Eye Tracking, Foveated Rendering, And SMI’s Quest For VR Domination

Status
Not open for further replies.
No way to show the difference? Why not use secondary display, one person using VR and other would (should) see the uality difference because they can focus on parts of scenes unlimited by tracker.

But anyway, good stuff, should help with performance, but also has other uses as already explained in the article..
 
This is one of those *absolutely mandatory* technologies for VR to be acceptable on modern or near future GPU hardware, with acceptable screen resolutions(8k+ per eye) and acceptable frame rates(min 90 per eye).

However, if someone has a breakthrough in GPUs that increases their processing power by about 50-100x(enough to render dual 16k screens at full quality at >90hz rates) then you don't need this technology. Because thats what its going to take to render full field of view for each eye without a screen door effect, and without nausea at modern polygon/effect counts. It will take more then that if we ever want real world quality images at those resolutions and frame rates.
 
Exciting news. You mention the hardware could be added to almost any HMD, but isn't there a need for a software component? Will each game developer need to modify existing code for this device, or is there a driver that would compensate no matter the content?
 
This is pretty much like comparing gaming with or without V-Sync in terms of GPU loading (lets ignore screen tearing, just look at the performance metrics). Most game titles that already run very high framerates on a high end GPU will be able to reduce your power and heat output by turning on V-Sync. This seems like it'll do the same thing, but as an added bonus, boost framerates due to reduced rendering on the peripheral vision. Ingenious!
 
Exciting news. You mention the hardware could be added to almost any HMD, but isn't there a need for a software component? Will each game developer need to modify existing code for this device, or is there a driver that would compensate no matter the content?

Developers would certainly need to utilize this.

And this, for example, is what I've been arguing Oculus and HTC should have focused on for the first iterations of their HMDs instead of cameras, controllers or headphones.
 
can we upgrade the CV models of Rift or Vive with this technology?

Well, you can't really buy this as a standalone product, and you probably will never be able to. SMI wants to show what it can do so that HMD makers will implement it in v.2 of their products.

But theoretically, HMD makers could add this to their respective HDMs fairly easily, according to SMI.
 
Exciting news. You mention the hardware could be added to almost any HMD, but isn't there a need for a software component? Will each game developer need to modify existing code for this device, or is there a driver that would compensate no matter the content?

Yeah, great question. I've been bugging SMI for more on that. 😀 They aren't saying too much at this point, but from what I've gathered, yes, the software is a big deal, and it's actually their core competency. They will have to work with some bigger players to implement it all, but it sounds like it shouldn't be especially difficult.
 
Status
Not open for further replies.