Canon EOS R7 Mark II: What We Know

After 200 000 exp on mechanical shutter, some thousands on electronic, and lots of video, most wildlife and birds, I`m done with the R7i. The price tag was too low, and it was not a successor of the 7Dii. In short, and its about shooting in demanding conditions shooting birds and wildlife: The AF misses a lot, and even report hits when it is a miss. AF also often jumps off after some shots in a burst, even in slower pace. Far too few keepers. Electronic shutter can`t be used on moving objects because of the very slow read out speed that gives a horrible rolling shutter. 1. curtain electronic works fine, but very noisy.

Else a very good camera which delivers excellent results under less demanding conditions. I've had no issues with them (we have two), and they have survived lots of sub arctic hard weather on shorter and longer trips all year round. But do not expect the R7 to compete with top shelf /full frame) cameras on challenging objects.

I have waited for the upgrade to come. And I hope it will be a top shelf camera in the APS-C class. A little sibling to R5ii. 32 MP is enough, but most important the camera has to deliver better focus, with AF in high class, have pre shooting in RAW, fast electronic shutter and keep the noise down, companied with eye tracking like the big 3 (plus R6iii? Guess it will be a standard on top cameras). APS-C format has some advantages Canon should lift to the top shelf. My big wish for the R7ii (or what its name will be) is to be a real companion together with my new R5ii.

I bought the R7 thinking it was the successor and came to the same conclusion. I found the noise was unacceptable for a modern camera. It missed a LOT, even in not particularly demanding situations. I also found that the shutter was incredibly noisy and jolting. It felt like mirror slap almost. I felt like I was holding a $300 entry-level model. I stuck with my R5 for closer subjects, and brough the R7 for if I was shooting stuff over the ocean, or raptors higher up. Even in those situations the R7 was tough to get good results from, and in anything other than perfect lighting the noise was disappointing. After about 4 months I set the R7 on a shelf and haven't touched it since. I've never used a body less than that one.
 
Upvote 0
I bought the R7 thinking it was the successor and came to the same conclusion.
The R7 was the successor, but to the 90D not the 7DII. The 7-series was Canon's red-headed step child, and the R7 is part of that series in numeral only. Perhaps the R7II will be the philosophical successor to the 7DII, but I am not holding my breath for that to happen.
 
  • Like
Reactions: 2 users
Upvote 0
I own a 5D3 and a 5DS, both of which are FF. The maximum possible ISO for the 5D3 is 51200 and I regularly used it a 6400 and occasionally at 12,800. The maximum ISO for the 5DS is 6400 and I never use it above 3200.
So? I used the 5D3 and 5Ds for years and found them essentially identical in ISO performance. I would personally top out at about 12800 on both.
 
  • Like
Reactions: 1 users
Upvote 0
So? I used the 5D3 and 5Ds for years and found them essentially identical in ISO performance. I would personally top out at about 12800 on both.
The 5DsR actually has more DR at lower ISO and the same DR at higher ISO. But people will go on believing that smaller pixels mean more noise. You can lead people to facts, but you can't make them comprehend.

5DIII-5DsR.png
 
  • Like
Reactions: 4 users
Upvote 0
So? I used the 5D3 and 5Ds for years and found them essentially identical in ISO performance. I would personally top out at about 12800 on both.
Here's to corroborate what you claim, using the 5Div rather than 5D3, from 7 years ago. Two shots of a male Bellbird (in New Zealand). One was taken by me with an EF 400mm f/4 DO ii on a 5Div at f/4 and iso 6400. The other taken by my wife at the opposite side at f/5.6 with an Ef 400mm f/5.6 II on a 5DSR at iso 6400, pushed through 1.56 ev in post, so equivalent to iso 18,900. They are bathed in green light as shaded deeply by green leaves. The 5DSR shot is reduced in size to that of the 5Div to post on site. Both were processed using the 2017 version of DxO. The 5Div shot is on top. The noise on the 5DSR is fine at an iso approaching 25,600.


bellbird_male_2B4A8000_DxO_CRiso6400.jpgbellbird_male_3Q7A8969DxO_bellbirdmale_50%.jpg
 
  • Like
Reactions: 6 users
Upvote 0
Two people believing something is true doesn’t make it so. The Flat Earth Society has lots of members.

Pixel size does matter…for resolution and diffraction. Pixel size matters for noise from an individual pixel, for those who are addicted to measurebation.

Pixel size does not matter for image noise (any more, though it did before the advent of gapless microlenses).

Same size sensor + smaller pixels = same noise floor = same DR (e.g., R6II vs R5II). Smaller sensor + same size pixels = higher noise floor = less DR (e.g., R5II in crop mode).

View attachment 223212
Our Headmaster (= US Principal) invited a member of the Flat Earth Society to lecture to us at school so we could get used to the cases put forward by those of such ilk. 60 years on, the experience is more useful than ever.
 
  • Like
Reactions: 1 users
Upvote 0
Here's to corroborate what you claim, using the 5Div rather than 5D3, from 7 years ago. Two shots of a male Bellbird (in New Zealand). One was taken by me with an EF 400mm f/4 DO ii on a 5Div at f/4 and iso 6400. The other taken by my wife at the opposite side at f/5.6 with an Ef 400mm f/5.6 II on a 5DSR at iso 6400, pushed through 1.56 ev in post, so equivalent to iso 18,900. They are bathed in green light as shaded deeply by green leaves. The 5DSR shot is reduced in size to that of the 5Div to post on site. Both were processed using the 2017 version of DxO. The 5Div shot is on top. The noise on the 5DSR is fine at an iso approaching 25,600.


View attachment 223219View attachment 223220
Then, I'd really like to know why Canon restricted the top ISO for the 5DS and 5DSr.

Update: I went back to my stored images of the Imaging Resources test bench. I had stored images of the 5Ds at ISO 6400 both with no noise reduction and with maximum in-camera noise reduction. I'll admit that the max reduction is highly effective but that raises the question: does using heavy noise reduction on a 50MP image result in something less than a 50 MP image? Normally, I use no noise reduction at all.
 
Last edited:
Upvote 0
Then, I'd really like to know why Canon restricted the top ISO for the 5DS and 5DSr.

Update: I went back to my stored images of the Imaging Resources test bench. I had stored images of the 5Ds at ISO 6400 both with no noise reduction and with maximum in-camera noise reduction. I'll admit that the max reduction is highly effective but that raises the question: does using heavy noise reduction on a 50MP image result in something less than a 50 MP image? Normally, I use no noise reduction at all.
I always use DxO PL on RAW files. In general, I use no additional noise reduction than the default ones. The 5DSR results are as clean as I get nowadays with my R5 and R5ii. I really loved the 5DSR, and don't know why they restricted the iso range.
 
  • Like
Reactions: 1 users
Upvote 0
Then, I'd really like to know why Canon restricted the top ISO for the 5DS and 5DSr.
Perhaps Canon felt that higher ISOs were more than the buffer/processors could handle without undesirable slowdowns in frame rate? At ISO 12800, 5Ds RAW file size is ~88 MB. Extrapolating from other cameras, at ISO 51200 files would be >110 MB.
 
  • Like
Reactions: 2 users
Upvote 0
Perhaps Canon felt that higher ISOs were more than the buffer/processors could handle without undesirable slowdowns in frame rate? At ISO 12800, 5Ds RAW file size is ~88 MB. Extrapolating from other cameras, at ISO 51200 files would be >110 MB.
Maybe. The bottleneck would have to be fairly early in the processing chain. The 5Ds can reduce its jpeg image size from 50MP (L) to 39MP (M1) or 22MP (M2). I mostly use it at M2, unless I expect to display or print the image extremely large. For 20X30, 22MP is enough. ISO 12800 is HI, which I've never used. The normal maximum is ISO 6400.
 
Upvote 0
User requests would help - it's been five months since the 56mm f/1.4 and the other four lenses came out.

After seeing so many articles saying that C1P was so much better than LR in terms of processing, I bought a license (a few years back).

Lo and behold it was missing SO many profiles - not even an eos M profile at all. Within an hour it was simply unworkable for me. :mad:

I gave it one more shot, and then gave up. I tried to held out from adobe's yearly fee, but it became untenable.

I've been sitting pretty ever since, never a worry. Sometimes (company) scale does matter.
 
Upvote 0
After seeing so many articles saying that C1P was so much better than LR in terms of processing, I bought a license (a few years back).

Lo and behold it was missing SO many profiles - not even an eos M profile at all. Within an hour it was simply unworkable for me. :mad:

I gave it one more shot, and then gave up. I tried to held out from adobe's yearly fee, but it became untenable.

I've been sitting pretty ever since, never a worry. Sometimes (company) scale does matter.
I've been using DxO for many years and have always found it does a better job than Lightroom in rendering RAW files, because DxO, a testing lab, creates their own lens profiles. Whenever I've had to go back to Lightroom - which I also have - I've found its results unsatisfying - which is why I'm so vexed by their foot-dragging on the Sigma APS-C lenses that they've already profiled in other lens mounts.
 
  • Like
Reactions: 1 users
Upvote 0
The 45 Mpx R5 sensor reads out twice as fast as the R7's 32 Mpx sensor. The R7 sensor is of an old design, and new 40 Mpx sensor could be at least twice as fast, I would presume. I agree about the change in ergonomics between the R7 and the R5 series is a bit of pain for those of us who have both.
I know, but the R5 was nearly about 3x the price of the R7. It is not a technical issue but knowing Canon's product policy over many years I am not sure that Canon will "allow" R7 II users to get the fastest readout speeds Canon can engineer with conventional sensors. Would be great, of course, and even more attractive to upgrade (I am already willing to do that).
 
  • Like
Reactions: 2 users
Upvote 0
I've been using DxO for many years and have always found it does a better job than Lightroom in rendering RAW files, because DxO, a testing lab, creates their own lens profiles. Whenever I've had to go back to Lightroom - which I also have - I've found its results unsatisfying - which is why I'm so vexed by their foot-dragging on the Sigma APS-C lenses that they've already profiled in other lens mounts.
I use DxO since some months and like some of its features very much, e.g. its noise canceling plug-ins are really impressive. But its lens profiles are at least partly underwhelming. One example is Tamron's EF 24-70mm f/2.8 "G1" lens that I still use: the with auto setting of the profile of this lens still pincushion distortion is massively visible on images with straight lines (architecture). Fortunately, you can change that manually, that's again a good feature of DxO. I will never return to LR, overall I decided to quit all Adobe products because of their subscription only policy. I prefer the classic product policy: pay once and after, say, two years for an upgrade or so. Much better way to keep costs under control.

The best package I ever used - of course IMO - was Darktable. For instance its intelligent auto correction of perspective distortion lines worked in many cases surprisingly well, much better than DxO's, obviously some experts in pattern recognition programmed this tool that DxO doesn't have. Because of its big scene of enthusiasts Darktable offered a huge selection of really great tools - much more than I ever managed to try. The downside was, like with all free apps, that its interface wasn't always user friendly (like Gimp) and needed a steep learning curve, plus it's instability - you could never be sure that your favorite tool plug-ins did always work. Anyway, unfortunately they stopped to support Apple's OS, so it's over now for Mac users like me.
 
Upvote 0
My reasoning against the 200-800 or even the 100-500 is I live in the northern US (Michigan), so half of my sports photography is indoors, and those lenses are simply too slow for shooting in field houses or gymnasiums.
I have the RF 200-800 and I learned to like this lens more than I expected when I bought it (with a tad of frowning whether this is a good decision). But I only use it outdoors with enough light, mostly for birding, then it is absolutely great, lot of fun to use. I never would recommend to use it indoors for sports in poor lighting. Plus, even if you shoot it at the short end of about 200mm, it only offers f= 6.3, so you do not get that 3D isolation of an athlete in front of a maybe closer background that I think a good sports photographer requires. So for that purpose the 200-800 would be definitely too slow.
 
Upvote 0
How does dual SD provide more security than CFe plus SD?

Note that sensor read out speed is one of the factors determining AF performance (at least, as far as subject tracking is concerned). A faster read out speed means better sampling of moving subjects.
That reminds me of another issue of the current R7 I hope Canon will improve with the Mk II: too small buffer plus a quite limited writing speed on UHS-II SD cards (here is a test that fits to quite well to my own experience: https://rfshooters.com/blog/cameras/canon-r7/memory-cards/). So, with 30 fps you will most probably miss the interesting part of an action scene, because the camera stops recording too early. I mostly use therefore 15 fps full electronic or electronic 1st curtain shutter speeds for action, but still the buffer/card writing bottleneck kicks in often too early.
 
Upvote 0
I agree that 40mp will be a bad idea. Diffraction will start to be noticeable at F5.6 or so, negatively affecting most people's telephoto lenses. Hopefully Canon is not so stupid as to pay attention to the idiots on social media who constantly clamor for more MP's and have no clue what diffraction is - or much else about actual photography..
I am with you. Paired with the RF 200-800 f/6.3-9.0, which is a really nice combo with the R7 Mk I, that combo already makes clear that physics is definitely working against even higher MP sensors. In fact, the current R7 I is already pushed well over the theoretical diffraction limit of such a combo, and I guess, the R7 is mostly a birder's camera, and that combo is getting quite popular I observed. The rule of thumb for the diffraction limit is "pixel pitch x 2 = max. aperture no. with no visible diffraction softening". The R7 I has a pixel pitch of 3.2 µm, so we get f = 6.4, which in fact is already well below f = 9. That said, diffraction softening is a quite good-natured loss of optical information: digital sharpening (enhancement of micro contrasts) doesn't bring that lost information back but makes the f/9 images look crisp and detailed enough for our eyes that we perceive them as sharp. I think that Canon quirks already RAW images look "better" that way, and that's okay. In the end, if an image looks good that's all we want from photography.

The growing selection of very to extremely slow supertele lenses that Canon brought (and maybe brings) out technically speaks against even higher MP APS-C sensors, based on the law of wave optics. But the marketing of course wants to make those happy who always think "more is better". Give the people wot they want. :cool:
 
Last edited:
Upvote 0
That reminds me of another issue of the current R7 I hope Canon will improve with the Mk II: too small buffer plus a quite limited writing speed on UHS-II SD cards (here is a test that fits to quite well to my own experience: https://rfshooters.com/blog/cameras/canon-r7/memory-cards/). So, with 30 fps you will most probably miss the interesting part of an action scene, because the camera stops recording too early. I mostly use therefore 15 fps full electronic or electronic 1st curtain shutter speeds for action, but still the buffer/card writing bottleneck kicks in often too early.
I find that Raw Burst Mode helps a lot.
 
Upvote 0
I own a 5D3 and a 5DS, both of which are FF. The maximum possible ISO for the 5D3 is 51200 and I regularly used it a 6400 and occasionally at 12,800. The maximum ISO for the 5DS is 6400 and I never use it above 3200.
I’m surprised. I’ve found that at the same output size as the 5Diii the 5DS knocks the socks off it at high iso performance.
 
Upvote 0
I bought the R7 thinking it was the successor and came to the same conclusion. I found the noise was unacceptable for a modern camera. It missed a LOT, even in not particularly demanding situations. I also found that the shutter was incredibly noisy and jolting. It felt like mirror slap almost.
Isn’t this noise shot (photon) noise, just physics of light - random photons etc coupled with the small capacity pixels ? Same complaints came from some people moving up to the 60mp Sonys. Imagine what the 100mp FF sensor would be like that all these mp hungry people want !
Regarding the shutter, I’m sure it’s just to remind you that you didn’t fork out for a R5. (A bit like the RP.)
 
Upvote 0