The Canon EOS R3 will be officially announced in September

Nov 2, 2016
849
648
There seems to be a correlation between low resolution and high ISO.
There does not appear to be a similar correlation with dynamic range.
It’s not the resolution per se. it’s the larger site size. Otherwise every sensor would have the same dynamic range if they were of the same generation, which they don’t.
 
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
It’s not the resolution per se. it’s the larger site size. Otherwise every sensor would have the same dynamic range if they were of the same generation, which they don’t.
Right, the current cameras all have vastly differing DR ... and the one with the highest DR is of course equipped with a low resolution ... right?


Right? :unsure:
 
Upvote 0
Feb 28, 2013
1,612
272
70
Canon need to get into the soap opera business they really seem to be stretching out their product launches. Why the hell tease something in early Spring that your not going to fully announce until September and sell from November / December.
Gone it seems are the days they keep something secret and then announce & start selling a month later.
It starts to become negative publicity.
 
  • Like
Reactions: 1 user
Upvote 0
Jul 21, 2010
31,088
12,851
I know very well how it works. You don’t seem to.
Do you? Then please use your vast knowledge to explain the position of the Canon R5 in the list below (which is from Bill Claff's website). The only cameras that exceed the DR of the Canon R5 have MF sensors (and the R5 is better than some MF sensors).

Camera ModelMaximum
PDR ▴
Low Light
ISO
Low Light
EV
Phase One IQ4 150MP13.112252412.82
Phase One IQ3 100MP13.061280012.00
Hasselblad H6D-100c12.89856511.42
FujiFilm GFX 10012.301041311.70
Hasselblad H6D-50c12.01640011.00
Hasselblad X1D-50c11.98905811.50
Phase One IQ25011.92640011.00
FujiFilm GFX 50S11.90624110.96
FujiFilm GFX 50R11.86621810.96
Canon EOS R511.85543510.76
Pentax 645Z11.77515710.69
Hasselblad CFV II-50c11.73469210.55
Sony ILCE-7R11.71410810.36
FujiFilm GFX 100(FF)11.68630710.98
Sony ILCE-7RM311.65578710.85
Nikon D85011.63411510.36
Sony ILCE-7RM411.62474410.57
Nikon D61011.61396910.31
 
  • Like
Reactions: 1 users
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
Do you? Then please use your vast knowledge to explain the position of the Canon R5 in the list below (which is from Bill Claff's website). The only cameras that exceed the DR of the Canon R5 have MF sensors (and the R5 is better than some MF sensors).

Camera ModelMaximum
PDR ▴
Low Light
ISO
Low Light
EV
Phase One IQ4 150MP13.112252412.82
Phase One IQ3 100MP13.061280012.00
Hasselblad H6D-100c12.89856511.42
FujiFilm GFX 10012.301041311.70
Hasselblad H6D-50c12.01640011.00
Hasselblad X1D-50c11.98905811.50
Phase One IQ25011.92640011.00
FujiFilm GFX 50S11.90624110.96
FujiFilm GFX 50R11.86621810.96
Canon EOS R511.85543510.76
Pentax 645Z11.77515710.69
Hasselblad CFV II-50c11.73469210.55
Sony ILCE-7R11.71410810.36
FujiFilm GFX 100(FF)11.68630710.98
Sony ILCE-7RM311.65578710.85
Nikon D85011.63411510.36
Sony ILCE-7RM411.62474410.57
Nikon D61011.61396910.31
Just to be fair, it is worth pointing out that the R5 tops the chart thanks in part to featuring baked in noise reduction at low ISO.

But of course, apart from Mr. Claffs tests specifically looking for this, I have not seen anybody noticing any downsides coming from the noise reduction technique Canon uses. And even without it, it is certainly not at any Form of disadvantage compared to vastly lower resolution sensors like the one found in the R6 and 1DX III.
 
Upvote 0
Jan 29, 2011
10,675
6,121
Of course we could get into the weeds with regards PDR, which is a normalization and where some manufacturers are cooking the RAW files to some extent, and the actual capacity of the sensor to capture a wider range of a Stouffer style step wedge. The we can move to the differences in practical and engineering differences in a noise floor!
 
Upvote 0
Sep 20, 2020
3,065
2,395
Just to be fair, it is worth pointing out that the R5 tops the chart thanks in part to featuring baked in noise reduction at low ISO.

But of course, apart from Mr. Claffs tests specifically looking for this, I have not seen anybody noticing any downsides coming from the noise reduction technique Canon uses. And even without it, it is certainly not at any Form of disadvantage compared to vastly lower resolution sensors like the one found in the R6 and 1DX III.
Even Camon states R5 has more DR than R6.
It would be pretty tough to prove that is all because of "baked in noise reduction".
Especially, since pretty much every camera has some degree of noise reduction.
Canon is by far not the worst offender.
 
  • Like
Reactions: 1 user
Upvote 0

entoman

wildlife photography
May 8, 2015
1,998
2,438
UK
I am officially announcing that I will NOT be purchasing this camera.

Barring any unknown lottery winnings, I am officially broke.... I can only dream of one day owning it.
Haha, I officially announced that I wouldn't be getting an R5.

6 months later, look at my gearlist...
 
  • Haha
Reactions: 1 user
Upvote 0

AlanF

Desperately seeking birds
CR Pro
Aug 16, 2012
12,343
22,519
Everywhere you look there are shortages of high-end toys, cars, watches. Partly due to unprecedented demand from central banks flooding liquidity due to COVID recesssion worries and the chip shortages. I have five RF lenses ordered and I am weeks and months from getting them. Maybe not so bad the R3 is announced in September.
Do you count the five RF lenses and R3 as high-end toys?
 
Upvote 0
Jan 29, 2011
10,675
6,121
And why would that be? I don't understand.

BSI eliminates the wiring that is blocking a part of the light in conventional sensors. As there has to be more of this wiring on a higher resolution sensor, it actually should reduce the difference between high end low resolution even more in terms of noise - and the difference is already negligible in my eyes:


And DR depends so heavily on noise. I know a single large pixel has higher DR than a single small one - but we are usually care about images, and therefore average multiple pixel together to compare image sections of the same area. Which results in the same light and noise per area, so where is the advantage for lower resolution sensors coming from?

To be clear I do expect Canon to improve both in terms of DR and low light performance with this new sensor. It marks a pretty strong shift in their manufacturing technology after all. I just don't see that improvement being related to the resolution and fully expect the upcoming, higher resolution bodies (R5s, R1?) to outperform it.
The issue with BSI for larger sensors is that the micro lenses replace a lot of the light/efficiency lost to the wiring.

BSI made a huge difference in small sensors, like phones and P&S's, but as the sensor gets larger the percentage of light loss due to the wiring is not as big, and, as I say, the microlens arrays over the sensors replace a lot of that loss. When Sony went to FF BSI the DR didn't increase at all from their previous generation non BSI sensors.
 
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
The issue with BSI for larger sensors is that the micro lenses replace a lot of the light/efficiency lost to the wiring.

BSI made a huge difference in small sensors, like phones and P&S's, but as the sensor gets larger the percentage of light loss due to the wiring is not as big, and, as I say, the microlens arrays over the sensors replace a lot of that loss. When Sony went to FF BSI the DR didn't increase at all from their previous generation non BSI sensors.
I guess you are referring to my last paragraph there, in explaining why we may see little to no gain in DR due to the new sensor tech? As this does not look like an attempt to explain why larger pixels should lead to more DR in an image.

If so, good point.
 
Upvote 0
Jan 29, 2011
10,675
6,121
I guess you are referring to my last paragraph there, in explaining why we may see little to no gain in DR due to the new sensor tech? As this does not look like an attempt to explain why larger pixels should lead to more DR in an image.

If so, good point.
The bigger pixel argument stems from a difference in the way DR is measured. Do you measure at a pixel level or do you normalize? If you measure at the pixel level one big pixel is less affected by some types of noise for a given exposure (depending on a billion caveats like same generation processing, etc etc) than a small pixel. Also a bigger pixel potentially has a bigger full well capacity which can be used to increase DR.

If you normalize across same generation then sensors are sensors and nothing beats area, and this has been true for many years.

I believe the meme originated back in the day before microlens arrays when on sensor wiring was comparatively inefficient and for a smaller pixel higher resolution sensor the wiring took up a bigger percentage of the sensor area than the same sized sensor with fewer pixels, this meant the former were not as efficient and prone to more noise. This has not been true for several generations of sensor.
 
Last edited:
  • Like
Reactions: 2 users
Upvote 0
Aug 7, 2018
598
549
I downloaded RAWs from this site https://www.dpreview.com/reviews/image-comparison/fullscreen at ISO 100 and increased the exposure five stops to compare the ISO noise in underexposed areas. Shouldn't the R6 and the 1D X III have the same sensor? The 1D X III has much lower noise. I wonder how the R3 will compete. I also saw that my 1D X is VERY bad in that regard. So I really need a new camera. WIth the 1D X you can't really recover shadows at all even at ISO 100 without introducing a lot of noise.
 
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
The bigger pixel argument stems from a difference in the way DR is measured. Do you measure at a pixel level or do you normalize? If you measure at the pixel level one big pixel is less affected by some types of noise for a given exposure (depending on a billion caveats like same generation processing, etc etc) than a small pixel. Also a bigger pixel potentially has a bigger full well capacity which can be used to increase DR.

If you normalize across same generation then sensors are sensors and nothing beats area, and this has been true for many years.

I believe the meme originated back in the day before microlens arrays when on sensor wiring was comparatively inefficient and for a smaller pixel higher resolution sensor the wiring took up a bigger percentage of the sensor area than the same sized sensor with fewer pixels, this meant the former were not as efficient and prone to more noise. This has not been true for several generations of sensor.
I am aware and fully agree with all of this. Hence, my emphasis on DR in an image, not in a pixel. Which I assume the typical consumer and certainly Canon is focusing on when it comes to choice of MP.
 
Upvote 0