because they do not have the sensor technology

i'll agree that article does support that view, but reading between the lines, the biggest, most profitable camera company can afford to do whatever it wants. If it's not developing market-beating sensors, I believe it's because they decided they don't need to.

It seems Sony went all-in on developing the sensor for the FF cameras which seems to have paid off

Again, that backs up my point. As you say it paid off... with a minor share of the mirrorless market and a minor share of the full-frame market, even if they did own the quite tiny mirrorless full-frame combination for a few years. Without the sensors they would have failed to do even that, right? So again, that's my point: "It doesn't matter if Sony, Nikon or whoever else has to (originally I said, feels like they have to) do it, Canon doesn't. Not right now. "

 

[Quirkz]

At 70mp, I would hope for a binned 35mp mode that provided better noise (and low light?) performance. Still plenty of resolution for most scenarios. But yeah, as an owner of an EOS R, I'm really hoping for a few things in a camera like this—the return of 5D/7D ergonomics (a few more CF buttons and a joystick), dual card slots, and better rendering.

On the last point, it's hard to describe, but the R's files have a tendency to look "crunchy" and not all that pleasing right out of camera. It seems like I have to massage them a lot more in post to get good results. This is compared to my original 7D and my Fujifilm X-E3. I don't know if it's just the way the camera processes the files or if it's a sensor thing or what, but the color and contrast seem off—not enough color saturation, too much contrast. I end up trying different camera profiles in LR and can usually get good results, but the files are always a bit 'meh' out of camera. Anyone else?

This might be a preference thing. I’ve got the Fuji x-e3 as well, and the 5d4 (I’m assuming there isn’t a big difference from the R to the 5d4, given the very similar sensor, but correct me if I’m wrong). I prefer the shots out of the 5d4 myself for that contrast punch, and I find the colors more natural. Love the x-e3 for the size to performance ratio, just prefer the image from the 5D4.

 

[Kit.]

Some of the newer Sony Semiconductor sensors are performing binning on the sensor level instead of the processor.

You cannot do it with a normal Bayer pattern, or you will lose more resolution than you expected. In IMX294CJK, Sony uses "Quad Bayer" pattern, which even without binning has half of the color resolution of the normal Bayer pattern.

 

[SwissFrank]

In IMX294CJK, Sony uses "Quad Bayer" pattern

Hey, veering totally off-subject now but you clearly know your stuff. Two questions:

1) Bayer has 1 R 1 B but 2 greens, because green is the most important channel to human vision. But the biggest challenge to shooting is low-light, which is usually very warm incandescent or even candlelight, where there's very little blue. Wouldn't an alternate with twice the blue therefore be better? It'd practically cut low-light noise in half. It'd double noise in the important green channel in high-light situations but we have sensitivity to burn in well-lit situations.

2) what about an alternate pattern with two channels of IR?

3) what about an alternate pattern with three channels of IR and one white pixel? It wouldn't capture color info of course but would add a couple stops sensitivity in low-light situation?

3a) or instead of three IR, have two IR, and double both green and blue?

4) or back to four colors: what about R G B and white, again using the white to get another couple stops of luminance?

 

[Del Paso]

Most probably it is just another camera to buy Canon a bit more of a time to come with some real sensor technology changes. First, 5DIV like sensor, second, 6DII like sensor, next - high res sensor. Would be suprprised, if it would be all new generation of sensors accompained with IBIS, etc.

Did you ever post a positive comment upon Canon?

 

[Kit.]

1) Bayer has 1 R 1 B but 2 greens, because green is the most important channel to human vision. But the biggest challenge to shooting is low-light, which is usually very warm incandescent or even candlelight, where there's very little blue. Wouldn't an alternate with twice the blue therefore be better? It'd practically cut low-light noise in half. It'd double noise in the important green channel in high-light situations but we have sensitivity to burn in well-lit situations.

The green channel is important for luminance resolution. Human eye is very poor at resolving fine details in blue color.

2) what about an alternate pattern with two channels of IR?

3) what about an alternate pattern with three channels of IR and one white pixel? It wouldn't capture color info of course but would add a couple stops sensitivity in low-light situation?

3a) or instead of three IR, have two IR, and double both green and blue?

4) or back to four colors: what about R G B and white, again using the white to get another couple stops of luminance?

Ideally, we would want our pixels to saturate at the same exposure. With RGBW, the W channel saturates early and provides little to no information in "normal" light. The W channel somehow helps in case of underexposure, but at the expense of color noise.

If we don't care about color at all, a monochrome sensor is obviously the best. If we do care about color recognition, but not by humans (machine vision), CYYM or CMYW would probably be the best. The infrared channel would add more information for subject recognition in machine vision, but not all lenses are apochromatic "enough" for it to be in focus.

 

[flip314]

The green channel is important for luminance resolution. Human eye is very poor at resolving fine details in blue color.

You can actually do some cool experiments in photoshop. Take any photo at screen resolution and split it into the R G + B channels. Resize the blue channel to 50% size, then 200% size (so it's the same as the starting resolution). This basically throws away a bunch of the blue information from the picture. Then recombine the channels. You will be hard pressed to notice any loss in resolution. You can probably even get away with it at 33%-300% or 25%->400% resize.

Try the same thing with green and you'll notice right away. Red kind of falls somewhere in the middle between green and blue for importance to humans.

 

[SwissFrank]

The green channel is important for luminance resolution.

OK, so a R B B G quartet would reduce color noise, esp. blue channel color noise, but have more luminance noise. With the current Bayer pattern, at extreme high-ISO shots with low black-body-temp light sources (candlelight, incandescent on dimmer) is color noise or luminance noise the worse problem?

Ideally, we would want our pixels to saturate at the same exposure.

It seems like one of the quad-Bayer applications is to have half the pixels relatively insensitive, to capture detail in highlights. That seems like a noble cause though I like Canon's alternative of switching between two accumulators to achieve the same ends (from the journal article mentioned on this site a couple weeks ago).

With RGBW, the W channel saturates early and provides little to no information in "normal" light. The W channel somehow helps in case of underexposure, but at the expense of color noise.

So my RBBG quartet idea is bad because it'd unacceptably cut color noise at the expense of luminance noise.
So my RBGW quartet idea is bad because it'd unacceptably cut luminance noise at the expense of color noise.
I didn't realize we were so well-balanced between color and luminance noise that a step in either direction was a step down :-D

In fact, yeah, I proposed RBGW is that in very low light, luminance is -6dB of that currently (if we assume an unfiltered white is getting 3x the photons of a filtered green). Color noise is only a bit worse though (-6dB for green channel but R and B unchanged). The worst case scenario would be that color noise is unacceptable but you can desaturate a bit and just enjoy the smooth luminance.

If we don't care about color at all, a monochrome sensor is obviously the best.

I figured that out in 2000 or so. I think Leica had models simply missing the Bayer filter and some site surely has a walkthrough on the resulting improvements.

The infrared channel would add more information for subject recognition in machine vision

I hadn't thought of that. My idea was just to be able to duplicate digitally the look of infrared B&W or even infrared color film at will. Naturally you wouldn't be able to focus all colors unless you had a truly apochromatic lens (and I think most "apo" lenses are only apo in visible light).

 

[Michael Clark]

I dont like shoot blind after couple shoot i wait viewfinder lag ends and recompose pic . They need get lag away before can increase fps

That's what your left eye is for. Try it sometime. It does take getting used to, but eventually you can train your brain to use one eye at a time when both are open.

At 70mp, I would hope for a binned 35mp mode that provided better noise (and low light?) performance. Still plenty of resolution for most scenarios. But yeah, as an owner of an EOS R, I'm really hoping for a few things in a camera like this—the return of 5D/7D ergonomics (a few more CF buttons and a joystick), dual card slots, and better rendering.

On the last point, it's hard to describe, but the R's files have a tendency to look "crunchy" and not all that pleasing right out of camera. It seems like I have to massage them a lot more in post to get good results. This is compared to my original 7D and my Fujifilm X-E3. I don't know if it's just the way the camera processes the files or if it's a sensor thing or what, but the color and contrast seem off—not enough color saturation, too much contrast. I end up trying different camera profiles in LR and can usually get good results, but the files are always a bit 'meh' out of camera. Anyone else?

Horizontal or vertical line binning doesn't really do much for improving noise performance. Full binning of a 70MP sensor would result in a 17.5MP image, not 30MP.

On the last point, it's hard to describe, but the R's files have a tendency to look "crunchy" and not all that pleasing right out of camera. It seems like I have to massage them a lot more in post to get good results. This is compared to my original 7D and my Fujifilm X-E3. I don't know if it's just the way the camera processes the files or if it's a sensor thing or what, but the color and contrast seem off—not enough color saturation, too much contrast. I end up trying different camera profiles in LR and can usually get good results, but the files are always a bit 'meh' out of camera. Anyone else?

LR is only guessing at Canon's demosaicing algorithms. Whatever you see when LR first opens a raw file is just one of many possible interpretations of the raw image data. There's no such thing as "the" raw file in terms of what you can see on a screen. It's always a processed interpretation.

50 megapixels or 75 megapixels isn’t much of a difference. I expect the big change in this camera will be computing power, look for dual (or quad) digic9 to handle the processing load.

How are they going to get rid of the heat from all of that processing? Where are they going to get the energy for more than 100-200 frames per battery?

I think it is possible that the 20mm flange distance was selected to create space in the body for future iterations of the electronics. E.G. FF 4K would generate a lot of heat. A thicker body could hold a larger heat sink or other strategies for dissipating heat. It could also house larger, more powerful processors for high rate FPS of high MP sensors. IBIS also takes room. IMO a larger area to incorporate it may allow them a more ambitious and effective design.
I think they though about the future a lot with this platform and tried to think of as many possibilities to avoid what they see is an increasingly difficult environment for expensive re-vamps of ILC lines.

But all of that can be done better behind and below the sensor. A longer registration distance just necessitates more empty space between the sensor and the flange.

 

[3kramd5]

That's what your left eye is for.

Yep; far better situational awareness. I spent many years behind rifle scopes, and shoot cameras the same way.

 

[Pape]

i believe i use both eyes ,but after 2 first shot flying bird seems to wander edge of frame ,cant compose with left eye
maybe my hands just not steady enough to follow steady
Or pressing shutter makes camera wobble so only first one is good centered.

To be really amazing action camera ,mirrorless should have one processor what only job is make viewfinder roll good so never any lag. i would like that more than higher fp speed.

 

[Don Haines]

How are they going to get rid of the heat from all of that processing? Where are they going to get the energy for more than 100-200 frames per battery?

I suspect that it is going to be a 1DX2 size body. Have you seen the latest Olympus? It’s huge for a 4/3 camera! I think that a “pro” mirrorless is going to be a large camera.....

 

[-pekr-]

As usual, you miss the point...either through intentional obtuseness or inability to comprehend. You’re the one who claimed that Canon is ‘buying time to bring real sensor technology changes’. Given the common DRumbeat on these forums that Canon’s sensors have been ‘behind’ for many years, and the fact that they haven’t lost any market share, I asked you to explain why Canon needs to ‘bring real sensor technology changes’. Clearly you had no cogent response, so you went with snark instead. Sad.

What do you really want me to say here? Once I am considering to buy some product, I rarely care about company's market share. Well - kind of anywy - would not buy from a potentially bancrupt one, risking some warranty issues, etc. So we've been waiting for the 6DII. And I did not like its featureset - from sensor, to dumbing down simple stuff like phone output, present on 80D, etc. What you call a DRumbeat, we felt as an important aspect for our flash-less wedding in-church shooting. So my question is - do you, in any situation, allow a customer to just say - no? Or should we buy anything Canon just throws at us?

So the story continued for us - saved a bit more (and remember, in some countries, you charge something like 300-500 for a wedding shoot), and went with the 5DIV, as we felt it will have some longer term lifespan for us. Then the R came. Generally, I like the concept very much. But so far, even when I thought we could buy it as a secondary body, I just said - not yet. R - 5DIV sensor, we've already got. RP - 6DII sensor, we don't want. So the next step from Canon I have expected, is some higher res body. Packing more pixels onboard does not necessarily mean the whole ecosystem upgrade - DR, IBIS, DPAF speed, etc.

So once again - we have the right to sit and wait. The next product for us is the 5DIV replacement in an R body. In the meantime we have bought Canon Image Prograf 1000 for in-house up to A2 printing. Thinking to buy M5 or M6 replacement, just for an amusement. But - once again versed (not necessarily negative) about the M vs R mount situation - no upgrade path apart from EF mount.

So, yes - Canon either surprises us with new sensor tech, or what we are going to see is mostly refurbished 5DS(R) sensor, along with awful tracking speed, etc. My guess is based upon what we've got with R and RP, thinking Canon needs more time to come-up with more general advancements. And I think it is going to be available with the 5DIV replacement. If it comes sooner, well then, will wait for the lower res version anyway - we don't need/want 70mpix anyway.

 

[neuroanatomist]

What do you really want me to say here? Once I am considering to buy some product, I rarely care about company's market share. Well - kind of anywy - would not buy from a potentially bancrupt one, risking some warranty issues, etc. So we've been waiting for the 6DII. And I did not like its featureset - from sensor, to dumbing down simple stuff like phone output, present on 80D, etc. What you call a DRumbeat, we felt as an important aspect for our flash-less wedding in-church shooting. So my question is - do you, in any situation, allow a customer to just say - no? Or should we buy anything Canon just throws at us?

So the story continued for us - saved a bit more (and remember, in some countries, you charge something like 300-500 for a wedding shoot), and went with the 5DIV, as we felt it will have some longer term lifespan for us. Then the R came. Generally, I like the concept very much. But so far, even when I thought we could buy it as a secondary body, I just said - not yet. R - 5DIV sensor, we've already got. RP - 6DII sensor, we don't want. So the next step from Canon I have expected, is some higher res body. Packing more pixels onboard does not necessarily mean the whole ecosystem upgrade - DR, IBIS, DPAF speed, etc.

So once again - we have the right to sit and wait. The next product for us is the 5DIV replacement in an R body. In the meantime we have bought Canon Image Prograf 1000 for in-house up to A2 printing. Thinking to buy M5 or M6 replacement, just for an amusement. But - once again versed (not necessarily negative) about the M vs R mount situation - no upgrade path apart from EF mount.

So, yes - Canon either surprises us with new sensor tech, or what we are going to see is mostly refurbished 5DS(R) sensor, along with awful tracking speed, etc. My guess is based upon what we've got with R and RP, thinking Canon needs more time to come-up with more general advancements. And I think it is going to be available with the 5DIV replacement. If it comes sooner, well then, will wait for the lower res version anyway - we don't need/want 70mpix anyway.

You can say and whatever you want. Buy...don’t buy. Complain on a forum. Canon doesn’t care. Their maintained ~50% market share says they continue to make decisions approved by the majority of ILC buyers. (Oh, and as for bankrupt companies, Samsung certainly isn’t one...but where is the NX series now? )

 

[Don Haines]

So once again - we have the right to sit and wait. The next product for us is the 5DIV replacement in an R body. In the meantime we have bought Canon Image Prograf 1000 for in-house up to A2 printing. Thinking to buy M5 or M6 replacement, just for an amusement. But - once again versed (not necessarily negative) about the M vs R mount situation - no upgrade path apart from EF mount.

Personally, I think that sit and wait is a fine option. If what you have does the job and there is no compelling “killer feature “ that you need, why not wait? Myself, I have yet to upgrade to the next model, I tend to skip versions. I think that the R series has great potential, but will not jump until I see a significant improvement over what I have.

Everyone is speculating about the future, but nobody knows what it will be except Canon, and they are not talking.

 

[neuroanatomist]

Personally, I think that sit and wait is a fine option. If what you have does the job and there is no compelling “killer feature “ that you need, why not wait? Myself, I have yet to upgrade to the next model, I tend to skip versions. I think that the R series has great potential, but will not jump until I see a significant improvement over what I have.

Waiting is perfectly fine. I see no need to get the 1D X II, my 1D X meets my needs just fine. I did get an EOS R because it was a significant improvement in certain ways over my current gear:

• Much smaller than my 1D X
• FF sensor which my M6 lacks
• Ability to easily use an ND filter (adapter) with my TS-E 17 and 11-24 which are very challenging to front-filter

My main use for the EOS R will be travel.

 

[Don Haines]

Waiting is perfectly fine. I see no need to get the 1D X II, my 1D X meets my needs just fine. I did get an EOS R because it was a significant improvement in certain ways over my current gear:

• Much smaller than my 1D X
• FF sensor which my M6 lacks
• Ability to easily use an ND filter (adapter) with my TS-E 17 and 11-24 which are very challenging to front-filter

My main use for the EOS R will be travel.

I am thinking about the R for travel too. So far, I don't see any lenses that strike my fancy, but the 24-105 comes close... At the moment there is not enough lens selection for me so it would end up with an adaptor for everything. Of the rumoured lenses to come soon, it is the 24-240 that gets me going. If it is reasonably good and at a reasonable price, I can see jumping in. As a one lens walk-about camera that has the ability to be an awesome pair.

Of course, all is speculation now, so I shall wait and see.