I understand and I also noticed they chose very favorable photos of the game(it looks almosttttttt identical) whilst the shots you showed definitely show some differences. Time will tell really. I'd love for it to be super close to a PC running ultra settings but I do acknowledge that the checkerboard rendering will create some artifacts that native 2160P does not contain. But hell even on 4K games on PC we still see flaws, aliasing, tearing etc lol.

Can you specify what is annoying?
This thread is full of people who have a hard time distinguishing between checkerboard and true 4k.
VR foveated rendering is talking about discarding 90% of info outside of narrow circle,
but the concept here is retaining the full 50% of checkerboarded info in that area.

If anything, some level of subsconscious awareness of difference in detail may re-enforce feeling of immersion,
by re-inforcing transfer of POV whereby you naturally use POV swivelling whenever possible when glancing around.
Obviously, there is no need for "sharp" borders of these areas, it is choosing specific pixels to render (vs. interpolate),
so the "full render zone" graduates outward (towards full 50% checkerboard, whose full-render pixels alternate each frame)

Anyhow, I'd be interested on other takes on what else could be "between" pure 50% checkerboard and true 4K,
since there is no longer a simple step like 720p->900p->1080p in standard upscaling.  
I guess with the color subsampling idea, color is checkerboard rendered, while brightness is rendered true 4k = 2/3 load?

really would like to see this because Im not impressed buy the up scaling of the Xbox one s

Chroma subsampling, or 4:2:0 downsampling only helps with compression. No point to that process for rendering a game. Same with pentile displays not giving any advantage to rendering even though color information is reduced to half.

Foveated rendering could certainly work in racing games, and for any portion of the screen that is going to be motion blurred or receive dof anyway. In racing games most of what you're flying by is heavily blurred, only the competing cars and distant track need to be fully rendered. In FPS it might be easier to spot, yet even so, human sharp vision is very narrow, only a few words wide in a sentence at a time. Assuming you're looking near the center 99% of the time is not a bad assumption.

Color is more visible away from the center, yet reducing color resolution just doesn't help rendering.

Futurproofing for what? A console should be relevant for 5/6 years. That's the norm. Thing is, if the Scorpio spends 2 years not selling because it's too expensive what's the point?

Especially when they will release a new system every 3 years or so... Then after the scorpio what happens? They have two choices, release a new model that will improve graphics as much but will again be way too expensive for the general public. Or make the smart decision of releasing a 400$ console that will basically be what the pro is to the PS4.

Anyway, besides the lack of 4K bluray, the PS4 pro is already " furturproofed" enough.

If your fov on a 4K tv is 20 degrees you won't notice the difference between 1080p and 4K anyway. 20/20 vision allows for about 60 pixels per degree, average adults can make out about 80 pixels per degree and perhaps you still feel a difference at 120 pixels per degree, yet 192 is really pushing it (4K at 20)
http://phrogz.net/tmp/ScreenDens2In.html

Anyway since full screen motion blur and dof are so popular nowadays, kinda waste to render all that in full detail to blur it again straight after.

Why shouldn't there be a simple step like 720p->900p->1080p in standard upscaling? If 900p is an option for the 8th gen and 640p was an option for 7th gen, anything between 1920x1080 and 3840x2160 with up- or downscaling is also an option.

The Nvidia driver f.e. offers DSR-factors of 2.00 (2715x1527 on a 1920x1080 display), 2.25, 3.00 and 4.00 (3840x2160 on a 1920x1080 display) to improve picture quality and reduce aliasing. Or a game could render in 2560x1440... with a display in that resolution connected to the PS4P it would use native resolution, otherwise it would downscale on an 1080p display (with DSR profits) or upscale on 4K-displays.

I'm not saying that these are better solutions that the checkerbox mode of the PS4P, but the option is there and it could look better or worse depending on the game and the art style.

Yeah racing was one of things in my mind, where you're much less likely to dwell your vision on stuff outside the center. 
Of course this is talking about a larger area than VR foveated vision, and less drop off outside the area.
I guess the other approach to an "in between" is simply randomly adding more full-render pixels thru-out the entire screen.
It does seem like for 1st person views, that focusing those towards the center will yield the most perceptible improvement.
I guess I had thought renderers could natively output YUV/HSV/etc imagery (logically using textures etc in those formats as well), but perhaps not?

He's right though. That can work dor VR because of how close the screen is to your eyes. What part of the image falls into what would have been your peripheral vision is never going to be in focus unless you deliberately roll your eyes all the way over to one side or the other. But usually you would b focused on the centre of the screen. 

On a TV however, even if you are looking at specific things the entire screen is within view and they wouldn't have any meaningful way to track your eye movement in relation to where exactly on the screen you are focused on. So where in VT they could track your eyes and give the part if the screen they are focused on Full rez, on a TV the best they could do is track your cursor... so what happens if you decide to focus your eyes on the edge of the screen without moving the cursor?

As for the color depth thing, the amount of processing required to go from a 8-10bit color depth to support HDR isnt that much. The problem has just been that TVs didn't support it all this time as the level of brightness a TV would need to make those differences noticeable werent achievable at the time. it's the reason why one of the key specifications of HDR support is a display that can go up to 1000nits of brightness and at least 400-500nits of 100% sustained brightness. 

And as for what's next between 50% 4kc and 4kn.... my guess would be anything that renders a larger number of real pixels per frame. As it stands the PS4pro renders 4M of the 8M pixels that makes up a native 4k image then using the checkerboard tech reconstructs the other 4M. If you could render 6M Pixels (75% 4kc) in a checkerboard pattern then that leaves you with only 2M pixels to reconstruct. The end product would be an image better than 50% 4kc.

This could be something that the scorpio could do with ease if it's psecs holds true. Though I have no idea exactly what sony's patents would stop them from doing. My guess is that it only stops them from making the same kinda customizations to their APU. But they should still be able to do it in software. 

This is nothing like the XB1s upscaling. That's a straight upscale where one pixel is basically attached to occupy 4 pixels then sharpened artificially amongst other things to improve sharpness and mitigate ghosting...etc. 

Yeah, obviously you can upscale between arbitrary resolutions,
which will have the characteristic of standard upscaling with smaller discrepancy in resolutions.
I was focused on taking advantages of the paradigm of checkerboard rendering which has a different characteristic visual effect than upscaling, and what intermediate steps or approaches might there be before full 4K?  
As far as I can tell, one can increase the percentage of full rendered pixels, e.g. 50%->60%->80%->100%, and the question is does one do so randomly, or focused in a certain way?  
Screen-center-bias seems like it would yield higher perceptual benefits for strongly screen-center focused games, e.g. 1st person POV racing, shooting, etc.  In other type of games, you would probably be just as well off equally distributing the "full pixels" across the whole screen, although I suppose one could exclude/include specific areas of interest or disinterest, e.g. areas adjacent to/ overlapping HUD elements, foreground weapons swinging around in lower middle zone etc.