RF Mount APS-C camera coming second half of 2021 [CR2]

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
Crop sensors are good for reach. That said, when I crop full frame photos for "more reach" the resolution is always better, at least on a screen. Full frame sensors allow 2X the light of cropped sensors, have better DR and ISO performance. The 7D is/was a great camera, but it is no 5Dxxx in any manner other than perhaps a few more frames/second.
When you crop a FF sensor, you throw away all of the "extra light" that was collected by the FF sensor due to its larger surface area. If the pixel pitch is the same and the technology is identical at the photosite (a/k/a pixel well) level, then there's absolutely zero difference between cropping the image from a FF sensor after the fact and using an APS-C sensor at the time of image capture.

The reason most FF sensors have better DR and ISO performance is because their photosites are larger than most APS-C sensors, thus allowing more light to be captured per photosite.

But if one takes, for example, an EOS 5Ds (with a low pass filter) and compares it to a 7D Mark II where both have the same pixel pitch and use the same generation of technology, there's no difference between the two apart from autofocus.

The wider baseline allowed by the wider mirror in FF cameras affects the performance of the AF system with PDAF using reflex mirrors. The narrower baseline required by the narrower mirror in APS-C DSLRs negatively affects AF performance compared to FF DSLRs. But when talking about mirrorless, which does not use microlenses to redirect light from the edges of the lens to AF sensors in the center of the AF focus array the way PDAF systems using reflex mirrors do, the difference between FF and APS-C in AF performance is negligible. That's why the 90D has much better AF in Live View than when using the OVF.
 
Last edited:

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
True, but the 90D is the closest thing to a 7D3 Canon has given us. I think Canon wants to bring back that 7 line asap. So many people badly wanted a 7D2 successor. It looks like it will be in RF. I really would be surprised, given the rumor of speed and performance in the first RF-C body, if it's not the R7
In terms of AF performance it will have the best available AF system physically possible, so more like the 7D Mark II in that respect. But in terms of durability and weather resistance due to the materials used for the exterior body as well as shutter durability, it will probably be closer to the 90D or R6. That's the gist of what I got several months ago from the last APS-C R body rumor before this one. Time will tell.
 
  • Like
Reactions: Fletchahh

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
a year isn’t exactly ASAP, but if Canon has only recently (last year ) awoken to the consumer desire for this camera, then a year is very fast to get it out the door. Probably overly ambitious.
There have been rumors of an APS-C EOS R body pretty much since the EOS R system was introduced in late 2018. It's just that most of the folks here could not hear it because they had already decided in their own minds that "EOS M is the ONLY APS-C mirrorless system for Canon" and "EOS R is a FF ONLY system."

What makes you think Canon just now "awoke" to consumer desire for this camera?
 

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
It isn't. The EF-M runs at a higher clock rate and a lower voltage than EF (when using native EF-M lenses.)

EF protocols - 1986
EF-M protocols - 2012
RF protocols - 2018

Yes, RF is more advanced. But I doubt that there's any reason to withdraw EF-M right now. After all, if you want a compact camera and compact lenses, EF-M is better than RF.
And cheaper, too.
 
  • Like
Reactions: stevelee

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
Speaking of size, RP is only 0.6 in wider than the M50, But is is also 0.3 in shorter than the M50. The RP is thicker than the M50 due to bigger grip. But with the lens being mounted on, the difference of thickness becomes meanlingless. In order for the APSC-R to be sucess, Canon has to come up with body and lenses that is comparable in size and price close to the M sytem. RP is already at $1000. The migration path is not as rosy as most of people think. if someone buys a APSC-R body and get the RF lenses, he will end up spending big bug and end up with big size( that he can migrate into the R later). The downward migration is easier. just put the big and expensive R lens on the APSC-R body. But that defeats the goal of having a smaller system. So it seems to me, APSC-R and R will almost becomes two sperate system with the common mount.
You are assuming an R7 or R90 will be made with the goal of being smaller and cheaper than the RP. That does not seem to be the primary motivation behind this camera. It seems to be more about giving current 7D MArk II and 90D users a mirrorless option.

The users of the 7D Mark II and 90D are far more concerned with pixel density and speed of performance than they are with size and weight, or with it being cheaper than the EOS RP.
 
  • Like
Reactions: Joules and SteveC

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
That is huge by M standard. Also that will be the field of view of 56 mm focal length on APS-C body. Pretty useless focal length.
Which is why the EOS M system and any potential APS-C EOS R body have absolutely nothing to do with one another.

As for the 56mm angle of view on FF being "pretty useless", how many millions upon millions of 135 format SLRs do you think were sold with a 50mm or 55mm as the "standard" lens?
 
  • Like
Reactions: stevelee and SteveC

Rocky

EOS R
Jul 30, 2010
956
56
Which is why the EOS M system and any potential APS-C EOS R body have absolutely nothing to do with one another.

As for the 56mm angle of view on FF being "pretty useless", how many millions upon millions of 135 format SLRs do you think were sold with a 50mm or 55mm as the "standard" lens?
Please tell me who is still making 55mm lens now.
 

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
I've tried 90D and the 32mp sensor, and I didn't like the high iso performance of it. I hope the "built around speed" saying in the RF rumor could mean something like a 24-25mp sensor, based on new technology like the R5's 45mp sensor which apparently have very fast readout (and image quality) despite the high megapixel count.
If you are comparing a higher density sensor to lower density sensors by pixel peeping at "100%", the higher density sensor will always look worse. This is because "100%" of a 32MP APS-C sensor is a higher magnification than "100%" of a lower resolution 20/24/26 MP APS-C sensor.

If you take a 32MP image and look at it at 100% on the same monitor that you look at a 20MP image at 100%, is the equivalent of looking at an 8x12 inch print from 20 MP sensor and comparing it to a 20x13 inch print from the 32 MP sensor. Of course you are going to be able to see every defect more clearly, just as you would if you magnified the 20 MP image to "160%" to give it the same enlargement ratio that you give the 32MP sensor at "100%"!

The R5 FF 45MP sensor cropped to APS-C is 17.8MP.
 

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
Well, build around speed sounds like something aimed at the 7D series segment and therefore, a high resolution would be advantageous for use with the large telephoto lenses. Keep in mind the R5 does 20 FPS electronic at 45 MP. You no longer have to chose between high res and high speed with Canon. I think one of the biggest arguments for a 7-series RF APS-C camera would be the shutter. Canon can offer a 12 FPS mechanical FF shutter at the R6 price point. With the lesser weight and travel length of an APS-C shutter, that may well match or exceed the 16 FPS shutter in the 1DX III at far lower cost.
The g-forces at startup and stopping the shutter curtains at the other side of the frame are the same for both a FF and APS-C shutter if one desires both to transit the sensor at the same speed in terms of millimeters per unit time. There are slight advantages to an APS-C shutter due to lower weight in terms of the amount of force needed to accelerate the shutter to the same speed as compared to a full frame shutter. But as Galileo demonstrated by dropping cannonballs of different weight/mass off the tower at Pisa, g-forces due to acceleration are independent of mass. The reason APS-C shutters are cheaper than FF shutters is because they can transit the sensor at a 1.6X lower speed and still have the same transit time as a FF shutter can.

One other thing to note is that the (what we assume is a) cheaper shutter in the R6 is rated at 300,000 actuations whereas the more expensive shutter in the R5 is rated at 500,000 actuations when both can run at the same number of frames per second. For most 7D Mark II type users, shutter durability would be more desirable than a slightly cheaper price. The typical use cases for an R7 would be similar to those for the 7D Mark II: sports, action, and birding. All three of those activities tend to involve high frame counts.

My 5D Mark III that was my primary FF body for five years before I got a 5D Mark IV has barely half as many shutter clicks on it as the 7D Mark II that I've owned for five years has, even though I use the FF bodies for everything I shoot and only use the 7D Mark II as the "long" body in a two camera setup for sports.
 
Last edited:

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
Even at 12fps mechanical on R5, it costs you some bit depth. 6 fps is the highest mechanical speed for 14 bit RAW shooting on the R5.



I've seen this theoretical argument many times, but I just have to say my personal experiences doesn't agree. I like the the 6400ISO photos with my 6 years old 7DII 20mp sensor much more than similar shots with the 90D's 32mp sensor (I tried it for a period). Also with the 90D's images downscaled to 20mp. I just couldn't get rid of white pixels in the deep shadows of the 90D images without turning noise reduction much higher up than I like to do. This is for RAW images processed in Adobe Camera Raw.
If you are processing raw Canon image files with Adobe, there's your problem, particularly with regard to noise reduction. Try Capture One or even Digital Photo Professional for much better conversion of Canon raw image files. Or even use third party NR products, such as Noise Ninja, after the fact of raw conversion. Adobe hasn't improved their noise reduction algorithms for processing raw files from Canon cameras for years (We won't even begin to talk about color demosaicing, which is even worse).
 

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
Please tell me who is still making 55mm lens now.
Close:

1599238461052.png


Close enough:

1599237677084.png


Not that there is any material difference between 50mm and 55mm, seeing as how lens focal lenses are never that exact. A lens marketed as a 35mm lens could very easily have an actuial FL of 33.33333mm which translates to 50mm at 1.5X. Lenses marketed as 50mm lenses can have actual FLs of 47-53mm.

Zeiss 55mm for Sony E-mount:

1599237998814.png


Zeiss Otus 55mm f/1.4 for a variety of mounts including Canon EF:

1599238104094.png


Zeiss presents this one as "The Standard Lens" on their own website.
 
Last edited:
  • Like
Reactions: Joules and SteveC

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
3,102
1,820
I've tried 90D and the 32mp sensor, and I didn't like the high iso performance of it. I hope the "built around speed" saying in the RF rumor could mean something like a 24-25mp sensor, based on new technology like the R5's 45mp sensor which apparently have very fast readout (and image quality) despite the high megapixel count.
The 12-bit limitation is only when using the electronic shutter at all frame rates.

For H+ (high speed continuous shooting plus) at 12 fps with mechanical shutter or electronic first curtain, the limitation is 13-bits.
 

BeenThere

EOS R
CR Pro
Sep 4, 2012
1,196
615
Eastern Shore
The g-forces at startup and stopping the shutter curtains at the other side of the frame are the same for both a FF and APS-C shutter if one desires both to transit the sensor at the same speed in terms of millimeters per unit time. There are slight advantages to an APS-C shutter due to lower weight in terms of the amount of force needed to accelerate the shutter to the same speed as compared to a full frame shutter. But as Galileo demonstrated by dropping cannonballs of different weight/mass off the tower at Pisa, g-forces due to acceleration are independent of mass. The reason APS-C shutters are cheaper than FF shutters is because they can transit the sensor at a 1.6X lower speed and still have the same transit time as a FF shutter can.

One other thing to note is that the (what we assume is a) cheaper shutter in the R6 is rated at 300,000 actuations whereas the more expensive shutter in the R5 is rated at 500,000 actuations when both can run at the same number of frames per second. For most 7D Mark II type users, shutter durability would be more desirable than a slightly cheaper price. The typical use cases for an R7 would be similar to those for the 7D Mark II: sports, action, and birding. All three of those activities tend to involve high frame counts.

My 5D Mark III that was my primary FF body for five years before I got a 5D Mark IV has barely half as many shutter clicks on it as the 7D Mark II that I've owned for five years has, even though I use the FF bodies for everything I shoot and only use the 7D Mark II as the "long" body in a two camera setup for sports.
Hmmm. G forces are not quite independent of mass unless the physics of:
F = M x A. (Force equals mass multiplied by acceleration) Is incorrect.
The old Galileo thing is that the rate of acceleration of any mass under the force of gravity is the same. That does not mean the force acting on the mass is the same.
 

SteveC

R5
CR Pro
Sep 3, 2019
1,622
1,479
When you crop a FF sensor, you throw away all of the "extra light" that was collected by the FF sensor due to its larger surface area. If the pixel pitch is the same and the technology is identical at the photosite (a/k/a pixel well) level, then there's absolutely zero difference between cropping the image from a FF sensor after the fact and using an APS-C sensor at the time of image capture.

The reason most FF sensors have better DR and ISO performance is because their photosites are larger than most APS-C sensors, thus allowing more light to be captured per photosite.

But if one takes, for example, an EOS 5Ds (with a low pass filter) and compares it to a 7D Mark II where both have the same pixel pitch and use the same generation of technology, there's no difference between the two apart from autofocus.

The wider baseline allowed by the wider mirror in FF cameras affects the performance of the AF system with PDAF using reflex mirrors. The narrower baseline required by the narrower mirror in APS-C DSLRs negatively affects AF performance compared to FF DSLRs. But when talking about mirrorless, which does not use microlenses to redirect light from the edges of the lens to AF sensors in the center of the AF focus array the way PDAF systems using reflex mirrors do, the difference between FF and APS-C in AF performance is negligible. That's why the 90D has much better AF in Live View than when using the OVF.
Amplifying (I believe) on your point here: I've seen people say something like "FF allows more light" which has always bothered me. It lets in more total light, sure, but it's spread out more, too, so the amount of light hitting a particular part of the sensor (inside the would-be cropped area) is the same. That light that hits near the outside of the FF image circle doesn't do a damned thing for the pixels near the center of the sensor. A pixel is not affected at all how many other pixels are around; it is affected by its own size, for the same reason astronomers like wider and wider mirror (or objective lens) telescopes.
 

Rocky

EOS R
Jul 30, 2010
956
56
Zeiss presents this one as "The Standard Lens" on their own website.
Thanks for introduce me to these lenses. The Sigma actually is a an APS-C lens not for FF. The Zeiss Sonnar is deigned for Sony E mount( original for APS-C). The Nikon is a 0.95 with a $8000 price tag. The Zeiss Otus is 1.4 with $4000 price tag. All the later three are extra ordinary design to obtain superior images quality, worth the brand of Zeiss. Due to smaller ideal throat opening( Sony) or clearance of the mirror box ( Sonnar and Otus) with fast aperture, all four have to use reverse telephoto design and bump the focal length to 55mm. Remember the Zeiss Biotar for Exakta and Practica?? That is also 55mm. Due to clearance of the mirror box. It is the same situation.
Just out of curiosity, what is the percentage of your photo are taken with 55mm lens?
 
Last edited:

Joules

EOS R
CR Pro
Jul 16, 2017
1,046
1,183
Hamburg, Germany
The g-forces at startup and stopping the shutter curtains at the other side of the frame are the same for both a FF and APS-C shutter if one desires both to transit the sensor at the same speed in terms of millimeters per unit time. There are slight advantages to an APS-C shutter due to lower weight in terms of the amount of force needed to accelerate the shutter to the same speed as compared to a full frame shutter. But as Galileo demonstrated by dropping cannonballs of different weight/mass off the tower at Pisa, g-forces due to acceleration are independent of mass. The reason APS-C shutters are cheaper than FF shutters is because they can transit the sensor at a 1.6X lower speed and still have the same transit time as a FF shutter can.

One other thing to note is that the (what we assume is a) cheaper shutter in the R6 is rated at 300,000 actuations whereas the more expensive shutter in the R5 is rated at 500,000 actuations when both can run at the same number of frames per second. For most 7D Mark II type users, shutter durability would be more desirable than a slightly cheaper price. The typical use cases for an R7 would be similar to those for the 7D Mark II: sports, action, and birding. All three of those activities tend to involve high frame counts.

My 5D Mark III that was my primary FF body for five years before I got a 5D Mark IV has barely half as many shutter clicks on it as the 7D Mark II that I've owned for five years has, even though I use the FF bodies for everything I shoot and only use the 7D Mark II as the "long" body in a two camera setup for sports.
I am not sure if I understand you here correctly.

Looking at a single blade on the shutter curtain, its width will be 1.6 times greater for a FF shutter compared to APS-C. If the other two dimensions remain the same, the mass will be 1.6 as great too.

To accelerate or decelerate the two shutter blades with their different masses to the same velocity v, I can either have the time it takes to reach v the same, requiring me to apply a greater force to the FF shutter, or have the applied force the same, requiring a greater amount of time for the FF shutter to reach this speed.

v = a*s where a=F*m

You even mention this. Why does it matter that the acceleration my be the same? Doesn't one take more powerful motors and more energy?

And for the two shutters traveling at the same speed, an APS-C one will complete a full transition of the sensor in 1/1.6 the amount of time. So if that were the only constraint in FPS, it could achieve a framerates that is 1.6 times as high.

I am not disagreeing with you here. Just not quite seeing what you wanted to point out. Maybe I got something wrong?

All I was going for with my point was this:

A mechanical shutter still has an advantage over electronic ones. An APS-C sensor imposes lesser challenges when driving the shutter to high fps. The M6 II already beats the R6 and R5 on paper with 14 vs 12 FPS, although at a lesser shutter rating (100,000). Leading me to believe that if they were to design an APS-C camera with RF design constraints, rather than EF-M Iines, they may well match or exceed the 16 FPS speed found in the 1DX III while maintaining a price that is not too far off what the 7D III crowd wants.

Achieving this same speed in the truly high res R (80 MP+) is unlikely to be achieved at this price point. Just in terms of electronic speed, it could match an R7 and outperform it in all other regards. But I think there may well be a reason to introduce an APS-C camera to the RF system because that specific niche that wants reach and speed is still best served by such a camera, unless the corresponding market is also willing to deal with pricing at or above the R5 level.
 

Joules

EOS R
CR Pro
Jul 16, 2017
1,046
1,183
Hamburg, Germany
Even at 12fps mechanical on R5, it costs you some bit depth. 6 fps is the highest mechanical speed for 14 bit RAW shooting on the R5.



I've seen this theoretical argument many times, but I just have to say my personal experiences doesn't agree. I like the the 6400ISO photos with my 6 years old 7DII 20mp sensor much more than similar shots with the 90D's 32mp sensor (I tried it for a period). Also with the 90D's images downscaled to 20mp. I just couldn't get rid of white pixels in the deep shadows of the 90D images without turning noise reduction much higher up than I like to do. This is for RAW images processed in Adobe Camera Raw.
But you usually don't shoot high FPS scenes at ISO 100. So you're loosing DR anyway, so that slightly reduced Bitrate should not translate to a real world disadvantage, right?

As for the low light performance of higher resolutions, that is nothing purely theoretical. I linked you to a nice demonstration where you can compare same sized sensors with different resolutions at the same magnification to see that the M6 II beats the 80D and the R5 and R6 are so similar, that I find myself flip flopping between which one looks better when starting at them for a while.

Unless you also set up a studio comparison with identical lighting conditions and settings to make your personal observations, any differences you see are hard to blame on the resolution differences. Of course, you have to compare at the same magnification, otherwise you make a comparison where two variables are changed. But you seem aware of that.

But you also say you like the 7D II results more than those from similar 90D shots. As light intensity is the greatest influence on noise, and it is not a linear one, you can't make proper comparisons with similar conditions. They have to be identical. As for white dots, that sounds odd. That's a type of noise I would expect for long exposures, or really hot conditions. As Michael pointed out, it may also be a technical difference.

Anyway, as I see it, an R7 only makes sense if it can make the most out of long lenses. So it must have a sufficient pixel density to exceed the R5 and compete with the eventual R5's / R3 high MP FF body.