Today at SXSW, two of the cofounders of Oculus and one of its lead hardware engineers took the stage for a quick on-the-fly Q&A session about the future of their burgeoning virtual reality headset.
If you (like, presumably, most people watching) were hoping for an update on when they might ship the retail version (or even a second developers-only version) of their virtual reality headset: no such luck*.
[* The topic actually got brought up a few times, but they actively dodged it. When an audience member asked “Do you know roughly when a new developer kit might ship?”, Palmer replied simply: “Yes.”]
They did, however, touch on a question that many have asked since the Oculus Rift started picking up steam: why now? Plenty of people have tried this virtual reality headset thing before, and it’s never taken off. Why might the Oculus Rift do any better?
(For reference, the original questioner mentioned the DisneyQuest Aladdin virtual reality ride that Disney built in the 90s — hence Palmer mentioning it specifically, below.)
Here’s what Palmer Luckey had to say:
The main issue was.. well, there were two issues [with past attempts.]
One, it was really, really expensive… All virtual reality units were tens of thousands of dollars each. The DisneyQuest stuff was super cool — but that headset cost hundreds of thousands of dollars to develop, and cost tens of thousands each. They were using custom CRTs, carbon fibre shells, these crazy spring mounts.
The other thing: quality really was not up to where it could be a consumer product. It could be a research project, easily, and you could do very cool things. People have been using virtual reality for research, but it’s only pretty recent that we’ve had mobile displays kind of… driving this revolution where we have very high pixel density displays that are a reasonable size, that are very light weight, and low cost.
Same thing with motion tracking. Back then, a lot of motion tracking was done mechanically. The VR headsets were using magnetic tracking. That actually worked okay for virtual reality… if you were standing up, and they’d isolated you from any ferrous objects [editor’s note: objects containing lots of iron and are often magnetic] and calibrated against any distortion. But it wasn’t ever good enough for a consumer solution, where it’s going to be planted in a lot of different places, in a lot of different environments. They did have some [other types of tracking] then, but it wasn’t nearly as cheap and was an order of magnitude less precise then what we have today.
The biggest part: no matter how good the VR hardware was at the time, the computing power has never really been there. Today, the average person’s PC, or even a $400 console can render… not a photorealistic environment, but a pretty good looking 3d environment. If you look at any of those old systems back then, they weren’t anything near photorealism. There were cool things you could do — you could render a wireframe, at high framerates, in 3d.. but it wasn’t anything consumers would latch on to.
Oculus hardware engineer Ryan Brown added:
A quote I like from Chris Anderson: we’re all benefitting from what he likes to call the “peace dividend of the [smartphone war]“. Everything we can package into these [head mounted displays], it’s because of the minituarization and monetization of these cell phone components.
And, as for the biggest topic touched on all night: if they can show off prototypes that do all sorts of things their current developer kits don’t (like the Crystal Cove prototype they showed at CES) what’s keeping them from shipping those?
“It takes a long time to go from ‘We have a working demo that works in one environment!’ to ‘We have something that is cheap, reliable, and would work in everyones home without any failure points.’
Some people are like ‘They got it working! They can ship it now!’
If we shipped it [right now], it would be a terrible product. It does fall apart in a lot of scenarios. We know how to fix a lot of those, and we’re working on fixing others. “
But what’s the biggest hurdle they see on the horizon? As I myself have waxed on about before, no one has really nailed input yet — that is, between keyboard/mouse, gestural interfaces, and game controllers, controlling what’s going on inside of the virtual reality world doesn’t really feel perfect yet.
Palmer: “The next big obstacle is going to be input. Input is something that hasn’t really been tackled yet in VR.”
Nate Mitchell: “Yeah, I agree on input.”
For what it’s worth, Oculus is already doing better than any past stab at a virtual reality headset. They’ve already sold upwards of 60,000 headsets — and that’s with the hardware still in a developer-only state. And, as Nate Mitchell pointed out on stage, more people have built software for the Rift in the past year than has been built for virtual reality (as a whole concept) in the past 20 years before it.