Canon EOS R3: What do we know, and what can it tell us about a future ‘EOS-R1’?

The EOS R3 clearly inherits aspects of Canon’s EOS-1D series of cameras, but what else can we learn from what Canon has told us?

Canon has just announced the development of what will be the highest-sped RF-mount camera yet, the EOS R3. The few details that have been released point towards a very interesting camera, but also something else; something even more capable, coming in the future.

So although Canon hasn’t given away a lot away in this announcement, it’s provided enough lines for us to read between them.

The name:

The most obvious implication of the R3 announcement comes from the product’s name. Canon’s flagship pro sports models have been named as some variant on ‘1’ as far back as the F-1 SLR which debuted in 1971. The EOS-1 came along in 1989, and the first flagship digital model, the EOS-1D, was released in 2001. The fact that the R3 isn’t branded as an ‘R1’ strongly suggests Canon has something planned to sit above it.

Back in the late 1990s Canon had the EOS-3 which sat between its EOS-5 and EOS-1 series models and, just as with the R5 and R6, I doubt this re-use of numbering is coincidental.

1998’s EOS-3 shared a lot of technology with the EOS-1V of 2000. What it didn’t share was quite same level of build and reliability that pros depend upon.

This could mean one of two things: firstly, in terms of price, it may mean that the R3 will sit between the $3,900 price point of the EOS R5 and somewhere in the region of the $6,500 price tag implied by the current range-topping EOS-1DX Mark III. Of course, the hypothetical R1 could also be even more expensive than the existing DSLR flagship, making it the preserve of agency shooters and sideline pros, with the R3 providing a more accessible option non-professionals. Just like the film-era EOS-3 did, back in the 90s and early 2000s.

Either way, the limited specs Canon has given so far: 30 fps shooting with continuous AF, improved AF for sports shooting and a dual grip design all suggest that on spec at least, the R3 may be a more powerful camera than the 1D X Mark III, despite its more lowly numerical title. That, in turn, sets even higher expectations for an R1 level model.

Stacked CMOS

The other factor that boosts our hopes for both cameras is Canon’s adoption of Stacked CMOS technology. We recently wrote that the use of Stacked CMOS is likely to be the key enabling technology underpinning almost every aspect of the performance of Nikon’s promised Z9 pro-level mirrorless camera. There’s every reason to suspect the same will be true for Canon.

Stacked CMOS is likely to be the enabling technology underpinning Nikon and Canon’s promised pro-level mirrorless cameras

Canon says its chip will be a Dual Pixel design, developed in-house. This means we can’t be certain of the approach it will use: in itself the Stacked CMOS approach doesn’t necessarily mean the use of in-sensor RAM, as is the case in Sony Semiconductor designs, but this is an obvious means of supporting the fast sensor readout that Canon is promising.

Interestingly, it means we’ll see Canon jump directly from conventional Front Side Illumination chips to Stacked CMOS, without passing through the intermediate step of making single piece (non-stacked) BSI chips.

Canon has a history of persisting with older manufacturing processes (such as larger component scale) for longer than most of the rest of the industry, squeezing everything possible out of its manufacturing investment before moving on. This is very different to companies that also make smartphone sensors, and regularly have fabrication equipment being passed down as their fine-scale small sensor products move on to the next cutting-edge process.

To Canon’s credit, the low light and readout performance of the EOS R5 sensor is extremely high for an FSI sensor, and it could be that Canon decided that there weren’t sufficient benefits to justify the cost of moving from FSI to BSI in the relatively large pixel sensors it makes. The jump to Stacked CMOS (essentially next-gen BSI) opens up more significant scope for performance improvements though, so it’ll be exciting to see what the R3 and hypothetical future R1 will do with it. This level of ambition would be consistent with the apparent leap forward we saw with Canon’s most recent R models.

Eye input AF

The other attention-grabbing detail in the R3 announcement is that it will have ‘Eye input AF,’ that will select an AF point based on where the photographer is looking. Canon users of a certain age are likely to be reminded of the ‘Eye Controlled Focus’ system from the 1990s that did much the same thing.

The EOS-3 featured the most ambitious version of Canon’s Eye Controlled Focus system. The 45 AF points that made up the system were all very tightly packed, and the EOS R3 is likely to need to work with much higher precision if it’s to be useful.

Eye Controlled Focus didn’t make the leap across to Canon’s digital cameras and people have speculated whether it could make a comeback ever since. Clearly similar sentiments persisted somewhere within Canon.

It’s been 17 years since the last camera to offer Eye Controlled Focus and AF systems have only become more complex in the interim, with many more selectable points spread over a wider area (the most points the old system ever had to cope with was 45; in the EOS-3, as it happens). But equally, in the meantime, sensing technologies and subject tracking algorithms have become significantly more advanced. As subject-aware AF tracking systems have shown, cameras have become very adept at identifying and tracking small moving subjects with a high degree of precision. This isn’t quite the same challenge, but it’s similar enough, conceptually, to raise our hopes that it’ll be up to the job. That said, subject tracking has become so good, there’s a question mark over how necessary it’ll be, at all.

Wrap-up

The EOS-1D X Mark III includes a large battery, a substantial image buffer and connections including Ethernet ports that pro sports shooters rely upon. The R3 won’t necessarily do so.

What the specs make clear is that the R3 is going to be an ambitious camera. Just over a year ago, we said that the 20 fps-in-liveview EOS-1D X III looked like a pretty capable mirrorless sports camera that just happened to still be trapped inside a DSLR body. But the R3, with its promise of 30 fps shooting and improved AF, makes it highly possible that Canon’s second-string camera may well outperform its current flagship. This in itself is an exciting prospect, even before you start to explore the implications for a future R1.

The EOS R3 clearly inherits aspects of Canon’s EOS-1D series of cameras, but what else can we learn from what Canon has told us?

Canon has just announced the development of what will be the highest-sped RF-mount camera yet, the EOS R3. The few details that have been released point towards a very interesting camera, but also something else; something even more capable, coming in the future.
So although Canon hasn’t given away a lot away in this announcement, it’s provided enough lines for us to read between them.
The name:
The most obvious implication of the R3 announcement comes from the product’s name. Canon’s flagship pro sports models have been named as some variant on ‘1’ as far back as the F-1 SLR which debuted in 1971. The EOS-1 came along in 1989, and the first flagship digital model, the EOS-1D, was released in 2001. The fact that the R3 isn’t branded as an ‘R1’ strongly suggests Canon has something planned to sit above it.
Back in the late 1990s Canon had the EOS-3 which sat between its EOS-5 and EOS-1 series models and, just as with the R5 and R6, I doubt this re-use of numbering is coincidental.

1998’s EOS-3 shared a lot of technology with the EOS-1V of 2000. What it didn’t share was quite same level of build and reliability that pros depend upon.

This could mean one of two things: firstly, in terms of price, it may mean that the R3 will sit between the $3,900 price point of the EOS R5 and somewhere in the region of the $6,500 price tag implied by the current range-topping EOS-1DX Mark III. Of course, the hypothetical R1 could also be even more expensive than the existing DSLR flagship, making it the preserve of agency shooters and sideline pros, with the R3 providing a more accessible option non-professionals. Just like the film-era EOS-3 did, back in the 90s and early 2000s.
Either way, the limited specs Canon has given so far: 30 fps shooting with continuous AF, improved AF for sports shooting and a dual grip design all suggest that on spec at least, the R3 may be a more powerful camera than the 1D X Mark III, despite its more lowly numerical title. That, in turn, sets even higher expectations for an R1 level model.
Stacked CMOS
The other factor that boosts our hopes for both cameras is Canon’s adoption of Stacked CMOS technology. We recently wrote that the use of Stacked CMOS is likely to be the key enabling technology underpinning almost every aspect of the performance of Nikon’s promised Z9 pro-level mirrorless camera. There’s every reason to suspect the same will be true for Canon.
Stacked CMOS is likely to be the enabling technology underpinning Nikon and Canon’s promised pro-level mirrorless cameras
Canon says its chip will be a Dual Pixel design, developed in-house. This means we can’t be certain of the approach it will use: in itself the Stacked CMOS approach doesn’t necessarily mean the use of in-sensor RAM, as is the case in Sony Semiconductor designs, but this is an obvious means of supporting the fast sensor readout that Canon is promising.
Interestingly, it means we’ll see Canon jump directly from conventional Front Side Illumination chips to Stacked CMOS, without passing through the intermediate step of making single piece (non-stacked) BSI chips.
Canon has a history of persisting with older manufacturing processes (such as larger component scale) for longer than most of the rest of the industry, squeezing everything possible out of its manufacturing investment before moving on. This is very different to companies that also make smartphone sensors, and regularly have fabrication equipment being passed down as their fine-scale small sensor products move on to the next cutting-edge process.
To Canon’s credit, the low light and readout performance of the EOS R5 sensor is extremely high for an FSI sensor, and it could be that Canon decided that there weren’t sufficient benefits to justify the cost of moving from FSI to BSI in the relatively large pixel sensors it makes. The jump to Stacked CMOS (essentially next-gen BSI) opens up more significant scope for performance improvements though, so it’ll be exciting to see what the R3 and hypothetical future R1 will do with it. This level of ambition would be consistent with the apparent leap forward we saw with Canon’s most recent R models.
Eye input AF
The other attention-grabbing detail in the R3 announcement is that it will have ‘Eye input AF,’ that will select an AF point based on where the photographer is looking. Canon users of a certain age are likely to be reminded of the ‘Eye Controlled Focus’ system from the 1990s that did much the same thing.

The EOS-3 featured the most ambitious version of Canon’s Eye Controlled Focus system. The 45 AF points that made up the system were all very tightly packed, and the EOS R3 is likely to need to work with much higher precision if it’s to be useful.

Eye Controlled Focus didn’t make the leap across to Canon’s digital cameras and people have speculated whether it could make a comeback ever since. Clearly similar sentiments persisted somewhere within Canon.
It’s been 17 years since the last camera to offer Eye Controlled Focus and AF systems have only become more complex in the interim, with many more selectable points spread over a wider area (the most points the old system ever had to cope with was 45; in the EOS-3, as it happens). But equally, in the meantime, sensing technologies and subject tracking algorithms have become significantly more advanced. As subject-aware AF tracking systems have shown, cameras have become very adept at identifying and tracking small moving subjects with a high degree of precision. This isn’t quite the same challenge, but it’s similar enough, conceptually, to raise our hopes that it’ll be up to the job. That said, subject tracking has become so good, there’s a question mark over how necessary it’ll be, at all.
Wrap-up

The EOS-1D X Mark III includes a large battery, a substantial image buffer and connections including Ethernet ports that pro sports shooters rely upon. The R3 won’t necessarily do so.

What the specs make clear is that the R3 is going to be an ambitious camera. Just over a year ago, we said that the 20 fps-in-liveview EOS-1D X III looked like a pretty capable mirrorless sports camera that just happened to still be trapped inside a DSLR body. But the R3, with its promise of 30 fps shooting and improved AF, makes it highly possible that Canon’s second-string camera may well outperform its current flagship. This in itself is an exciting prospect, even before you start to explore the implications for a future R1.Read MoreArticles: Digital Photography Review (dpreview.com)

Leave a Reply

%d bloggers like this: