Does anyone with scalar chip knowledge know what the difference between the 8K at 2560x1440 (upscaled) per eye and the 8K-X at ~.6667 Super Sampling(assuming it can scale this way), which is close to downscaled 2560x1440 per eye, would look like? I know the the Pimax crew said there isn’t much of a difference in general, but I was wondering what the differences could be.
1 Like
More Pixels. That’s the only difference.
But isn’t it rendering the same amount of pixels in my case?
upscale vs native. No its not the same thing.
Two things immediately spring to mind, that should differ, in favour of the X:
-
Even though you have only rendered at 1440, the barrel distortion is done at full native resolution. This means it will be more spatially correct. This includes correction for chromatic abberation.
-
Every time you resample an image, it will be a bit softened (which can be ameliorated in a number of ways, but those, too, are destructive processes). In the case of the X this only happens once, whereas with the regular it does twice.
1 Like
Isn‘t his point that he wishes to understand if there is any substantial difference between the following two scenario‘s:
-
8K: GPU renders at 2560x1440, submits the data to the 8K upscaler chip, which transforms it to 2x4K.
-
8K(X): GPU renders 2x4K, but with SS of approx. 0.667 which would suggest that the internal true rendering of the scene happens at lower resolution resembling 2560x1440, which then is just upscaled by the GPU to 2x4K to be sent to the 8K(X).
So it would appear as if in both cases the scene is only rendered at effectively 2560x1440, and in one case it is upscaled by the 8K upscaler chip while in the other case it is upscaled by the GPU. If both upscaling processes where performed at equal quality, you would expect comparable picture quality.
So his question effectively comes down to comparing the upscaling quality of the upscaler chips utilized in the 8K with the GPU upscaling.
The obvious downside of the GPU upscaling on the 8K(X) is that it increases the workload of the GPU while the 8K takes care of that part. The upside is that it creates flexibility, with powerful future GPU‘s you can increase the SS above 0.667 all the way up to 1.0 (native resolution).
3 Likes
I’m not talking native here, I’m saying the image rendered would be the same due to a SS value of .6667 on the X.
So I noticed a bit of a softened image on some of the “through the lens” videos, although I know those are hard to judge without actually using the thing. On one hand native would be better to keep the image crisp, but on the other hand, maybe the softness will alleviate some of the awful aliasing in some games. Hmm.
Yes, I guess I was asking about the upscaler chip quality on the rendered image vs a native panel/GPU “upscaling” from 0.6667 SS, or the same resolution that the upscaler goes up to.
I hope the upscaling isn’t too bad. It’s a bit frustrating that people will put down the 8K because “it’s not 8K, it’s upscaled,” when it’s the only headset with two 4K screens in it available for normal consumers. The upscaled resolution is also higher than any other headset out there for normal consumers. It’s weird that people will ignore that and put it down any way they can. I can’t wait until all headsets have good FOV and quality and we’re just buying the “name brand” that we like the most.
The chip upscaling should be pretty good. 2560x1440 to 3840x2160 is an easy case. It’s exactly 1.5 scaling.
1 Like
Please, what is SS ?
so you guys saying that we will get the 80+ hz with hdmi 2.0? from what i underastand is not 4k but upscalling to 4k because if it is 4k then from what i know max hz would be 60hz with 2.0, however being 2560x1440 might give more hz , correct me if i’m wrong. im totally lost with how that work.
Super Sampling. It’s a technique to enhance or degrade the image resolution, but then the GPU fits it to the physical resolution of the screen. If the number is >1, it helps fight aliasing and makes the image more crisp. In my case, however, I’m talking about multiplying by less than 1, so the image becomes degraded down to the same input resolution of the 8K.
1 Like
Pimax is hoping to get 90 Hz using DisplayPort 1.4, which supports that frequency at a resolution of 2560x1440. Note that NVidia cards do NOT support DSC compression, which limits the resolution below the maximum DP 1.4 data rate.
As the resolution increased (to say, 4K), the refresh frequency drops to 60 Hz or even 30 Hz for 8K. My NVidia 980Ti only supports DP 1.2, so I’ll be limited to 75 Hz refresh. NVidia 10xx cards support DP 1.4 (without DSC compression).
This is complicated, so here are some links:
3 Likes