The R7 Mark II with 39MP: It Sounds Spot on – and here’s why it makes sense

[Richard CR]

Okay, so the rumors are sounding as if it appears that the Canon EOS R7 Mark II is indeed set to arrive in the first half of 2026 with a new 39MP APS-C sensor of BSI variety, either stacked or unstacked. As someone who enjoys delicious megapickles, I'll be happy with 39 versus 32.5, even […]

See full article...

 

[Maximilian]

Whatever pixel count the Mark II will get, whatever sensor design, the performance will be very interesting.
Can't wait what readout speed and what DR and s/n it will deliver. I hope your predictions are right @Richard CR.

Yes, I really want this lens for my next APS-C camera, as Canon doesn’t think we deserve a modern fast kit zoom.

And I can tell you, you will love the 18-50 DC DN. At least I do.

 

[Nemorino]

IMO the question is if the DPAF performance will suffer from the smaller pixels.

 

[Richard CR]

IMO the question is if the DPAF performance will suffer from the smaller pixels.

*assuming stacked sensor*

The efficiency gained by going stacked from front side illumination should be enough to offset the small efficiency loss by slightly smaller pixels.

In theory, of course.

 

[swingman]

It’s all depends on readout speed. If same or slower than mark I, no birders would want it even pixies on the birds are higher

 

[Planterz]

I'm holding out for hope of a stacked sensor. If it's not stacked, it would need to at least have a fast enough readout speed, but I'm not sure it'd be fast enough to satisfy those of us who want, above all else, a fast readout speed. For me it's not just about warping bird wings or propeller/rotor blades, but also the longevity/durability of the camera. The more I can rely on electronic shutter and less on mechanical, the longer the camera (or at least the shutter mechanism) will last, as well as retaining possible resale value if that comes to pass. I'd honestly rather save up for an R5 II instead (not an impossibility regardless) than buy an R7 II that I wear out the mechanical shutter on because I suck and rely on spray-and-pray instead of skill.

 

[Sporgon]

Regarding the comments on a R5s or R3 I could see that from a marketing and sales position it would be beneficial to have the ‘high resolution’ model as a separate model - the R3ii. Also would then make the positioning in the line up clear.

 

[padam]

The more I can rely on electronic shutter and less on mechanical, the longer the camera (or at least the shutter mechanism) will last, as well as retaining possible resale value if that comes to pass. I'd honestly rather save up for an R5 II instead (not an impossibility regardless) than buy an R7 II that I wear out the mechanical shutter on because I suck and rely on spray-and-pray instead of skill.

If the shutter is well-designed to begin with, it's going to last for hundreds of thousands of actuations and not an issue at all.
Even if it does, it is considerably easier to replace than in a DSLR. Apart from the EOS R failing sometimes (first one with sensor protector), I don't know of other issues with Canon mirrorless regarding shutter wear.

 

[Richard CR]

If the shutter is well-designed to begin with, it's going to last for hundreds of thousands of actuations and not an issue at all.
Even if it does, it is considerably easier to replace than in a DSLR. Apart from the EOS R failing sometimes (first one with sensor protector), I don't know of other issues with Canon mirrorless regarding shutter wear.

Agreed - sort of.

the mirrorbox assembly was extremely complex but that was usually seperate from the shutter assembly outside of timing.

but I think Canon and others have done alot of work around mirrorless to make it smaller and easier to manage.

and yes, it must be certainly easier to replace as well.

 

[AlanF]

I'm holding out for hope of a stacked sensor. If it's not stacked, it would need to at least have a fast enough readout speed, but I'm not sure it'd be fast enough to satisfy those of us who want, above all else, a fast readout speed. For me it's not just about warping bird wings or propeller/rotor blades, but also the longevity/durability of the camera. The more I can rely on electronic shutter and less on mechanical, the longer the camera (or at least the shutter mechanism) will last, as well as retaining possible resale value if that comes to pass. I'd honestly rather save up for an R5 II instead (not an impossibility regardless) than buy an R7 II that I wear out the mechanical shutter on because I suck and rely on spray-and-pray instead of skill.

I really don't understand readout times. The R7 can be truly bad for bird and insect wings but I have never had any problems whatsoever with the R5, which is only twice as fast. The R5 does have rolling shutter effects on panning against backgrounds with vertical lines that come out sloping, but not with wings flapping.

 

[GMCPhotographics]

The issue with ultra high pixel density sensors is optical resolution. Most of Canon's current lenses are engineered to an optical resoution to match around 50mp. Any sensor that resolved higher than 50mp will make soft images because the lenses (when shot wide open) are not capable to resolving that level of micro detail. Stopping down should sharpen up the results....but diffraction starts to occurr. With a bright lens like a RF 50mm f1.4, there's a long way before diffraction starts to take effect. However, take the RF 200-800 which has a max aperture of f9 it's no wonder that there are reports of it being slightly soft wide open at 800mm on a R5 and noticably soft on a R7. Stoppping down on a R7 isn't going to help because the level of aperture related increase in sharpness is being undone by diffraction. A R7's diffraction starts at only f5.2. The R5's is f7.1and the R6ii is f9.6. The native aperture of the 200-800 is f9. So diffraction is already causing softness issues on a R5 (slightly) and quite a lot on a R7.
If our lenses could resolve at a higher MFT level, then diffration becomes less of an issue too. The only solution is sharper and brighter lenses.
The only camera test site that has sensor resoluton test charts is this polish site https://www.optyczne.pl/testy_aparatów.html and it makes for interesting opservations between brands and specific cameras. it gives is a sensor based MFT value and is roughly comparible to our lens MFT charts. if your sensor is lower than your lens MFT the you are probably going to see sharp results.
For example, my R6ii has a tested MFT of .64 lpmn. The R5 has .82 lpmn and the R7 a lofty .92 lpmn.
According to Canon's MFT charts, the RF 200-800's highest value is about .78 lpmn at 800mm. Where as the RF 100-500L has a value of about .83 lpmn at 500mm.
If we campare a very sharp lens like the RF 135mm f1.8 LIS, it measures a very high value of around .95 lpmn, which is rated at f1.8 and has a lot of stopping down sharpness potential until sensor diffraction starts to rob the optical sharpness. Many of the big whites fall into this super resolution and high brightness category.
The other big question is how many megapixels does the scene that we are photographing actually require. I get the whole "45mp gives me more cropability", but that only holds true if the scene has that much detail (heat haze etc) and if the lens can also resolve that level of data too. landscapers are always asking for more observable blades of grass and the current stae of the art is never enough. I think even the legendary Ansel Adams with is massive plate camera used to comment on "counting enough ants". But for a lot of photography there are limits in terms of useful reach and limits in the amount of detail that our phtotgraphable subjects can present to us in average light.
Sure you can use AI and LR sharpening routines, but we are talking about sensors and optics here, not post processing tweaks. Post Production can re-generate and hide a multitude of issues.

 

[Planterz]

If the shutter is well-designed to begin with, it's going to last for hundreds of thousands of actuations and not an issue at all.
Even if it does, it is considerably easier to replace than in a DSLR. Apart from the EOS R failing sometimes (first one with sensor protector), I don't know of other issues with Canon mirrorless regarding shutter wear.

Granted, and I understand that a camera's shutter mechanism should be considered a replaceable part with finite life. But in my case, with my Sony a6700, I feel like I'm wearing it out too quickly. I got it in Oct 2024, and have used only on a relatively few real outings.

1 day at the Davis-Monthan AFB to learn to use the camera taking pics of aircraft flying overhead during a training exercise weekend.
2 days planespotting at Tucson International Airport (shared with AZANG base).
4 days (3 1/2 really) photographing the Heritage Flight training at DM.
2 days at the airshow.

My shutter count is over 25K, which is 1/10th the rated life of the shutter mechanism. I feel like I hit that mark way too quickly, even though, with that math, it still means many years of life left in the camera. A fast enough electronic shutter readout that could replace the majority of my shutter actuations would extend the life of the camera. More to the point, maybe, it'd alleviate the anxiety/guilt I feel when I click away. I realize that I simply need to "get over it" and not worry about the mechanical shutter, but that doesn't mean I wouldn't feel better firing off bursts of 100+ with an electronic shutter instead just to maybe get 1 or 2 keepers. Know what I mean?

I really don't understand readout times. The R7 can be truly bad for bird and insect wings but I have never had any problems whatsoever with the R5, which is only twice as fast. The R5 does have rolling shutter effects on panning against backgrounds with vertical lines that come out sloping, but not with wings flapping.

Perhaps the way wings bend when they flap mask a bit of warping due to rolling shutter, and it still looks natural rather than artificially distorted.

 

[David - Sydney]

Integrated grip body vs mp is tricky marketing for Canon. There was the 1D/1Ds but also the 5Div/5DS(R).
Nikon gives 2 almost identical cameras with the choice of body which seems to be well received.
Canon brings out 3 grips for the R5ii rather than standard and integrated body versions.
Same for R3. Is it a baby R1 or a souped up R6. Will it become a high mp sensor but only in integrated body?
Will the new R7ii come in both standard and integrated body? My guess is not as they didn't release a native grip for the R7.

Canon have done some weird marketing in the past but it isn't clear to me how they can find their way out of these choices.

 

[P-visie]

Most of Canon's current lenses are engineered to an optical resoution to match around 50mp.

What is your source for this statement?

 

[padam]

My shutter count is over 25K, which is 1/10th the rated life of the shutter mechanism. I feel like I hit that mark way too quickly, even though, with that math, it still means many years of life left in the camera. A fast enough electronic shutter readout that could replace the majority of my shutter actuations would extend the life of the camera. More to the point, maybe, it'd alleviate the anxiety/guilt I feel when I click away. I realize that I simply need to "get over it" and not worry about the mechanical shutter, but that doesn't mean I wouldn't feel better firing off bursts of 100+ with an electronic shutter instead just to maybe get 1 or 2 keepers. Know what I mean?

Yeah, but how much your are willing to pay for that privilege?
The extra cost of a stacked sensor is way more than a shutter replacement.
This wouldn't be anywhere near the price of a R7 or A6700.

And to contradict that, you might end up with way more pictures than you need, and it needs way more time to cull through, reducing time spent on taking good pictures, which is not the same as spending time actually taking pictures. It's of course different for sports photographers, etc.

Seeing that I can use a film camera and end up with a fairly ok rate of keepers, I have cut down massively on digital, and yet I still feel I end up taking more than I actually need with new cameras - just too easy. And I rarely use burst mode to begin with, pretty easy to make multiples in single-shot (even if I do, it's most likely at the lowest setting).
I use my old DSLRs in a way that is somewhat closer to film, since I need to watch that focus instead of just firing away, but I do have a lag-free viewfinder so have some chance with the timing.
Same with my Leica SL2 with vintage lenses and no AF. I do use it a handful of times, but not take a lot of shots (and always turning it off, since that pretty finder eats batteries).

So my point is: yes, it may get more convenient is some ways but also more complicated in others as well as being arguably unnecessary for most people and more expensive.
Personally, for photography, I'd value a good EVF more than electronic shutter performance (although they are interconnected since the sensor provides the live feed). Canon's are ok, but not the best.

 

[AlanF]

Perhaps the way wings bend when they flap mask a bit of warping due to rolling shutter, and it still looks natural rather than artificially distorted.

This is what I have seen with ES on the R7, and nothing at all with the R5, which is nominally only twice as fast.

 

[GMCPhotographics]

What is your source for this statement?

General comments from press release and articles reported here by Canon's techs over the years. The EFII lenes were the first drive to match the coming higher resolution cameras, the resolution of the RF lenses are generally similar to those of the EFII's. In some cases the RF lenses are actually less sharp that the EFII's. If Canon were upping their optical resolution to match their comming higher resolution cameras then Canon new the approx mp density they were developing and knew their old mkI EF lenses were built to a lower resolving standard. it is reasonable to assume therr highest resolution camera, the 5DSR, rated at .79 lpmn on the link I quoted earlier. Camera's like the 22mp 5Diii were resolving at around .45 lpmn on this test bed. The R6ii totally eclipses the 5DIII with a score of .64lpmn. I found when I migrated from my 5DIII's to a R6ii that manyof my older EF (mk1) lenses were giving my soft results. This confirms my statement. R5 slightly eclipses the 5DSR with a value of .82lpmn. It seesm that newer pixels resolve better than older ones.
But it is reasonable to assume that Canon's legacy statements about needing sharper lenses wasn't arbitory but had a specific MFT value in mind that would match their MFT sensor targets.

 

[AlanF]

General comments from press release and articles reported here by Canon's techs over the years. The EFII lenes were the first drive to match the coming higher resolution cameras, the resolution of the RF lenses are generally similar to those of the EFII's. In some cases the RF lenses are actually less sharp that the EFII's. If Canon were upping their optical resolution to match their comming higher resolution cameras then Canon new the approx mp density they were developing and knew their old mkI EF lenses were built to a lower resolving standard. it is reasonable to assume therr highest resolution camera, the 5DSR, rated at .79 lpmn on the link I quoted earlier. Camera's like the 22mp 5Diii were resolving at around .45 lpmn on this test bed. The R6ii totally eclipses the 5DIII with a score of .64lpmn. I found when I migrated from my 5DIII's to a R6ii that manyof my older EF (mk1) lenses were giving my soft results. This confirms my statement. R5 slightly eclipses the 5DSR with a value of .82lpmn. It seesm that newer pixels resolve better than older ones.
But it is reasonable to assume that Canon's legacy statements about needing sharper lenses wasn't arbitory but had a specific MFT value in mind that would match their MFT sensor targets.

It's not the newer pixels per se that resolve better, it's the weakening or improvement of AA-filters. Canon has redesigned them to give much less blurring, and sometimes they are asymmetric, blurring in only one direction.

 

[GMCPhotographics]

This is what I have seen with ES on the R7, and nothing at all with the R5, which is nominally only twice as fast.

View attachment 227633

I've seen similar ruined images from the R5 too from a few friends images. It's a symptom of the confusion between sensor read out speed, the frequency of the wing flap and the size of the wing in the frame. The R7 is particaulary prone to this becuase it's ES readout speed is so poor, but it can affect R5 and R6 images too.

I've always considered the Electronic shutter to be completely useless with action shots on the R7. It's fine ofr static objects, which is anathema...who needs 30fps on a static object? The AF can't really keep up with the 30 fps, the complaints about the R7's AF are notorious and the general advice from Canon is to slow the frame rate in lower light. The R7's buffer is too small to consistently use 30fps and the sensor is nearly 30ms readout. the camera has a relly good electronic first curtain shutter of 15fps, which has a fantastically fast readout speed of 2.4ms (which is the fastest readout speed of any current Canon camera).
When you consider the AF latency, the buffer and the sensor readout, the R7 is really good upto 15fps in 1st curtain, only 5fps slower than the original R5's 20 fps ES.

I've not seen many of these read out artifacts on my R6ii with my visits to Sri lanka, but then again, I used 1st Curtain shutter a lot and not many Electronic Shutter mode. So I was seeing a readout speed of 3.4ms, only beaten by the R1's stacked sensor readout speed of 2.7ms.

 

[AlanF]

I've seen similar ruined images from the R5 too. It's a symptom of the confusion between sensor read out speed, the frequency of the wing flap and the size of the wing in the frame. The R7 is particaulrly prone to this becuase it's ES readout speed is so poor.
I've always considered the Electronic shhutter to be completely useless with action shots on the R7. The AF can't really keep up with the 30 fps, the buffer is too small and the sensor is nearly 30ms readout. the camera has a relly good electronic first curtain shutter of 15fps, which has a readout speed of 2.4ms.
When you consider the AF latency, the buffer and the sensor readout, the R7 is really good upto 15fps, only 5fps slower than the original R5's 20 fps ES.

With the R5ii on the other hand...