5DsR image quality

[Frodo]

There have recently been some strongly held opinions on image quality of the 5DsR presented in this forum. Some of these opinions have stated that:
- There is little difference in image quality between the 5DsR and other 5D cameras
- The 5DsR has lots of noise at high ISOs.

As a relatively recent owner of a 5DsR in addition to my 6D (and I had owned a 5DII before then), this did not match my perceptions.

So in order to do a fair comparison of detail at ISO 100 (which is where I shoot landscapes) and noise at ISO 3200 (commonly used at events), I:
- Downloaded Raw files of the 5DIII, 5DIV, 5DsR and 6D of DPreview's "studio test scene"
- I processed these in LR 6 with no adjustments
- I cropped the ISO 100 files at 1800 pixels on the 5DsR files and copied the crops onto the other ISO 100 files and exported at 1800 pixels along the long edge. This would have upscaled all the files except 5DsR
- I cropped the ISO 3200 files at 1800 pixels on the 5DIII files and copied the crops onto the other ISO 3200 files and exported at 1800 pixels along the long edge. This would have downscaled the 5DsR file, the 5DIV somewhat, and slightly upscaled the 6D file.

These are attached. In case the file names don't show they are 4 x ISO 100 files: 5DIII, 5DIV, 5DsR, and 6D, then the ISO files in the same order.

IMHO:
- The ISO 100 5DsR files are noticeably sharper - look at the bristles of the right hand brush and the fine text on the paint tubes.
- The 5DsR is slightly pinker, the 5DIV is slightly yellower, but I note that the lighting (as shown by reflections) differs between the images)
- In the ISO 3200 images, the 5DIII has the most noise, the 5DIV the least. The 5DsR has more noise than the 5DIV, but the grain is very fine and it did not bother me. There was some false colour in the 5DIII image. The pink colour cast in the 5DsR image bothered me more than the noise.

Anyway, this comparison focused on the two extremes of my work. I did not compare images at extremely high ISOs and did not compare dynamic range. If they wish, someone else can do that.

This confirmed my view, that the 5DsR produces excellent image quality in terms of detail and noise.

 

[Sporgon]

Downsampling the 5Ds / sr is the key to a level playing field, as you have done. Because the 5DsR has been reduced the most the image will be sharper through this alone.

Similar situation with noise at most ISOs; at the same output size the 5Ds isn't that bad.

Theoretically the higher resolution of the 5Ds should help to overcome the colour definition 'limitations' of the Bayer array sensor, and by down-sampling to a reasonably manageable file size - such as the 5,700 long-side px on the 5Diii you should still benefit from colour definition and tonal graduation. I never open my 5Ds RAW files at full size unless the final output warrants it, which is rare, especially for panoramics.

However I have to say I'm not convinced by all the mp hype and I don't think that the difference at output sizes other than those that are much larger than native for other, smaller mp cameras, is that much different. I remember when the Nikon D800 and Sony A7r came out, and some people on the web were saying how important the extra resolution was in even modest, A3 size outputs. I disagreed then and still do now. The fact is, 12 mp is still a lot.

 

[Frodo]

Hi Sporgon, I agree with your comments. Valid point about MP differences at modest print sizes. I bought my 5DsR with the expectation of using mRaw most of the time and full Raw when necessary. I am very disappointed with the malleability of mRaw files, especially when lifting shadows. Don't know if the finger should be pointed at Canon or Adobe, but mRaw is effectively unusable.

 

[deleteme]

As I am also a new owner of the 5DsR I have certainly heard the naysayers on the subject of IQ.

My experience was that the resolution advantage is slight at almost all presentation sizes both in print and absolutely on the web.
Color is not an issue as I make custom camera profiles though I notice that LR offers decent ones for this camera.

What has made a difference is that well exposed, sharp images allow dramatic crops that leave a lot of file size left to use as needed.
As my images are also enlarge to 8+ feet wide for architects offices, the increased detail is welcome but even then not a mind blowing event.

In short, the MP bump afforded by the 5DsR is similar to the bump from 12MP to 22MP. Noticeable in PS and LR at magnified levels but not really game changing for most people or purposes.

 

[aceflibble]

I am very disappointed with the malleability of mRaw files, especially when lifting shadows. Don't know if the finger should be pointed at Canon or Adobe

Canon. 'M' raw is still taking the full readout of the sensor, but instead of saving it directly 1:1 (as with 'full' raw), the processor is scaling the image down. When it does this it is throwing away data as it merges information downwards; less data means less to play with in additional post processing. It's not just the obvious reduction in resolution, either. As data is merged down you'll get more limited colour reproduction and colour shifts, on top of the potential for more colour noise. (Though luminance noise typically does reduce.) Adobe (or any other post-processing software) can't do anything about it 'cause the camera itself already chose which data to keep and which to get rid of.

For all intents and purposes, you can think of it like using .jpg instead of raw, in that you're telling the camera to save space by reducing file complexity; as a result you're not going to get the full quality the camera can produce.
Canon can't really be 'blamed' for this, since expecting the in-camera CPU to process the file as well as you can on a desktop computer with dedicated software just isn't a reasonable expectation. If you want the speed and convenience of having the less powerful, simpler, in-camera CPU handle the downsampling for you then you're going to pay the price in image quality.


As a general rule, higher pixel counts only matter if you are either making the full use of the files at 1:1 (let alone upsizing larger) or if you're reducing the files in post production, where dedicated image processing software (such as Lightroom, Capture One, Photoshop, etc) can perform smarter file resizing and compression. The 5DS R's image quality at 50mp is fantastic by 35mm standards (of course larger-sensor cameras with similar or higher pixel counts do generally beat it in most regards) but if you're not actually using all 50,000,000 pixels then that image quality isn't going to be apparent. If the 'm' resolution is actually enough for you in terms of print/viewing size then something like the 5D4 or Sony α7R may be more suitable; using a lower-resolution sensor fully is going to give you better image quality and processing options than having a higher-resolution camera clumsily scaling itself down. (Using the higher-resolution camera at its full potential and then scaling down the files yourself will continue to be better, of course.)


It's the same deal as the Sony α7R vs α7S cameras, and in many ways it's the same as modern APS-C bodies vs older 35mm bodies. A lot of people still use the original 5D because if 12mp is enough for you and you want results right out of the camera, the 5D still equals or beats most current 24mp+ APS-C bodies; only through using the full resolution of newer cameras fully (i.e. don't let the camera do the resizing for you) do the newer sensors beat the older ones.

 

[AlanF]

It's not just the extra megapixels, it's the absence of the low-pass filter as well. When I want the highest resolution, I use my 5DSR. When I want better AF for action or for birds in flight, I use my 5DIV.

 

[Zeidora]

I've seen the mRAW comment previously and I don't get it. Why buy a 5DsR and not use it for its intended purpose??? Boggles my mind. It's like complaining that a sledgehammer is difficult to use for picture hanging.
Get a 5DIV and be done with it.

 

[Valvebounce]

Hi Zeidora.
Thanks for that, today's laugh out loud.

Cheers, Graham.

It's like complaining that a sledgehammer is difficult to use for picture hanging.

 

[Frodo]

I've seen the mRAW comment previously and I don't get it. Why buy a 5DsR and not use it for its intended purpose??? Boggles my mind. It's like complaining that a sledgehammer is difficult to use for picture hanging.
Get a 5DIV and be done with it.

My Italian motorcycle can ride at 130 mph, but most of the time I'm happy with 70mph. Perhaps I should have bought a 125cc Honda that is more fuel efficient.

I do not need 50MP all the time. Usually 20MP FF is fine. But I bought the camera for the times when I (a) want to produce large prints or (b) crop heavily (like having a built-in APS-C sensor). This is what I believe "the intended purpose" of the 5DsR is.

Or are you suggesting that Canon requires me to get both a 5DIV and a 5DsR?
In fact I already have a 6D and 5DsR. Today I am going on an overnight hike and will only take one DSLR with me. I would like to have the option as to what resolution my Raw file should be. Is it unreasonable for me to expect that the mRaw provided by Canon is actually usable?

 

[Sporgon]

I would like to have the option as to what resolution my Raw file should be. Is it unreasonable for me to expect that the mRaw provided by Canon is actually usable?

Have you tried converting the mraw in Canon’s own DPP ?

 

[Frodo]

Have you tried converting the mraw in Canon’s own DPP ?

Hi Sporgon

I have tried this. This avoided the awful green cast in pulled shadows, but I had less dynamic range (i.e. ability to pull shadows) compared to LR operating on full Raw files. I think I posted the result of this test earlier.
Going back into DPP made me realise how much I have become accustomed to the LR workflow (and tools). So one option would be to create tiffs in DPP and then import into LR. But this extra step is a hassle and creates extra tiffs that would be close to the same size as full size Raw files.

I am becoming used to capturing and processing large files. Storage is not an issue, given the cost of hard drives, although backups take a little longer.

Interestingly, my 6D has had little use since I got the 5DsR. It is a necessary backup for when I shoot events, but I prefer the 5DsR as my "serious" camera and my M3 as my "casual" camera.

 

[bholliman]

There have recently been some strongly held opinions on image quality of the 5DsR presented in this forum. Some of these opinions have stated that:
- There is little difference in image quality between the 5DsR and other 5D cameras
- The 5DsR has lots of noise at high ISOs.

As a relatively recent owner of a 5DsR in addition to my 6D (and I had owned a 5DII before then), this did not match my perceptions.

...

This confirmed my view, that the 5DsR produces excellent image quality in terms of detail and noise.

Thanks for posting! The opinions expressed above are coming from people who have never used the camera or worked with its files. I sold and gifted my 5D MkIII and 6D after acquiring a 5DsR since I found the 5DsR images to be better for almost all uses.

 

[aceflibble]

I do not need 50MP all the time. Usually 20MP FF is fine. But I bought the camera for the times when I (a) want to produce large prints or (b) crop heavily (like having a built-in APS-C sensor). This is what I believe "the intended purpose" of the 5DsR is.

Well what you "believe" is categorically wrong. Canon have never marketed the 5DS/R in the fashion you suggest. If you ever go to a trade show you'll find Canon routinely refer to the 5DS/R as being a competitor to medium format cameras in conversation. They intend for it to be used at the full resolution, by people who require that full resolution. They have always expected, and continue to expect, the 'average' user of the 5-series to use the 5D4 if they do not need the full 50mp resolution of the 5DS/R. This is how they have always talked about and promoted the camera.


Far too many people blindly think that how they expect to use every camera is how every camera is expected to be used by everybody.
Canon intend, design, and market the 1D series for people who need durability.
The 5DS/R is for people who need resolution. (Just as the old 1D/1Ds divide used to be.)
The standard 5D series is your all-rounder intended for the majority of working professionals.
The 6D is primarily just for mitigating the used sales of older 5D series, which is mostly made up of the richer end of the enthusiast market.
The 7D is for professionals who need speed but not necessarily the heft of the 1D.
The XXD are for the majority of serious enthusiasts.
XXXD are for your standard user.
XXXXD are for your absolute beginners.

That's how Canon reps talk about bodies whenever you talk shop with them, and even when they're speaking to the public.

And yes, they do expect professionals to use multiple bodies. A regular 5D and then either a 5DS or 1D are what they expect most pros to be using, other than sports & wildlife photographers where they expect two 7Ds or a 7D and a 1D.

Bear in mind that, thanks to internet 'power creep', most people have a different view of what is 'pro' and 'enthusiast' compared to the manufacturers themselves. Canon don't think L lenses are for non-pros, either, yet look at how many people insist on using 24-70 and 70-200 f/2.8 L zooms on a 5D4 just to capture their kids' sports games.

If you've come up with a different use for a product than the one Canon intended when designing and marketing it, and it works out for you, great. Good for you. Keep doing it. But don't assume that how you use something is how the manufacturer expected it to be used, or that why you bought something is the reason they thought you'd be buying it. There is every chance that why any particular person bought any given product, and how they use it, is not what the manufacturer expressly intended.

 

[Frodo]

"Categorically wrong" are strong words...

I agree that is how Canon marketed the camera. As most advertisers would, you stress the points of difference, especially in relation to the previous/existing model. But I believe that the purpose of the camera was broader. Otherwise why would Canon provide an mRaw setting (and even a green square auto setting!). Simple additions that broaden the use of the camera. The 7D mk II is marketed as "capturing the action faster", but not everyone wants to shoot 10 fps so Canon also provides a lower frame rate.

But that's not my point. I started this thread to demonstrate the details at low ISO settings and low noise at high ISO (3200) settings. I maligned the inability to push shadows in LR. That simple function would increase the utility of the camera, whether or not that was the purpose of the camera.

 

[Sporgon]

I do not need 50MP all the time. Usually 20MP FF is fine. But I bought the camera for the times when I (a) want to produce large prints or (b) crop heavily (like having a built-in APS-C sensor). This is what I believe "the intended purpose" of the 5DsR is.

Well what you "believe" is categorically wrong. Canon have never marketed the 5DS/R in the fashion you suggest.

I believe that you believing that Frodo's beliefs are categorically wrong is quite unbelievable given that the 5Ds has a 1.6 * built-in crop mode

And also has two smaller RAW options......

And some pretty small jpegs....... ;D

 

[KeithBreazeal]

It's really not a good action camera...
except when you get rivets using the old 100-400L

Thunderbirds Diamond LR screenshot Minden-Tahoe 2015 2685 © Keith Breazeal by Keith Breazeal, on Flickr

 

[Maiaibing]

For raw picture quality the 5DS/R cannot be beat by any other Canon DSLR.

To me the 5DS/R and a Dx camera is the perfect match. With the 5DIV being a good compromise between the two if you want a little more fps and the live view/video options it delivers over the 5DS/R and can live with the relatively paltry 7 fps compared to a fast shooting Dx.

Since I prefer maximum IQ and very often need the ability to crop the 5DS/R gets almost all my time.