Is a Canon EOS R7 Mark II coming this year? [CR2]

Quack said:

1. The R7 has some additional AF features compared to the R5, like pre-capture and subject tracking in all AF zones, but whether it is "better" is debatable.
2. Yes, the R7 should resolve more detail than the R5 with 1.4x TC.
3. Depends on how you define image quality, but in my opinion, the answer to this would be no.

4-7. all valid, in my opinion.

Click to expand...

The pixel pitch of the R7 is 3.2µ and the R5 4.39µ, so the R7 potentially resolves 4.39/3.2 x the detail, ie 1.37x. Put a 1.4x on the R5 and it gives 1.4x in theory, which is pretty much a wash, in theory. But, on the one hand the 1.4x will degrade resolution to some extent and on the other the R5 sensor has a much weaker AA-filter and better resolution. Optyczne measures the R7 sensor to have only about 15% more resolution than the R5 with an f/4 lens, rather than 37%, pretty much in line with my own experience. And that decreases with increasing f-number. In practice, the R5 with 1.4xTC outresolves the R7.

@neuroanatomist I looked at the Dynamic Range versus ISO charts comparing my R7 to the APS-C R6 Mark II results - and the R5 Mark II APS-C as well -and I see that they're all pretty much the same, showing that my proposed kludge of putting my Sigma RF-S 56mm f/1.4 onto my R6 Mark II wouldn't improve noise performance, and by implication that giving an R7 II a 20-24mp sensor wouldn't help either.

I also compared the R5 II and R6 II full-frame results and saw those were pretty comparable to each other and about one stop better than the R7.

I bought the R6 II (instead of the R5) not only for better noise than the R7 but also because I wanted to stick with dual V90 SDXC cards, instead of having one CF card, which I don't need since I'm not a videographer or high frame-rate shooter (I'm happy with 3fps, since the only action I photograph is at chorus concerts).

Buying the RF-S 56mm f/1.4 Sigma is about the best I can do to improve low light performance from the R7 - and it's a nice sharp 90mm equivalent lens, to boot, in a tiny package, I'm very happy with it, just impatient for DxO to release the profile for it. Until then I'm using 45% pincushion and 65% vignette settings in DxO PhotoLab.

I'm attaching a few shots from this weekend the day after I got the lens, taken in a common living room area concert with not a lot of light, with the R7 at ISO 400.

Philnick said:

This does not address using the same lens on a crop sensor versus a full-frame sensor at the same distance from the subject.

Click to expand...

If you do that, to get the same picture you would need to crop the resulting FF image to the framing of the crop sensor, in which case you’d have equivalent noise and DR in the two images (more noise and less DR than the FF sensor can deliver, as shown above when using a FF camera in crop mode). You’d probably have more MP in the APS-C image…but that wouldn’t affect noise or DR.

I’m not a fan of DPR, but they did a decent job on this topic: The effect of pixel size on noise.

The best explanation of these concepts I’ve run across is this one: Equivalence.

Philnick said:

Exposure and noise are characteristics of each pixel, not global across the entire sensor…

Click to expand...

As I stated, exposure is determined by light per unit area (at the pixel level, if you prefer) but noise is not. Just compare the image noise at something fairly high, e.g., ISO 6400, on FF vs. APS-C – much higher on APS-C even though the exposure is the same.

The noise at ISO 2500 on APS-C is about the same as the noise ISO 6400 on FF. Quantitatively, the difference is a bit over 1⅓-stops [technically, it’s log₂(2.56) since the area of a FF sensor is 2.56-times larger than APS-C (for Canon)].

After reading this thread and the above linked articles, if you want to keep believing pixel size determines noise, go right ahead. Like those who believe the earth is flat and won’t let facts influence their belief, you are free to believe what you want even if it’s wrong.

neuroanatomist said:

If you do that, to get the same picture you would need to crop the resulting FF image to the framing of the crop sensor, in which case you’d have equivalent noise and DR in the two images (more noise and less DR than the FF sensor can deliver, as shown above when using a FF camera in crop mode). You’d probably have more MP in the APS-C image…but that wouldn’t affect noise or DR.

I’m not a fan of DPR, but they did a decent job on this topic: The effect of pixel size on noise.

The best explanation of these concepts I’ve run across is this one: Equivalence.


As I stated, exposure is determined by light per unit area (at the pixel level, if you prefer) but noise is not. Just compare the image noise at something fairly high, e.g., ISO 6400, on FF vs. APS-C – much higher on APS-C even though the exposure is the same.

The noise at ISO 2500 on APS-C is about the same as the noise ISO 6400 on FF. Quantitatively, the difference is a bit over 1⅓-stops [technically, it’s log₂(2.56) since the area of a FF sensor is 2.56-times larger than APS-C (for Canon)].

After reading this thread and the above linked articles, if you want to keep believing pixel size determines noise, go right ahead. Like those who believe the earth is flat and won’t let facts influence their belief, you are free to believe what you want even if it’s wrong.

Click to expand...

My last post was my surrender on this point. The Dynamic Range vs. ISO page you pointed me to settled the matter - the APS-C results were the same for the R5 II, R6 II, and R7, and the R6 II 24mp and R5 II 45mp full frame results were about equivalent and each about a stop better than the R7 results.

It seems that buying fast glass is the best thing for the R7. Good thing Sigma has just been allowed to bring its very popular APS-C DC DN lenses to RF-S. If anything can make the R7 popular, those lenses will be it. I'm using their RF-S 18-50mm f/2.8 DC CN and their RF-S 56mm f/1.4 DC DN. They've also released an RF-S 10-18mm f/2.8 and about three other RF-S f/1.4 primes.

My two RE-S Sigmas are each about 10oz in weight. The RF-S 18-50mm f/2.8 normal zoom (29-80mm equivalent) is about the size of the R7's kit lens, and the 56mm f/1.4 prime is even smaller.

I can't imagine shooting at ISO 2500 on my R7 - maybe because I've got an old PC and saving JPGs out of DxO Photolab with its most advanced noise reduction running can take half an hour per picture! (I used no noise reduction on the three shots attached to my prior post.)

I'm an old Kodacolor 400 shooter, and that's where I try to keep it, since that's its highest base ISO.

neuroanatomist said:

The best explanation of these concepts I’ve run across is this one: Equivalence.

Click to expand...

I just bookmarked that equivalence link.

I have always used this one.

Less verbose and friendlier formatting.

Philnick said:

My last post was my surrender on this point. The Dynamic Range vs. ISO page you pointed me to settled the matter - the APS-C results were the same for the R5 II, R6 II, and R7, and the R6 II 24mp and R5 II 45mp full frame results were about equivalent and each about a stop better than the R7 results.

Click to expand...

There's a reason so many articles are written about this!

Philnick said:

It seems that buying fast glass is the best thing for the R7. Good thing Sigma has just been allowed to bring its very popular APS-C DC DN lenses to RF-S. If anything can make the R7 popular, those lenses will be it. I'm using their RF-S 18-50mm f/2.8 DC CN and their RF-S 56mm f/1.4 DC DN. They've also released an RF-S 10-18mm f/2.8 and about three other RF-S f/1.4 primes.

Click to expand...

Yes, the faster the better. I really liked the 85/1.2L II on my 7D (I had the 85/1.8 and replaced it with the f/1.2), and my standard zoom for APS-C DSLRs was the EF-S 17-55/2.8. But honestly, the only reason I see to use APS-C cameras anymore is for the smaller size of the kit, mainly for travel. If shooting in bright light absent the need for shallow DoF, my EOS M6II does fine, and at a pinch it's ok indoors with the M22/2. I also travel with a full spectrum converted M6.

The advantages of APS-C are basically limited to smaller kit size and lower kit cost. People say there's an advantage when 'focal length limited', but IMO that's still about cost and size/weight. My birding lens is a 600/4 used with a 1.4x TC for 840mm f/5.6, and at typical subject distances atmospheric effects (air distortion) are significant and basically outweigh the benefit of a crop sensor. If I needed more MP, I'd use an R5II instead of an R3/R1. But being able to get to 800mm framing with a 100-500 on APS-C means less weight to carry and a lot lower cost. Of course, the latter means losing ~2 stops (2/3 in aperture and 1-1/3 in ISO noise). That's not a problem in good light, but at sunrise/sunset it can be. Equivalence means no free lunch.

Philnick said:

I can't imagine shooting at ISO 2500 on my R7 - maybe because I've got an old PC and saving JPGs out of DxO Photolab with its most advanced noise reduction running can take half an hour per picture! (I used no noise reduction on the three shots attached to my prior post.)

Click to expand...

DxO DeepPrime XD2 is excellent, but it does take CPU/GPU power to run. On my prior Mac (2019 Intel Core i9), processing a batch of 20-30 RAWs with DeepPrime took a while and the fans sounded like a jet taking off. On my new M4 Pro MacBook Pro, it takes a minute or two and the fans stay off.

Philnick said:

I'm an old Kodacolor 400 shooter, and that's where I try to keep it, since that's its highest base ISO.

Click to expand...

I didn't do slide film, I shot Tmax 100 and 400 and developed the negatives and prints myself. But with a FF camera and DxO, I regularly go to ISO 25600 on my R3, and I'll see if I can get away with ISO 51200 on my R1. I shoot events indoors or under high school night field lighting, and even with f/2.8 (24-105 and 100-300), needing action-stopping shutter speeds means pushing the ISO way, way up.

neuroanatomist said:

. . .
DxO DeepPrime XD2 is excellent, but it does take CPU/GPU power to run. On my prior Mac (2019 Intel Core i9), processing a batch of 20-30 RAWs with DeepPrime took a while and the fans sounded like a jet taking off. On my new M4 Pro MacBook Pro, it takes a minute or two and the fans stay off.


I didn't do slide film, I shot Tmax 100 and 400 and developed the negatives and prints myself. But with a FF camera and DxO, I regularly go to ISO 25600 on my R3, and I'll see if I can get away with ISO 51200 on my R1. I shoot events indoors or under high school night field lighting, and even with f/2.8 (24-105 and 100-300), needing action-stopping shutter speeds means pushing the ISO way, way up

Click to expand...

Kodacolor was color negative film - kind of like shooting RAW.

Kodak's slide films were Kodachrome (neutral colors) and Ektachrome (saturated colors). I preferred saturated, but mostly stuck to negative film for its greater dynamic range and thus could safely be over or under exposed by several stops, while slides had to be right in camera.

I've always preferred to shoot negatives and post process to taste. I got started with photography at the age of 8 when my father taught me how to develop and print black and white pictures. (I even used the short-lived home color printing process in the 1970s.)

The EF-S 17-55mm f/2.8 was my main lens from when I started with the Rebel XT up through my 70D. When I went mirrorless with the R7 I fell in love with its similarity in size and weight to the Canon FT-QL 35mm film SLR I used from 1968 to 2006 with an FL 85mm f/1.8 lens.

Going digital had meant that every body was bigger and heavier than the last until the R7 took me back to the small rig I had used for film for nearly 40 years.

The EF-S 17-55mm plus EF to R adapter was too big and heavy for that on the R7 and had to go. The Sigma RF-S 18-50mm f/2.8 DC DN is its reincarnation in a much smaller and lighter package.

Similarly, for reasons of size and weight, I traded down from my EF 24-70mm f/2.8L II USM plus EF to R adapter to the new RF 28-70mm f/2.8 IS STM, which is about half the weight and two thirds the length - and half the price of the RF 24-70. (Though Canon says the new 28-70 f/2.8 has "L quality optics and full weather sealing" they don't call it an L. No free lens hood, no red ring, but half the price! So I bought a $10 JJC clone lens hood, festooned a green rubber band with a few red twist ties, put them around the front of the lens barrel and said "it's a joyous no L.")

The Sigma RF-S 56mm f/1.4 on the R7 (90mm equivalent) competes with the Canon RF 85mm f/2 Macro on the R6 II.

When I shoot a concert I'll have the RF 28-70mm on the R6 II and the Sigma 56mm - or pehaps the 85mm f/2 - on my R7

For wide-angle shots I'll use my tiny and inexpensive RF 16mm f/2.8 on the R6 II. Can't cost justify any of the 10mm lenses for the R7, since I don't see that way.

Philnick said:

Kodacolor was color negative film - kind of like shooting RAW.

Kodak's slide films were Kodachrome (neutral colors) and Ektachrome (saturated colors). I preferred saturated, but mostly stuck to negative film for its greater dynamic range and thus could safely be over or under exposed by several stops, while slides had to be right in camera.

Click to expand...

Thanks for the correction. Should have remembered Paul Simon's song...

neuroanatomist said:

Thanks for the correction. Should have remembered Paul Simon's song...

Click to expand...

Of course he was really singing about Ektachrome, but non-photographers probably wouldn't have recognized the term.

AlanF said:

....And that decreases with increasing f-number. In practice, the R5 with 1.4xTC outresolves the R7.

Click to expand...

In my practice, comparing perhaps a 100 images, hand held, taken with both the R7 and R5 with 1.4x TC, the R7 outresolved the R5 in most, if not all. So, clearly, there are many factors in play (f stop, atmospheric conditions, hand held vs tripod, lens sharpness, 1.4x sharpness, etc.). I always urge folks who are trying to decide on buying a TC to buy one, test and compare, and then find out for themselves. Always buy from stores that offer returns.

Quack said:

In my practice, comparing perhaps a 100 images, hand held, taken with both the R7 and R5 with 1.4x TC, the R7 outresolved the R5 in most, if not all. So, clearly, there are many factors in play (f stop, atmospheric conditions, hand held vs tripod, lens sharpness, 1.4x sharpness, etc.). I always urge folks who are trying to decide on buying a TC to buy one, test and compare, and then find out for themselves. Always buy from stores that offer returns.

Click to expand...

I always give the same advice - try for yourself as there is copy variation of both extenders and lenses. Here are actual measurements of the resolving powers of the R7 and R5 measured by optyczne.pl. The R7 sensor has a very powerful AA-filter and underperforms its theoretical resolution compared with the R5 and its new generation AA-filter. An R7ii with a modern Canon sensor would be a real improvement.

Quack said:

In my practice, comparing perhaps a 100 images, hand held, taken with both the R7 and R5 with 1.4x TC, the R7 outresolved the R5 in most, if not all. So, clearly, there are many factors in play (f stop, atmospheric conditions, hand held vs tripod, lens sharpness, 1.4x sharpness, etc.). I always urge folks who are trying to decide on buying a TC to buy one, test and compare, and then find out for themselves. Always buy from stores that offer returns.

Click to expand...

I have actually posted a thread on the RF 200-800mm on the R7 and R5, showing my images of charts. You can see clearly that this relatively narrow lens that is not the sharpest gives no more resolution on the R7 than on the R5 but just puts twice the number of pixels on target at 800mm f/9. Adding the 1.4x RF TC doesn't improve the resolution of the R5 but again doubles the number of pixels.


I've dug through my other tests and here are two charts, from slightly further away (19m vs 16.7m). The wider aperture and very sharp RF 100-500mm + 1.4xTC on the R5 at 700mm outresolves the bare lens on the R7. Your results may differ, and wider apertures and light intensity will have effects. But, I do not find the R7 outresolves the R5 with the same lens plus 1.4xTC.