I Have an R5 MK II. AMA.

[Nrbelex]

I have a production R5 MK II. I'll be conducting some testing of my own, and I'm happy to take requests and upload RAWs (I'm not too concerned about burning my source), but note that my bandwidth is limited, so answers may not be immediate. My lenses on hand are the EF 24-70 f/2.8 (v1), EF 70-200 f/2.8 (v1), and EF 50 f/1.8 (v2), obviously all adapted. My prior camera I can compare against is a 5D Mk III. Adobe Camera Raw will process the RAWs, though this obviously may be preliminary or generic. I may experiment with DPP.

Test images will be available here:
https://drive.google.com/drive/folders/1Vkdc_YV1WtWw9FRYJ-Mv5uxKc4xlH1ll?usp=drive_link

As I start making real photos I'm happy with, you'll see them appear on Flickr here: https://www.flickr.com/photos/nrbelex/

PS - Over the coming days, and if he's interested, I'll try to provide Bill Claff what he needs for dynamic range testing per https://www.photonstophotos.net/Collaborations/Dynamic_Range_Collaboration.htm Hopefully my adapted EF glass is sufficient for this purpose.

PPS - Note that I'm cross-posting on both DPReview and Canon Rumors.

PPPS - If you're an established forum member (avoiding Canon trying to catch the source) in NYC and want to purchase the camera now from a Canon authorized retailer, contact me directly and I'll share my source.

 

[Nrbelex]

For those who struggled with Eye Control AF on the R3, as Canon mentioned, it is improved on the R5 MK II, but not solved. I could barely register an eye on the R3. On the R5 MK II, I can consistently register my eye, but the target does not perfectly follow it throughout the entirety of the frame. I'm not surprised, but it's a bit of a bummer. I wear contacts and have blue eyes, for what it's worth.

 

[Jethro]

My prior camera I can compare against is a 5D Mk III

Really interested in your analysis. I think you'll find a pretty dramatic improvement compared to the 5D3!

 

[Quarkcharmed]

Thank you - it will be interesting to see measurements in photonstophotos.

For now, DPR put the studio scene comparison images.

The dynamic range and overall high ISO performance took a hit it seems and look a bit disappointing.

Note that the second gain kicks in at ISO 400 in the R5 and at ISO 800 in the R5II. Also the R5 doesn't apply the infamous noise reduction in raw from ISO 800. So we can compare ISO 100 and ISO 800 between those cameras.

Looking at the deep shadows:

https://www.dpreview.com/reviews/im...23&x=0.03777228857777911&y=0.5007271810640153

... and comparing the EFCS vs EC in the R5II:

https://www.dpreview.com/reviews/im...23&x=0.03777228857777911&y=0.5007271810640153

we can tell the R5II has a significantly lower dynamic range at ISO 100 (say around 1 stop worse) and worse high ISO performance.

The ES may be 14 bits but it's still noticeably worse than the EFCS mode.

 

[Nrbelex]

I've uploaded RAW and JPEG shots taken with both the electronic and first curtain shutter, all at ISO 100. The first thing I've learned is that nothing in the EXIF data distinguishes electronic and mechanical shutter shots that otherwise have the exact same settings, which doesn't make this all that helpful a comparison. That said, I'm reasonably confident I can identify the electronic shutter shots as they seem to have more noise in the shadows when boosting the exposure. If that's right _88A0031 is electronic while _88A0030 is mechanical.

Update: If you open the RAWs in ExifTool, you can see the shutter type.

 

[Nrbelex]

Three big take-aways so far (all of which match my expectations going in):

1. While the image quality is certainly better than my 5D MK III, particularly at high ISOs, it's not a revolution.
2. The AF system, on the other hand, is a complete revolution. Having waited a few generations, I suddenly have things like eye detect AF and a plethora of options and modes that I'm confident will increase my keeper rate substantially.
3. IBIS is giving these old EF lenses a new lease on life. I'm able to hand-hold shots I'd never bother with previously.

P.S. My understanding is the R5 II without the cooling grip has no active fan. It's interesting then that it makes an audible sound in a quiet room, almost like a whirring, whenever powered on, even with IBIS turned off. Update: Apparently the R5 made such a sound as well, and it's very likely the IBIS unit (even though it was turned off).

 

[koenkooi]

How are you liking the pictures in ES mode? Does it do what you want it to do?

 

[Nrbelex]

How are you liking the pictures in ES mode? Does it do what you want it to do?

It's been raining this evening, so no shots from the real world yet. But pixel peeping seems to show a marginal difference in noise when the shadows are pushed. I expect to use ES mode in virtually all circumstances.

 

[Quarkcharmed]

I've uploaded RAW and JPEG shots taken with both the electronic and first curtain shutter, all at ISO 100. The first thing I've learned is that nothing in the EXIF data distinguishes electronic and mechanical shutter shots that otherwise have the exact same settings, which doesn't make this all that helpful a comparison. That said, I'm reasonably confident I can identify the electronic shutter shots as they seem to have more noise in the shadows when boosting the exposure. If that's right _88A0031 is electronic while _88A0030 is mechanical.

Update: If you open the RAWs in ExifTool, you can see the shutter type.

Thank you!
Any chance you could shoot a long exposure dark frame (with the lens cap on), with mechanical shutter and 30-60 seconds duration, and without the "long exposure noise reduction" on?
The noise is slightly higher than that of the R5, but I was wondering if Canon addressed the dark current/hot pixel problem. The hot pixels will be seen in a long exposure dark frame. We will also see if there is non-random noise, e.g. banding.
Dark frames at a short exposure, like 1/30s, will also be useful (both EFCS and ES).

Thanks!

 

[Viggo]

A stop worse DR? Wow, that is NOT good..
What was the big deal with a BSI-sensor if not better DR/noise control?

 

[koenkooi]

A stop worse DR? Wow, that is NOT good..
What was the big deal with a BSI-sensor if not better DR/noise control?

BSI+Stacked is only a thing when Canon doesn’t have it.

 

[AlanF]

A stop worse DR? Wow, that is NOT good..
What was the big deal with a BSI-sensor if not better DR/noise control?

The big thing about stacked BSI is better rolling shutter.

 

[AlanF]

Three big take-aways so far (all of which match my expectations going in):

1. While the image quality is certainly better than my 5D MK III, particularly at high ISOs, it's not a revolution.
2. The AF system, on the other hand, is a complete revolution. Having waited a few generations, I suddenly have things like eye detect AF and a plethora of options and modes that I'm confident will increase my keeper rate substantially.
3. IBIS is giving these old EF lenses a new lease on life. I'm able to hand-hold shots I'd never bother with previously.

P.S. My understanding is the R5 II without the cooling grip has no active fan. It's interesting then that it makes an audible sound in a quiet room, almost like a whirring, whenever powered on, even with IBIS turned off. Update: Apparently the R5 made such a sound as well, and it's very likely the IBIS unit (even though it was turned off).

The sensor on the R5 is in a different league from that of the 5Diii and is more than just "certainly better": it has 3 stops more DR at iso 100 and is better even at high iso, twice as many pixels, and a much better AA-filter so that it has far higher resolution.

 

[LSXPhotog]

Thank you - it will be interesting to see measurements in photonstophotos.

For now, DPR put the studio scene comparison images.

The dynamic range and overall high ISO performance took a hit it seems and look a bit disappointing.

Note that the second gain kicks in at ISO 400 in the R5 and at ISO 800 in the R5II. Also the R5 doesn't apply the infamous noise reduction in raw from ISO 800. So we can compare ISO 100 and ISO 800 between those cameras.

Looking at the deep shadows:

https://www.dpreview.com/reviews/im...23&x=0.03777228857777911&y=0.5007271810640153

... and comparing the EFCS vs EC in the R5II:

https://www.dpreview.com/reviews/im...23&x=0.03777228857777911&y=0.5007271810640153

we can tell the R5II has a significantly lower dynamic range at ISO 100 (say around 1 stop worse) and worse high ISO performance.

The ES may be 14 bits but it's still noticeably worse than the EFCS mode.

I read that article on DPReview and then I used my own eyes on the samples (imported them into Lightroom as well) and I walked away with a much different opinion on these results. "1 stop worse" Are we looking at the same pushed images in the link? R5 shadows shift dramatically magenta and bleeds colors on contrasting edges. The R5II shows hardly any magenta shift and no bleeding of colors over contrasting edges while preserving details better. There is more to usable dynamic range beyond read noise, it's also color preservation and detail. With the +25 color noise applied as standard import into Lightroom, I argue that it's producing a more usable image in the original R5. This also allows for very effective Enhanced Noise Reduction in Lightroom because the original color accuracy is preserved better.

R5II vs. R5 Vs. R3 vs. R6II pushed 6-stops at ISO 100 in EFC

The sensor does appear to be designed to perform very well at ISO 800 for video CLOG2/CLOG3.

I await a true dynamic range performance test. Because the current best dynamic range from any full frame sensor is the R3 (14.7 stops on DxO) and those images look worse than the R5II in these tests...when it's all said and done, you should only care about how the camera is to use and how the files are to edit for you. I greatly prefer images from my R3 and R6 Mark II over the image quality off my R5, but I prefer the resolution of the R5 for all the commercial photography I do. The R5II will likely perform similarly to the R3 sensor but do so with 45mp.

 

[neuroanatomist]

A stop worse DR? Wow, that is NOT good..
What was the big deal with a BSI-sensor if not better DR/noise control?

To repeat (response to a similar post on another thread):
Stacking does nothing for image quality, only read out speed. The benefits of BSI are essentially gone at full frame sensor sizes (they really only provide an image quality boost for sensors with pixels smaller than 2 µm).

 

[Viggo]

The big thing about stacked BSI is better rolling shutter.

But, it wasn’t that bad on the original R5, and after 4 years, the tech with BSI everyone raved about and waited for and it is even 14 bit in ES now, and STILL gets a lot worse IQ, because of the speed? Then the point kind of goes away though…

 

[Viggo]

To repeat (response to a similar post on another thread):
Stacking does nothing for image quality, only read out speed. The benefits of BSI are essentially gone at full frame sensor sizes (they really only provide an image quality boost for sensors with pixels smaller than 2 µm).

Didn’t see this before I posted and answer to AlanF, but yeah, confirms what I thought. Then I really don’t get the point with either the speed or the BSI tech. In most cases for most people the original R5 will have better IQ, how is that possible..

 

[AlanF]

But, it wasn’t that bad on the original R5, and after 4 years, the tech with BSI everyone raved about and waited for and it is even 14 bit in ES now, and STILL gets a lot worse IQ, because of the speed? Then the point kind of goes away though…

The R5 is actually very good for rolling shutter and only marginally worse than the R6, R6 II and R8 that have to transfer only half the the data. The rolling shutter can be seen when panning against vertical lines, which has occasionally affected me. I am a very happy R5 user and am not complaining about it. The faster read out should help out when tracking BIF, but again I am not complaining.

 

[koenkooi]

The R5 is actually very good for rolling shutter and only marginally worse than the R6, R6 II and R8 that have to transfer only half the the data. The rolling shutter can be seen when panning against vertical lines, which has occasionally affected me. I am a very happy R5 user and am not complaining about it. The faster read out should help out when tracking BIF, but again I am not complaining.

For me, the faster readout in the R5II will allow me to mix flash with ES.

 

[neuroanatomist]

Didn’t see this before I posted and answer to AlanF, but yeah, confirms what I thought. Then I really don’t get the point with either the speed or the BSI tech. In most cases for most people the original R5 will have better IQ, how is that possible..

I'd suggest waiting for a bit more information before concluding anything about the IQ of the R5II. Relying on information from one source is not a great idea, especially when that source has known biases.