Canon, when they have bothered to say much about the subject have always stated that they wanted image quality to improve from model to model on their cameras.
Even though Canon’s definition of image quality at times was a bit suspect (for example JPEG-only output, etc), for the most part, they have held to that, with minor variations (for instance the R5 Mark II being slightly worse than the R5).
Speed also has an impact on image quality, from being able to capture movement without any distortion, to panning without any rolling shutter distortion. The faster the sensor is, the better.
In early 2024, Sony delivered the Global Shutter based A9 III, and it took a while but the results from DXO are in. I didn’t really expect much better from the A9 III (though it wasn’t the first full frame global shutter sensor – it was though the first full frame global shutter camera).
In the ultimate need for speed, the A9 III certainly delivers but there’s a cost and that’s across the board image quality and the base ISO.
If we look at the DXO results, we see that the A9 III scores low on both colour reproduction and also dynamic range. It should also be noted that the A9 III has a base ISO of 250.
| Model | Overall score | Color Reproduction | Dynamic Range | High sensitivity |
| α9 III | 87 | 24.1 | 13 | 3224 |
| α9 II | 93 | 25 | 14 | 3434 |
| α1 | 98 | 25.9 | 14.5 | 3163 |
| α7 IV | 97 | 25.4 | 14.7 | 3379 |
| EOS R3 | 96 | 25 | 14.7 | 4086 |
| Z9 | 98 | 26.3 | 14.4 | 2451 |
The high ISO results more or less catch up because at higher ISO’s the A9 III deficiencies are no longer important.
So what we have is a very narrow focused camera, more so than what people are suggesting for the R1 amusingly.
While I think the R3 Mark II would be a perfect choice, a few things are standing in the way.
The Nomenclature
When Canon implemented a pellicle mirror on the EOS-1N they specifically came out with a “RS” model (EOS-1N RS) that made it clear that this model was different and make sure you are aware of the differences before you purchase. Much like the 5D Mark III and the 5Ds and 5DsR naming conventions.
I know this is semantics, but it’s basically… Canon wouldn’t do the same tomfoolery that Sony has done.
I could see Canon releasing a R3 GS and the R3 line shifts from the R3 to a R3 GS, with a lineage progression from that model onward. Then there is less consumer confusion over what exactly you are purchasing. But my guess is that if Canon does use the R3 as a Global shutter introduction, it won’t be an R3 Mark II.
The Sensor Performance
But would Canon do this at all? The drop in image quality is high from the R3 to the R3 with a similar global shutter to that in the A9 III. Not to mention removing ISO 100 up to 250 removes a lot of utility that the camera may have outside of high ISO fast movement and action shooting.
This is something that traditionally Canon is very averse to doing, especially the amount of dynamic range, and color fidelity that we are talking about here. So, if Canon implements a Global Shutter, I think they will work on the problems that Sony for some puzzling reason decided not to.
From Canon (2024)
There are negative points and the image quality is affected. Sensitivity management is not very good and the dynamic range is less extensive. And I think we would have difficulty optimizing this, because of the very complex structure of this type of sensor.
If we have to integrate this into our flagship product, our flagship , we want to have complete technology. Philosophy.
https://www.canonrumors.com/canon-executive-interview-at-cp-2024/
Sony Electronics *does* have better performing global shutter technology, but that was not used in the A9 III. Perhaps cost was a factor, but with the Sony A9 III getting a $1500 price hike for “something” – you’d think that would cover increased fabrication complexity costs. And since they didn’t go all out, there’s hardly any fabrication difference between a normal stacked sensor to this global shutter sensor (so don’t ask me what Sony users are paying an extra $1500 for).
I would assume that Canon’s global shutter sensor will be far more expensive than Sony’s because they will work on image quality versus marketing messages.
That also may cause Canon to take a considerable amount of time to come out with such a camera.
Global Shutter Sensors are the Future
Just because Sony was the first to put it in a camera, really is a little meaningless, Global Shutter sensors have been here a while, but it’s optimally performing sensors that’s what the industry really needs.
I do like the benefits of Global Shutter and I think Canon (or Sony) will decide to implement fabrication methods that lower the image quality loss and improve the costs of these sensors.
Once this happens en mass, then we’ll see an industry adaptation to sensors and cameras with absolutely no mechanical “bits” outside IBIS as the rest of the internal guts will be all electronic.
Normal sensors with mechanical shutters will disappear as “yesterday’s technology” much like DSLRs did in the past.
I’m all for that happening sooner than later.
|
When you purchase through links on our site, we may earn an affiliate commission. Here's how it works. |
See full article...
SPAD sensor?
Still, market wise, I would think there's a bigger market for a higher megapixel camera in the R3 format. May not make sense for the R3 line though.
Maybe an R2 (or R1s) as a higher megapixel top-end camera, like 65MPx but with the same AF pixel distribution and viewfinder as the R1, for portrait/wildlife shooters/enthusiasts; and a 20-24MPx global shutter R3 because... they can?
I'm still skeptical of the market size for a global shutter R3 versus what they've already targeted with the R1.
Fixed, no idea what happened.
This seems like the best idea I have heard yet. I'm also extremely confident Canon is not going to release such a sensor until they are confident there is minimal image quality loss compared to say the EOS R1.
But I can take a look...
And I see the following hazy comment future:
"This should be called the R1II as the R3 should have been called R1 and the R1 should have been called R3. But canon wasn't keeping up with the competition."
Clouded by the dork side the future is.
No
After 2 years with the R5, I switched to 100% electronic shutter (haven't used strobes in a while). Even with the R7 I have been 100% on ES. With the R3, I never used the mechanical shutter, however, when I check the exposure release count it says <1000, because I do have it set to close the shutter when the camera is tured off.
It does feel good to have the shutter closed when changing lenses in the field, but I always wonder if it makes too much difference. If the shutter protects dust by collecting (at least some) of it, then once the lens is attached, the camera turned on and the shutter opens, presumably some of that dust will just be there and get to the sensor anyway.
So, is it useful to have a mecanical "protective shutter", cheaper to implement since it doesn't have any real performance requirement? Or is the benefit not large enough to justify having one?
My point being that the direction that camera manufactures go may not necessarily be the most technically innovative. Rather, 1) you either have to be so technically innovative that the innovation will sell a lot of cameras or 2) you have to save the manufacturer money.
Now that we have sub 3 ms readout speeds, does a GS do either?
I use my R3 in silent mode a lot, and when I don't remember to go back to normal shooting before putting the camera away, I have to clean the sensor more frequently.
Probably different weighting of the subjects in favour of the Sony.
FWIW, while DxO's algorithm is a black box I have seen it estimated as:
DxOMark_Sensor_Score = 58.8 + 4.3*(ColorDepth-21.1) + 3.4*(DynamicRange-11.3) + 4.4*log2(ISO/663)
Take that with a grain of salt.
A 4 year cycle for the R3 would be ~Nov-2025.
It could be as simple as the R5ii in the R3 body similar to Z9/Z8 once the sales of the R5ii have eased in a year.
It could be to introduce a new stacked sensor faster than the R1 (or GS) which would push the price of the R3 GS closer to R1 and perhaps cannibalising R1 sales. Might not be as bad as it seems once the initial R1 sales have been made.
It could be a very high res sensor but I would imagine that the R5s would be a more natural evolution similar to 5Dr(s). Landscapers would not use portrait orientation as often as landscape but portrait shooters would use the integrated grip orientation frequently.
I am sure that there will be (or has been) focus groups run by Canon for ideas but as the R3 doesn't have the same continuous history as 1/5 series then they won't necessarily be constrained.