fredmiranda.com Login
This topic is in the 2020 archive. You can read it and like its posts here. It is read only.

Archive 2020New Canon R5 Rumor

  
 
rbeasley
Offline
Details
p.6 #1New Canon R5 Rumor
arbitrage wrote:
RoamingScott wrote
guidostow wrote
RoamingScott wroteI'm definitely in the camp of, if you're going to try to go over 60mp, do so in a 44x33 format and not full frame. The R4 is showing us that even at 62mp, full frame is starting to burst at the seams.
I don't follow the sony boards enough to understand what you mean by "burst at the seams"
Could you elaborate?
The pixels are getting so small that high iso looks considerably worse than the R3 sensor, more noise in general but more worryingly, a lot more false color in that noise.

Where the R3 has grain at high ISO, the R4 has lots of green and magenta within the noise. It's very noticeable in controlled tests.

Most of the Sony apologists on that board say "well that's at 100% crop, it looks fine when viewed full size!", which...is true, until you do a deep crop on a high ISO image. That's the point of all this resolution in the first place right? We want to carry shorter lenses and crop instead, very very few of us are printing building sized advertisements.
Yes, my 2 weeks with the R4 showed me it was very difficult to realize the benefits of all those pixels. There were a lot of color artifacts (blues and purples) in dark feathers and darker backgrounds showed more color noise and blotches (green, magenta). Similar color artifacts in feathers that I had found with my 5DSR back in the day. And yes for birds, wildlife (maybe sports) we are really buying high MP cameras to crop and use shorter lenses or even just to crop with our longest lenses when birds are tiny and skittish. But time and time
...Show more →

Yes but it is the camera of the decade is it not?

Jan 30, 2020 at 04:41 PM
arbitrage
Offline
Details
p.6 #2New Canon R5 Rumor
rbeasley wrote:
arbitrage wrote
RoamingScott wrote
guidostow wrote
RoamingScott wroteI'm definitely in the camp of, if you're going to try to go over 60mp, do so in a 44x33 format and not full frame. The R4 is showing us that even at 62mp, full frame is starting to burst at the seams.
I don't follow the sony boards enough to understand what you mean by "burst at the seams"
Could you elaborate?
The pixels are getting so small that high iso looks considerably worse than the R3 sensor, more noise in general but more worryingly, a lot more false color in that noise.

Where the R3 has grain at high ISO, the R4 has lots of green and magenta within the noise. It's very noticeable in controlled tests.

Most of the Sony apologists on that board say "well that's at 100% crop, it looks fine when viewed full size!", which...is true, until you do a deep crop on a high ISO image. That's the point of all this resolution in the first place right? We want to carry shorter lenses and crop instead, very very few of us are printing building sized advertisements.
Yes, my 2 weeks with the R4 showed me it was very difficult to realize the benefits of all those pixels. There were a lot of color artifacts (blues and purples) in dark feathers and darker backgrounds showed more color noise and blotches (green, magenta). Similar color artifacts in feathers that I had found with my 5DSR back in the day. And yes for birds, wildlife (maybe sports) we are really buying high MP cameras to crop and use shorter lenses or even just to crop with our longest lenses when birds are tiny and skittish. But time and time again, I keep reaching for that holy grail and always end up disappointed when I get to those really high pixel densities and try to use them to crop big. I'm sure for printing large landscapes shot at base ISO they would be phenomenal. But not for my use case.

As far as higher MPs cameras go, I've been most happy with the D850 at 46MP so that is why I'd choose 45 over 70+. I also really liked the 5D4 at 30MPs. So I think 35-45 would be great.
Yes but it is the camera of the decade is it not?



The original A7/R won the DPReview poll for camera of the decade. Followed by D850 and iPhone4. I think A7/R was honored because it started an entire new segment of FF MILCs that has grown ever since.
Jan 30, 2020 at 04:53 PM
ChrisMak
Offline
Details
p.6 #3New Canon R5 Rumor
arbitrage wrote:
RoamingScott wrote
guidostow wrote
RoamingScott wroteI'm definitely in the camp of, if you're going to try to go over 60mp, do so in a 44x33 format and not full frame. The R4 is showing us that even at 62mp, full frame is starting to burst at the seams.
I don't follow the sony boards enough to understand what you mean by "burst at the seams"
Could you elaborate?
The pixels are getting so small that high iso looks considerably worse than the R3 sensor, more noise in general but more worryingly, a lot more false color in that noise.

Where the R3 has grain at high ISO, the R4 has lots of green and magenta within the noise. It's very noticeable in controlled tests.

Most of the Sony apologists on that board say "well that's at 100% crop, it looks fine when viewed full size!", which...is true, until you do a deep crop on a high ISO image. That's the point of all this resolution in the first place right? We want to carry shorter lenses and crop instead, very very few of us are printing building sized advertisements.
Yes, my 2 weeks with the R4 showed me it was very difficult to realize the benefits of all those pixels. There were a lot of color artifacts (blues and purples) in dark feathers and darker backgrounds showed more color noise and blotches (green, magenta). Similar color artifacts in feathers that I had found with my 5DSR back in the day. And yes for birds, wildlife (maybe sports) we are really buying high MP cameras to crop and use shorter lenses or even just to crop with our longest lenses when birds are tiny and skittish. But time and time
...Show more →

Could not agree more. 45mp FF and 20-22 mp 1.5x crop are the sweet spot i.m.o. The Nikon D500/Z50 and the Nikon D850/Z7 both hit the sweet spot. That is why the coming Canon R5 sounds like the first Canon body that truly interests me, although a crop version would even be better.
Jan 30, 2020 at 05:11 PM
alundeb
Offline
Details
p.6 #4New Canon R5 Rumor
ChrisMak wrote:
arbitrage wrote
RoamingScott wrote
guidostow wrote
RoamingScott wroteI'm definitely in the camp of, if you're going to try to go over 60mp, do so in a 44x33 format and not full frame. The R4 is showing us that even at 62mp, full frame is starting to burst at the seams.
I don't follow the sony boards enough to understand what you mean by "burst at the seams"
Could you elaborate?
The pixels are getting so small that high iso looks considerably worse than the R3 sensor, more noise in general but more worryingly, a lot more false color in that noise.

Where the R3 has grain at high ISO, the R4 has lots of green and magenta within the noise. It's very noticeable in controlled tests.

Most of the Sony apologists on that board say "well that's at 100% crop, it looks fine when viewed full size!", which...is true, until you do a deep crop on a high ISO image. That's the point of all this resolution in the first place right? We want to carry shorter lenses and crop instead, very very few of us are printing building sized advertisements.
Yes, my 2 weeks with the R4 showed me it was very difficult to realize the benefits of all those pixels. There were a lot of color artifacts (blues and purples) in dark feathers and darker backgrounds showed more color noise and blotches (green, magenta). Similar color artifacts in feathers that I had found with my 5DSR back in the day. And yes for birds, wildlife (maybe sports) we are really buying high MP cameras to crop and use shorter lenses or even just to crop with our longest lenses when birds are tiny and skittish. But time and time again, I keep reaching for that holy grail and always end up disappointed when I get to those really high pixel densities and try to use them to crop big. I'm sure for printing large landscapes shot at base ISO they would be phenomenal. But not for my use case.

As far as higher MPs cameras go, I've been most happy with the D850 at 46MP so that is why I'd choose 45 over 70+. I also really liked the 5D4 at 30MPs. So I think 35-45 would be great.
Could not agree more. 45mp FF and 20-22 mp 1.5x crop are the sweet spot i.m.o. The Nikon D500/Z50 and the Nikon D850/Z7 both hit the sweet spot. That is why the coming Canon R5 sounds like the first Canon body that truly interests me, although a crop version would even be better.


This scenario is no surprise, the effective resolution on the sensor is noise limited and not pixel limited. Translated to the shooting conditions, you are light limited and focal length limited at the same time, a typical situation. When you have already maximized you other parameters, it is not possible to eek out more the easy way. The same goes for adding another teleconverter, the problems are very similar.

In addition to the obvious scenarios where we are not light limited, like some landscape photography, there are also situations when I find it possible to realize the benfits of a deep crop, and that is when I use shorter and faster lenses, typically f/2.8 lenses without teleconverters. In those situations it would also have worked to add a teleconverter, but I want the instant flexibility of not doing so. Simplified, making a 300 mm f/2.8 prime an instant 300 - 600 mm f/2.8 - f/5.6 effective digital zoom.

Do you find it difficult to utilize excess resolution also in those situations?
Jan 31, 2020 at 01:44 AM
technic
Offline
Details
p.6 #5New Canon R5 Rumor
arbitrage wrote:
mdvaden wrote
arbitrage wroteIt is a lot easier to stabilize a smaller sensor and Olympus has been regarded as the top company in achieving combined IBIS/OSS and I think has had it longer than anyone. From my research they are maxing out at 6.5 stops combined. If Canon has pulled off 7-8 on a FF sensor then that is surely impressive but I'll believe it when I see it..
There's minimal to see, but plenty to review.

Canon notes up to 5 stops on the Canon RF 24-105mm lens. If an IBIS can't eek out 2 more stops, what value would the IBIS be in any camera. Olympus for comparison is irrelevant to what Canon has or can do. Canon is the global juggernaut for cameras and lenses combined. I'm sure they can accomplish plenty of potential in future bodies too.

I'm going to guess the maximum for specs may relate to RF lenses on RF bodies.
The main value of IBIS is to use non-stabilized lenses and have stabilization. Combined numbers have never been pure addition of the specs taken on their own. Only 3 out of 6 RF lenses have IS currently.


I remember reading that the Olympus system offering combined 6.5 stops IS is limited by effects due to earth rotation. I haven't checked the details but if true it is highly unlikely that an FF system from Canon would perform better than those 6.5 stops.

Also, IS specs often only apply in specific (limited) conditions. My 55-250 IS STM has very good IS at the short end (3-4 stops?), but at the long end and for closeups the IS is 1-2 stops at best, close to useless. My 100-400II has far better IS than the 55-250 STM even at 400mm setting, but it is also 4x heavier and many times bigger. In practice, IBIS might sometimes be able to compensate for the weak parts of lens-based IS and provide cover for lenses where lens-based IS is close to impossible (like very bright standard/tele primes).
Jan 31, 2020 at 03:13 AM
greenfield
Offline
Details
Jan 31, 2020 at 05:04 AM
arbitrage
Offline
Details
p.6 #7New Canon R5 Rumor
But I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...
Jan 31, 2020 at 05:26 AM
technic
Offline
Details
p.6 #8New Canon R5 Rumor
arbitrage wrote:
But I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...


Yes, but the Canon R has much bigger lenses and sensors than m43 gear, so getting 7-7.5 stops would be really pushing it. I think in practice it will be more like those 20 fps framerates in recent ML cameras and come with quite a few strings attached.
Jan 31, 2020 at 06:15 AM
arbitrage
Offline
Details
p.6 #9New Canon R5 Rumor
technic wrote:
arbitrage wroteBut I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...
Yes, but the Canon R has much bigger lenses and sensors than m43 gear, so getting 7-7.5 stops would be really pushing it. I think in practice it will be more like those 20 fps framerates in recent ML cameras and come with quite a few strings attached.


Yes, that is why I said from the beginning that the 7-8 stops was a little unbelievable for a FF sensor and a company that has never done this before....but who knows, we shall see?
Jan 31, 2020 at 06:17 AM
alundeb
Offline
Details
p.6 #10New Canon R5 Rumor
The original EOS R system white paper describes the IS architecture. Unilke EF lenses, RF lenses transmit gyro information to the camera. The camera then combines the gyro information with image sensor information (live movement detection) and sends the actuator commands to the lens.

This is unlike EF lenses image stabilization which is autonomous in the lens. I think the architecture allows for a very efficient combination of lens and body stabilization, and that they might acheive very high performance. The highest performance will likely only apply to RF lenses and not EF lenses.
Jan 31, 2020 at 06:29 AM
gdanmitchell
Offline
Details
p.6 #11New Canon R5 Rumor
"Limited" to 6.5 stops. Heh.

Let's see. With the old "rule of thumb" of 1/(focal length) and a 100mm lens:
1/100 w/o IS
1 stop: 1/50
2 stops: 1/25
3 stops: 1/12
4 stops: 1/6
5 stops: 1/3
6 stops; 2/3
6.5 stops; a bit more than 1 second.

Twice as long with a 50mm lens.

Close to a quarter second with a 400mm lens?

Handheld? Theoretically as stable as 1/100 without IS?

That's quite a bit, no?

Regarding the effect of the earth's rotation, I don't know anything about that... but it seems like it would be a constant value (at least while the camera position is constant) for which a compensatory value could be applied.
Jan 31, 2020 at 09:33 AM
highdesertmesa
Offline
Details
p.6 #12New Canon R5 Rumor
gdanmitchell wrote:
...Regarding the effect of the earth's rotation, I don't know anything about that... but it seems like it would be a constant value (at least while the camera position is constant) for which a compensatory value could be applied.


Compensation might have vary depending on altitude, which would require built-in GPS.
Feb 01, 2020 at 01:06 PM
lighthound
Offline
Details
p.6 #13New Canon R5 Rumor
Are you people serious about the earth's rotation having an effect on me hand holding a shot at 1 sec of a boulder setting in my yard?
I say boulder because at that low of a ss, anything alive including a tree or bush or snail would likely exhibit some amount of subject motion blur. Even if mounted on a tripod.

12 days to go!
Feb 01, 2020 at 04:16 PM
dclark
Offline
Details
p.6 #14New Canon R5 Rumor
technic wrote:
arbitrage wrote
mdvaden wrote
arbitrage wroteIt is a lot easier to stabilize a smaller sensor and Olympus has been regarded as the top company in achieving combined IBIS/OSS and I think has had it longer than anyone. From my research they are maxing out at 6.5 stops combined. If Canon has pulled off 7-8 on a FF sensor then that is surely impressive but I'll believe it when I see it..
There's minimal to see, but plenty to review.

Canon notes up to 5 stops on the Canon RF 24-105mm lens. If an IBIS can't eek out 2 more stops, what value would the IBIS be in any camera. Olympus for comparison is irrelevant to what Canon has or can do. Canon is the global juggernaut for cameras and lenses combined. I'm sure they can accomplish plenty of potential in future bodies too.

I'm going to guess the maximum for specs may relate to RF lenses on RF bodies.
The main value of IBIS is to use non-stabilized lenses and have stabilization. Combined numbers have never been pure addition of the specs taken on their own. Only 3 out of 6 RF lenses have IS currently.
I remember reading that the Olympus system offering combined 6.5 stops IS is limited by effects due to earth rotation. I haven't checked the details but if true it is highly unlikely that an FF system from Canon would perform better than those 6.5 stops.

Also, IS specs often only apply in specific (limited) conditions. My 55-250 IS STM has very good IS at the short end (3-4 stops?), but at the long end and for closeups the IS is 1-2 stops at best, close to useless. My 100-400II has far better IS than the 55-250 STM even at 400mm
...Show more →

---------------------------------------------

greenfield wrote:
something about the gyro sensors ... https://petapixel.com/2016/09/26/olympus-says-earths-rotation-limits-image-stabilization-6-5-stops-max/


---------------------------------------------

arbitrage wrote:
But I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...


---------------------------------------------

technic wrote:
arbitrage wroteBut I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...
Yes, but the Canon R has much bigger lenses and sensors than m43 gear, so getting 7-7.5 stops would be really pushing it. I think in practice it will be more like those 20 fps framerates in recent ML cameras and come with quite a few strings attached.


---------------------------------------------

arbitrage wrote:
technic wrote
arbitrage wroteBut I think the latest is that the Olympus Mk III coming next week gets to 7.5 stops combined....

EM1X claims 7 stops already...
Yes, but the Canon R has much bigger lenses and sensors than m43 gear, so getting 7-7.5 stops would be really pushing it. I think in practice it will be more like those 20 fps framerates in recent ML cameras and come with quite a few strings attached.
Yes, that is why I said from the beginning that the 7-8 stops was a little unbelievable for a FF sensor and a company that has never done this before....but who knows, we shall see?


---------------------------------------------

alundeb wrote:
The original EOS R system white paper describes the IS architecture. Unilke EF lenses, RF lenses transmit gyro information to the camera. The camera then combines the gyro information with image sensor information (live movement detection) and sends the actuator commands to the lens.

This is unlike EF lenses image stabilization which is autonomous in the lens. I think the architecture allows for a very efficient combination of lens and body stabilization, and that they might acheive very high performance. The highest performance will likely only apply to RF lenses and not EF lenses.


---------------------------------------------

gdanmitchell wrote:
"Limited" to 6.5 stops. Heh.

Let's see. With the old "rule of thumb" of 1/(focal length) and a 100mm lens:
1/100 w/o IS
1 stop: 1/50
2 stops: 1/25
3 stops: 1/12
4 stops: 1/6
5 stops: 1/3
6 stops; 2/3
6.5 stops; a bit more than 1 second.

Twice as long with a 50mm lens.

Close to a quarter second with a 400mm lens?

Handheld? Theoretically as stable as 1/100 without IS?

That's quite a bit, no?

Regarding the effect of the earth's rotation, I don't know anything about that... but it seems like it would be a constant value (at least while the camera position is constant) for which a compensatory
...Show more →

---------------------------------------------

highdesertmesa wrote:
gdanmitchell wrote...Regarding the effect of the earth's rotation, I don't know anything about that... but it seems like it would be a constant value (at least while the camera position is constant) for which a compensatory value could be applied.
Compensation might have vary depending on altitude, which would require built-in GPS.


---------------------------------------------

lighthound wrote:
Are you people serious about the earth's rotation having an effect on me hand holding a shot at 1 sec of a boulder setting in my yard?
I say boulder because at that low of a ss, anything alive including a tree or bush or snail would likely exhibit some amount of subject motion blur. Even if mounted on a tripod.

12 days to go!


This thread was interesting and got me to thinking about this problem.
The limits of a gyro stabilized system is not difficult to calculate.
A gyro stabilized system will keep the camera pointed the same direction in an inertial reference frame; that is a frame that is not rotating.
Standing on the Earth our reference frame is rotating at 360 deg per 24 hours which is about 0.004 deg/sec.
If we have a camera that is stabilized in a non-rotating reference frame, the image will sweep through the FOV at that rate.
Consider a 85mm lens on a 24MP FF sensor. Each pixel subtends about 0.004 deg.
So if you take a photo with a shutter speed of 1 sec, the image is smeared about one pixel. Wider angle lenses, less smear, longer lenses more smear. Higher pixel density sensor more smear measured in pixels.
The direction the camera is pointing will determine the direction of the smear. At the equator, pointing horizontal toward the east, the smear is upward. At the north pole pointed horizontal in any direction, the smear is to the right. You can figure out other directions. It does not depend on altitude.
7 stops of stabilization may be interpreted to mean you can take photos with a gyro stabilized system with a shutter as slow as 1 sec and see the same results as a photo taken at 1/128 sec with an unstabilized system (see GDan's comment above).
What about taking a picture at 10 sec to see the results that would have seen at 1/13 sec? I don't think so. The stabilization system rolls off low frequency corrections. I am not sure it works as low a 1 Hz let alone 0.1 Hz.
I don't use IS to try to get sharp photos at 1 sec SS. I hope to get sharp photos at 1/100 sec since I am pretty shaky, not even close to being an inertial reference frame. I can't blame the Earth's rotation for my many blurry photos.

Dave

P.S. You may want to check my arithmetic.
Feb 01, 2020 at 06:20 PM
gdanmitchell
Offline
Details
p.6 #15New Canon R5 Rumor
Anyone who makes an exposure without first calculating a Coriolis compensation coefficient isn't a real photographer.
Feb 01, 2020 at 07:25 PM
davidnumrich
Offline
Details
p.6 #16New Canon R5 Rumor
This would be insane if true and assuming the electronic shutter is as good as the Sony a9's, it would make me consider moving to the EOS R system, but I have a hard time believing that Canon would release a camera that is as good or better in almost every regard as their upcoming 1DX Mark III, not to mention some of their dedicated cinema cameras.
Feb 01, 2020 at 08:01 PM
Lotuselite
Offline
Details
p.6 #17New Canon R5 Rumor
Anyone who makes an exposure without first calculating a Coriolis compensation coefficient isn't a real photographer.

You forgot to include Schuler oscillations Dan.
Feb 02, 2020 at 12:08 AM
RottenTheCat
Offline
Details
p.6 #18New Canon R5 Rumor
gdanmitchell wrote:
Anyone who makes an exposure without first calculating a Coriolis compensation coefficient isn't a real photographer.


I thought that algorithm was implemented in Digic III and later......

Feb 02, 2020 at 12:49 AM
lighthound
Offline
Details
p.6 #19New Canon R5 Rumor
Sounds like I finally have a legitimate excuse for my blurry pictures.
Feb 02, 2020 at 06:16 AM
gdanmitchell
Offline
Details
p.6 #20New Canon R5 Rumor
Lotuselite wrote:
Earlier quoteAnyone who makes an exposure without first calculating a Coriolis compensation coefficient isn't a real photographer.
You forgot to include Schuler oscillations Dan.


Sorry, but I only had 84 minutes to post that... ;-)
Feb 02, 2020 at 08:48 AM
FM ForumsCanon ForumJoin Upload & Sell
  
 

You are not logged inLog inRegister