Canon EOS R7 Mark II to Have Stacked 40MP Sensor?

I'm still a bit concerned because the original R7's noise performance was quite bad above 3200, and with the greater pixel density it could be even worse if they don't do something else. Also, is this the highest density pixel sensor that Canon has(or will) released?
johnchio, I am with you. Overall my R7 surprised me with a much better IQ than I expected for such a 30+ MP sensor. But, like you say, above 3200 noise kicks in very visibly (fortunately, todays photo apps with good RAW converters such as DxO can "restore" noisy images surprisingly well, but there are limits). 40 MP on an APS-C sensor means in particular for birders/wildlife shooters that if you want to freeze action with a long tele lens on the pixel level (what you want if you have such a camera), you need to go to very high shutter speeds (or you just have to forget about 40 MP and render all images down to, say, 20 MP). That means that you have to accept high ISOs in many settings if you go for maximum resolution. If such a sensor is stacked, the photo-active pixel area is even smaller than on a conventional sensor, so this boosts noise in addition - and shrinks the dynamic range again. Plus, diffraction blur kicks in at very low f-stop numbers already with such small pixels.

Therefore, like you, I'd really prefer a sensor with at least the same pixel count than the original R7 (I doubt Canon's marketing will accept a step back to bigger pixels). Stacked would then be very welcome, because the slow sensor readout really limits the useability of the R7's electronic shutter.
 
  • Like
Reactions: 1 user
Upvote 0
40Mpx is only 11% more linear resolution than 32.5Mpx, and each 40Mpx pixel is only 10% narrower so it's not going to make much difference in the DLA, the demands on lens resolution etc. It will make no difference to the S/N and high iso when you look at the whole image, because they are independent of pixel size. And if you pixel peep and look at the engineering S/N, each pixel will have only 10% less S/N, ie, less than 1/6th stop.
What me bothers more is the fact, that, if these rumors turn out to be true, 40 MP comes with a stacked sensor. That means an additional loss of dynamic range, and because these pixels are already quite small, I guess that this could be a huge trade-off. If it will be a BSI sensor like rumored, then the trade-off wouldn't be that dramatic, because it can use at least the maximum photo-active pixel area possible. So let's hope for BSI coming true.
 
Upvote 0
I don´t believe in a 40mp stacked sensor because that camera would really go upmarket towards R6iii pricing or maybe even more. I do believe in a 40 MP sensor or a 30/32 Mp stacked or partially stacked sensor and price point somewhere around 2+ K €.

Pricing in general will be very interesting because it kind of tells us where Canon sees the R7 in the line-up. I´d figure it'll be higher than R8 e.g. from an ergonomics POV with a joystick and scroll wheel, the LP-E6 battery family compatibility and with a really good sensor somewhere for around 2.299 € (or 2.399 €). It leaves room for discounts (200-300 €) and still has a good gap between R8 and R8.

I don´t believe in "mini-R5" and higher pricing than R6iii.
 
Upvote 0
If we see a high resolution R7 II, I dare say, for marketing reasons at the very least, we'll see a higher resolution R1 or R5.

Will be interesting to see the readout speed, as I believe the fast readout speed of the R1 has compromised the low light capabilities vs the R3. Th R7 II would be an interesting companion to the R1 in the same way the 7D II was to the 1DX.

Surely the #1 requested feature for the R7 II is a battery grip, though?
 
  • Like
Reactions: 1 user
Upvote 0
Sure, makes sense. For the same focal length FF and APS-C size is the same regardless of the image circle. But not for the same effective FL. So a wildlife shooter on APS-C can shoot his 800mm equivalent using a 500mm lens. It is thus much smaller but because of the focal length, not so much the size of the image circle.
Still not a reason to make an APS-c specific 500mm lens, which was the original point. Besides, the main reason many of us use the R7 is to get the extra reach with that 800mm lens which now looks like 1240mm from a framing perspective, or more like 1120mm from a pixel density perspective when comparing to an R5. So an R7 with an 800mm gives almost identical pixels on the bird as an R5 with the same 800mm and a 1.4 extender. The R7 rig is lighter and avoids the optical losses in the extender.
 
  • Like
Reactions: 1 user
Upvote 0
Absolutely true. And, the Canon RF 100-400mm f/8 (which fits FF) is much lighter than the OM 100-400 f/6.3 to prove your point about lighter by being slower.
Very handy lens, that 100-400 even if it is f/8 😁 and it is still quite decent at 560mm f/11 with the 1.4 extender. Contrast is a little low with the extender, but the detail is still there, so if you don't push the ISO too far, the results are very nice.
 
  • Like
Reactions: 1 user
Upvote 0
Depends how you look at it. As it stands Canon essentially outsources its APSC line of lenses to Sigma, which offer 4 fast primes (14, 23, 30, 56mm f1.4) and 4 zoomes (10-18mm f2.8, 18-50mm f2.8, 17-40mm f1.8, and 16-300mm) for Canon RF-S. And they are good quality. That covers the general use pretty good, such as family, travel, portrait, etc.. For wildlife people like me would be pleased to use my full frame lenses such as the RF 100-500mm. I have an R5. The R7ii would provide better reach with the same lens and would add to my kit, but not replace a full frame body. But it is true Canon has no high quality APS-C long lens. Many will be happy to use full frame lenses in those scenarios. Maybe sigma will introduce a long fast lens to RF-S.
Just for completeness, there are 5 fast primes, all F1.4: 12mm, 16mm, 23mm, 30mm and 56mm.

I own seven of those Sigma lenses in RF mount so I mostly agree with you. Disagreements: (1) It's 17-40 f/1.8 (2) Sigma needs to add a small 50-135 or 140 f/2.8 and (3) Sigma needs to redesign their 16 f/1.4 to make it much smaller and add a control ring and maybe add a control ring to their other older APS-C primes as well.
In addition to the small F2.8 zoom you mention, I'd be really interested in an RF upgrade to the EF 50-100mm F1.8, especially if they can acheive the same weight and focal length improvements they did when upgrading the EF 18-35mm F1.8 to the 17-40mm F1.8. Say they could get a 45-115mm F1.8 lens that weighs less than the EF 50-100mm. That would be fantastic!
 
  • Like
Reactions: 1 users
Upvote 0
Sigma needs to add a small 50-135 or 140 f/2.8
Sony users have long lamented the lack of a ~45-135mm f2.8 APS-C lens to match the common 70-200m f2.8 telephoto zoom. I think once upon a time Sigma was rumored to be working on one, but clearly that never came to be. Fuji has a 50-140mm f2.8, but they're the only APS-C maker to do so. Now that there's potentially 4 mounts (X, E, Z, RF-S), maybe it'll finally make sense for Sigma or whomever to make one.
 
Upvote 0
I could imagine that Canon, through the use of backside illumination, has kept the actual sensitive photosite area constant or near constant compared to the R7's, offsetting the effects of increasing the resolution to 40MP (which on its own would decreases the linear size by 15%). This would allow to maintain or improve the dynamic range through advances in sensor technology since the R7 (which, iirc, uses the sensor technology from the 90D, so from almost 7 years ago? Please correct me if I am wrong!), even despite stacking.

I doubt they would release a camera to the market that has worse low-light performance than the predecessor, especially since this new model is aimed to shift up-market a bit, also increasing buyers' expectations regarding image quality (and the target audience being wildlife photographers).
 
Last edited:
  • Like
Reactions: 1 user
Upvote 0
I could imagine that Canon, through the use of backside illumination, has kept the actual sensitive photosite area constant or near constant compared to the R7's, offsetting the effects of increasing the resolution to 40MP (which on its own would decreases the linear size by 15%). This would allow to maintain or improve the dynamic range through advances in sensor technology since the R7 (which, iirc, uses the sensor technology from the 90D, so from almost 7 years ago? Please correct me if I am wrong!), even despite stacking.

I doubt they would release a camera to the market that has worse low-light performance than the predecessor, especially since this new model is aimed to shift up-market a bit, also increasing buyers' expectations regarding image quality (and the target audience being wildlife photographers).
Why in the name of God do you even need 40MP on an APS-C? It's already 'cropped' in, the pixel density is quite high and for what?
 
  • Haha
  • Like
Reactions: 1 users
Upvote 0
Why in the name of God do you even need 40MP on an APS-C? It's already 'cropped' in, the pixel density is quite high and for what?
More pixels on the bird/mammal/... is always welcome, if it doesn't negatively impact IQ compared to its predecessor. Sure, a 24MP R1 will have vastly superior low-light and noise performance, but for the rest of us who cannot afford that and/or extremely expensive fast telephoto primes, a higher-resolution R7 is "good enough". It seems that Canon shares that view.
 
  • Like
Reactions: 1 user
Upvote 0
More pixels on the bird/mammal/... is always welcome, if it doesn't negatively impact IQ compared to its predecessor. Sure, a 24MP R1 will have vastly superior low-light and noise performance, but for the rest of us who cannot afford that and/or extremely expensive fast telephoto primes, a higher-resolution R7 is "good enough". It seems that Canon shares that view.
We shall see, more pixels isn't good if there's more noise tbh
 
  • Like
Reactions: 1 user
Upvote 0
I could imagine that Canon, through the use of backside illumination, has kept the actual sensitive photosite area constant or near constant compared to the R7's, offsetting the effects of increasing the resolution to 40MP (which on its own would decreases the linear size by 15%). This would allow to maintain or improve the dynamic range through advances in sensor technology since the R7 (which, iirc, uses the sensor technology from the 90D, so from almost 7 years ago? Please correct me if I am wrong!), even despite stacking.
A pixel in a 40 Mpx sensor is only 10% smaller than in a 32.5 Mpx one of same size, as I replied to your post on wednesday. And the dynamic range of the whole image at high iso is unaffected as it depends on the area of the entire sensor, not individual pixels, and the noise is primarily due to statistical fluctuations in the number of photons at high iso.

40Mpx is only 11% more linear resolution than 32.5Mpx, and each 40Mpx pixel is only 10% narrower so it's not going to make much difference in the DLA, the demands on lens resolution etc. It will make no difference to the S/N and high iso when you look at the whole image, because they are independent of pixel size. And if you pixel peep and look at the engineering S/N, each pixel will have only 10% less S/N, ie, less than 1/6th stop.
 
  • Like
Reactions: 2 users
Upvote 0
And for your purpose, open gate video will give you a big step up without the need for an 8K-capable sensor.
Only if they redesign the RF-S7.8MM F4 STM Dual to project light onto the extra sensor height 😉. But with the current design, in stills mode, that extra height of the 3:2 image is just blackness that doesn't aid the image.
Flowers-Sample-Picture.webp


When taking a 33MP still on the current R7, after processing it ends up being 4MP (2514x1676). When taking 4K videos on the R7, after processing it ends up being 1472x828, barely better than 720p. Also I have to choose between 4K Fine (oversampled) at 30p or 4K (binned) at 60p, which is tough because 3D looks better at higher framerates.

With this lens, you lose a lot of resolution because first there's smaller image circles so a lot of the pixels are just black, then you lose more when you crop those circles into rectangles, and finally each eye only gets 1/2 the remaining resolution, so in the end you've basically lost 85%-90% of your resolution.

When it's all said and done, 8K video recording would provide approximately a 720p -> 1080p jump, and any TV buyer will tell you that's totally worth it. On the other hand, open gate recording will gain me a bunch of black pixels that get cropped out.
 
  • Like
Reactions: 1 user
Upvote 0
I've been using the R7 (along with my R6) since it was released. Funny that I think about it because I remember people on this forum and in DPR saying APS-C was a thing of the past and no reason for Canon to continue crop sensors into the R line. I shoot a lot of macro and apsc is, in ways, better suited so I kept my 80D alongside my R6 and was so relieved with the R7s release. I bought it immediately.

My wishlist for the R7 mkii is pretty simple.


1) Focus bracketing *with flash* ability. M 4/3 cameras have taken so many macro photographers for this reason alone. This alone would make me upgrade.

2) Fix the focusing issues while in high speed.

3) format buttons/controls to match the R/6 and R/5. It would be mich easier muscle memory if things were in same place between different cameras.

** Just to add - If the mechanical shutter goes, the readout will have to be much, much better. I, and a lot of people, use the R7 for birds in flight. Just the lens IS alone gives the wobbles along with rolling shutter.
 
Upvote 0