I'm going to add my long-winded $0.02
First, this is yet more vindication for Canon that the AA filter is still a good thing for PRO photography. It is a relief that, while Canon takes a beating on the internet, they still retain some sense of practicality for real professional photography and do not completely appeal to the hordes of electronics enthusiasts across the internet who are amateur photographers at best.
The moire test images clearly shows the effects of moire in real world images. Hey, where's all the Nikonians who constantly preach that "moire doesn't happen anymore with higher megapixels" ...especially certain popular Youtube video bloggers. Really? Canon has 50MP, and moire still exists, so there goes that load of BS. When it came to Nikon and their fan base, there was widespread acceptance of the loss of the AA filter with pretty much ZERO comparisons or real discussion showing the consequences of removing it. Instead, they boasted of the extra "sharpness"...
In the Canon world, we see a more honest, professional and open debate with real examples.
I do think that the 5DSR sales will easily surpass. But that's the marketplace. And Canon is going to feed it. But we should be happy they are offering the regular 5DS.
NOW...onto the comparison.
On the test images for IQ - I think they are so close that it makes no sense to go with the 5DSR.
If there was a considerable difference, something easily noticeable - then one could argue about the added IQ on the R. But they are almost identical. The gain is practically zero. Only someone with O.C.D. might let it bother them.
They are so close in fact, that I'm willing to bet differences in LENS will have more impact on image-to-image comparisons than the sensors themselves.
In other words, you can buy two 70-200's, identical models - mount them and there is a good chance differences in the lenses will have more impact. How the AFMA is set will also matter. Shooting test charts is the best way to analyze since more variables can be controlled, but even then the slightest vibration, being 1mm off - just about anything tiny can have more of an impact. I'm not bashing their tests. I'm sure these were locked down on steady mounts and many test shots taken. Just illustrating how minimal the difference is.
In the real world, the wind blowing a little will change the moisture in the air and make a bigger difference on a landscape shot.
I don't know, maybe it's me - but the 2nd midframe image looks better for the 5DS than the 5DSR. The rest the 5DSR has a very minuscule advantage on.
Ok, all that said, for my money - the 5DS is the better choice. I get all the resolution, all the awesomeness, and I get to dodge moire in certain situations. Plus it is a little cheaper too. What is not to like?
After all, is this not supposed to be a studio and landscape camera?
If one is taking a landscape photo to blow up large, can you imagine having that ugly moire shown on the bricks in that one image? That is unacceptable. All these megapixels, all this and that goes out the window when an image is made amateur, and is ruined by ugly artifacts like that. I don't know how some people can go on and on about things like "dynamic range" and endlessly debate all this BS about how important that stuff is for their landscapes - but they do all their shooting with a non-AA filtered camera. Dynamic range to them is the end all be all. Meanwhile, they either have ugly moire in their images OR they have to photoshop it - mangling the image, OR they just have to choose a different shot entirely to avoid moire. So much for high resolution when you have to essentially damage the detail doing a very difficult fix on moire.
Or how about for studio shots of models? These can be blown up to large posters. Many fabrics cause moire.
The 5DS is the winner.
I think the real pros out there have already bought the 5DS and are happily shooting away. The huge wait lists and demand for the 5DSR is from the amateur, electronics enthusiasts out there. Well played Canon. Serving both interests.
As for the 5DSR, I think the question should be reversed. These tests show how the AA filter is an advantage. It should be shown (If it can), how the 5DSR can help in a practical way in terms of increased detail with the cancellation or no-AA.
I haven't seen it yet. And judging from the test images so far, the advantage is almost nil.
As they say, better to have and not need, then need and not have.
Particularly when you lose nothing.