Discussion in 'Video Cards & Monitors' started by summoner, Jul 9, 2018.
*Reviews of the Valve Index have appeared in the wild*
So if you buy the Vive Pro bundle kit, did you always get the V1 lightboxes and V1 controllers? Because that’s what you get now....
they really need to do eye tracking with foveated render because we really need 8k panels at 90Hz+.
the main thing that kills it for me is gazing in to the distance.
- the resolution is terrible,
- the rendering engines do not have enough depth queues to make long distances seem real. eg. atmosphereic scattering, global illumination.
the first can be solved by hi res panels, the second by high quality renderers using foveated for efficiency
What the he'll kind of home system is going to run 8k at 90hz? Or are you talking 8k per eye which is even more ridiculous
The Valve Index (ultimately better than the pro) bundle kit comes with V2 trackers, but is backwards compatible with V1.
Any machine. Look up foveated rendering. It only renders what the eye is focused on.
There is a lot of hype for foveated rendering but there is no implementation for us to look at the pros and cons and if we are even going to see it anytime soon.
Much better to wait until we have an implementation in my opinion. I'm not a fan of calling things the solution to huge problems as most of the time it doesn't live up to the hype or the implementation is so slow that its not useful.
Seems strange as I thought the bundle had V2 gear as well.
My guess is they have a shortage of v2 sensors.
They want $310 AUD per V2 Controller separately, so it might also be cost, HTC are greedy.
of course you have to wait for an implementation. its the correct play. we need internal eye trackers though. pretty easy to do.
need lots of memory bandwidth in the gpu though, even to render a blank wall at 8k lol. but its definitely within range of high end cards now.
No, not at all.
1. The hardware has to perform perfectly, for everyone
2. The game needs special support for it
3. Once support is enabled, that doesn't mean it will be any good. There are game/engine limitations that can exist that will make the feature ineffective.
Foveated rendering becomes pointless if its noticeable and gives a bad experience. This is why we don't have foveated rendering right now.
We actually have the hardware and the tech right now in some VR demos, its just not very good right now.
What this can also mean is that even if Valve release a perfect foveated rendering solution, we might only see certain future games implement it, and its much less likely in simulators and game ports, which is where you need the performance increase most of the time.
In VR alone we have a long list of very nice VR features from Nvidia that improve performance drastically and many games don't have these features due to the engine, engine version or simply the rendering method they have used.
In Tech alone we hype up features like DX12, multicore CPU support and Vulkan, all of which have very poor implementation. Yes most games do not have proper CPU multicore support, only partial. And they have had many many years to support it.
true just because its easy doesn't mean people will implement it.
eye tracking is orders of magnitude faster already than the required speed.
so its a matter of implementation.
Just because we can capture the eye at 1000fps, it doesn't mean we can figure out what it means, as well as make changes in a game, before your eye will notice it.
The huge potential performance savings figures are assuming insanely low latency of the game reaction as well as simply not noticing the reduction in quality in your view.
I see what theyve done - if you order the $1900 consumer kit you get the V1 stuff.
If you order the $2200 commercial kit you get the V2 stuff.
Youre right, thats a bit of a gouge. If the Index is as good as it seems to be then no one in their right mind is going to pay even more money for an inferior kit. Even compared to the Rift S, is the Vive Pro really 3 times better? I think not...