{"id":85209,"date":"2016-01-12T13:30:18","date_gmt":"2016-01-12T11:30:18","guid":{"rendered":"http:\/\/mygaming.co.za\/news\/?p=85209"},"modified":"2016-01-12T13:43:13","modified_gmt":"2016-01-12T11:43:13","slug":"foveated-rendering-may-drastically-reduce-the-minimum-requirements-for-vr-headsets","status":"publish","type":"post","link":"https:\/\/mygaming.co.za\/news\/hardware\/85209-foveated-rendering-may-drastically-reduce-the-minimum-requirements-for-vr-headsets","title":{"rendered":"Foveated rendering may drastically reduce the minimum requirements for VR headsets"},"content":{"rendered":"

Virtual reality, when done right, certainly isn\u2019t cheap, requiring some of the best display technology currently available, among a number of other things.<\/p>\n

That\u2019s why virtual reality (VR) headsets like the Oculus Rift will set you back a substantial $599<\/a>, or R9, 500, and that\u2019s before tax and shipping.<\/p>\n

It\u2019s expensive, sure, but cutting edge technology almost always is, so many are willing to front the cost for VR \u2013 then reality hits.<\/p>\n

The fancy displays, high refresh rates and latency reducing technology used to power this new VR age will only play nice with equally expensive PC hardware. SensorMotoricInstruments (SMI) argues otherwise.<\/p>\n

\"Foveated<\/a><\/p>\n

SMI has developed a new rendering technique, dubbed \u201cfoveated rendering\u201d, that promises to reduce the graphical performance required to render to two high-resolution displays.<\/p>\n

Foveated rendering uses a similar methodology to occlusion culling, otherwise known as z-culling<\/a>. Z-culling is a method that increases performance of an application by calculating whether you, the observer, can see a pixel or not. If you can\u2019t see it, usually because it\u2019s behind another opaque object, is there a point in rendering it?<\/p>\n

By eliminating unnecessary pixels, you save resources and reduce the load on the hardware \u2013 foveated rendering does something similar, sort of.<\/p>\n

SMI figured that the focal points of our eyes are a lot smaller than the entire image rendered by any VR device, so why not lower the resolution of the parts of the image you\u2019re not looking direct at?<\/p>\n

The image above is an example of how foveated rendering works. It reduces the resolution of the image immediately outside of the observer\u2019s focal point by 40%, and the area outside of that by 20%.<\/p>\n

In doing so, using eye tracking to determine what the observer is looking at, significant resources can be saved, allowing for less powerful hardware to run the same applications you might originally have required a beefier GPU for.<\/p>\n