Google Reveals Standalone Headsets at I/O Event
Google held its annual I/O developer event last week where it revealed that HTC and Lenovo will be building new standalone Daydream headsets that don’t require a separate phone to work. Based on the Snapdragon 835 SoC, the headsets will feature inside-out positional tracking, similar to Microsoft’s Hololens and Windows MR technology that we covered last week.
The tracking algorithm evolved from Google’s Tango project and has been christened “WorldSense”. An announcement video shows a young girl using it to sharpen her real-world dodgeball skills. Unlike Microsoft’s solution, there are no announced plans for position-tracked controllers. Instead, the devices will use the standard Daydream controller, which only senses rotation. It seems plausible that future controllers could use the same WorldSense tech to track themselves, but it may be prohibitively expensive at this point, since each controller would require a significant amount of computing power to run the algorithm.
VR Digest co-founder Ian Hamilton had a chance to try out WorldSense on behalf of Upload VR. In his hands-on he found the system to be quite impressive.
HTC’s standalone headset will be called the Vive Standalone. Pricing details have not been announced (and Lenovo hasn’t given out any details at all about their version), but Clay Bavor revealed to Steven Levy that the product cost will be “multi-hundreds of dollars”, in line with the Rift and Vive but without the cost of a separate computer. Bavor wrote an essay of his own in a Medium post, explaining the Google team’s philosophy about VR:
“...We should remember that as advanced as they may seem, today’s VR and AR devices are largely made from repurposed smartphone components. It’s like we’re building airplanes from bicycle and car parts. You can do it — it’s how the Wright brothers started — but it’s hardly where things converge.
...One day, we’ll wonder how we ever got along without computing that works like we do — computing that’s environmentally aware, that displays information to us in context, and that looks, feels, and behaves like the real world.”
More Announcements from Google I/O
Google announced their own light field rendering solution dubbed “Seurat” after the famous painter and draftsman Georges-Pierre Seurat. Unfortunately, they offered very little detail about how it actually works aside from a fourteen-second animation. A demo ginned up by ILMxLab shows a star wars docking bay being rendered with real time position tracking on a phone, showing off raytraced reflections glancing realistically off 50 million polygons as the perspective shifts. The scene takes an hour per frame to render traditionally on a workstation. No word on how long it takes to bake the lightfield.
Google claims that Seurat’s light fields can be compressed to just a few megabytes, which would seem to be a pretty amazing achievement. The tech is being marketed as a way to close the gap between PC graphics and mobile graphics, but it should be noted that there is nothing standing in the way of light field technology being used to increase fidelity in desktop based experiences.