Virtual reality is assuredly no longer the domain of charlatans and fools. Both John Carmack and Gabe Newell have heralded the Oculus Rift as a triumphant second coming of this much-maligned technology, and its 20-year-old creator, Palmer Luckey, has quickly ascended to the status of industry maven. $2,437,429 of Kickstarter cash suggest that others approve, too. We caught up with the talented Mr Luckey to talk about the manufacturing hiccups for his VR dev-kits, his dealings with Valve and other devs, and where the technology will head next.
So how'd you start in this?
Modding consoles into mobiles. You remember Benjamin Heckendorn ? He ran our forum, but he became an absentee administrator. All those mods he was making, he never posted them on his own forum. He would go straight to the news and bypass his community. Then he went, “I'm not modding any more, I'm going to get a real job.” So people got pissed at him about that. Then, somehow, me, this incredibly lucky 15 year old kid said, “Well, I'm going to make my own forum,” which has never worked in the history of ever, but this time it did - and we had 900 members in the first week and a couple of thousand after the first few months, and then we destroyed that community, because everyone left.
That's what I did before this, I was always a console modder. As for this particular design, I have the world's largest head-mounted display collection, 46 units now. I've taken almost every single one apart, learned everything about it, and put it back together. So it was a combination of trial and error and copying what I knew worked.
Do you have time for anything else? Do you ever watch TV?
I used to watch G4TV, before it became terrible, which it has been for quite some time now. I watch Survivor because my whole family likes it, even though I know it's all staged and cut. It's a guilty pleasure.
How about the Chinese side? How do you actually find people to make your dream device?
That wasn't even my doing. When I was getting this running, I was in a position to work with Jack McCauley (opens in new tab) ; he's our lead engineer and he was also lead engineer on Guitar Hero. He designed and manufactured all of the Guitar Hero controllers and peripherals and everything. He's also worked with other peripheral. He's the guy who's managed to get this made, at such a high quality, in China.
I wouldn't have gone to China. That would have been a comedic adventure, 'Palmer Goes To China.' It's not even as simple as getting a contact, it's doing it in a way so that you don't get ripped off. You can look at it and say, “We can make this part way cheaper,” but you have to be able to speak their language and work with their engineers. And that's Jack.
Wow. He's essential.
Yep, he's very essential. He was working for Activision, then he quit and started his own business, consulting. And he stopped consulting when we hired him.(opens in new tab)
You had some delays recently. Why was that?
Yeah, that was unfortunate. Two things. People are saying, “You must have known about these delays, you must have known that there was no way you could have made December." It's simply not true. We had them designed and ready, and we could have started manufacturing then shipping in December. The problem was the screens we were using, I had quotes for 5000 units, which was no problem. The problem was as soon as our Kickstarter ended we had almost double that - we had 9000 backers - so I asked for 10,000 units. Then they said, “No, we can't do that. This is an end-of-life product and we only have about 5,000 panels left.” Which they didn't tell us. And we never checked because we never thought we'd hit 5,000, it seemed the upper limit of what we might get. I mean that's over a million dollars, which no Kickstarters got.
We needed a screen, so we had to find another screen, which meant changing the entire design; we had to change the control ports, change the lenses, basically the whole thing. We were rushing because Jack knew this factory really well, that they could run with our aggressive schedule and still hit our December ship date. The fact that it was already going into manufacture, this was around Thanksgiving, meant we could have said, “Yeah, go ahead, get shipping them.” We would have made it. The problem was then with minor issues with the manufacturing. They made some samples for us, final ones coming off their machinery, which meant we had to pull the plug and go for another round of revision, and the only way to do that was delay. It would have been much worse to ship on time a product with glaring obvious flaws that we knew about.
Yeah, developers wouldn't work on dodgy hardware; though Notch has said that he's going to work on it for 0x10c.
Yeah, even if he has to write the Java himself! We have tons of people on board, a very good relationship with Valve for example.
Did you know about the Steam box before it was announced?
I can't say. But I'm really glad that it's been announced.
They weren't building a Steam box 12 months ago when we asked them.
They're doing a lot of things that they're not saying that they're doing. There's a lot of cool stuff they're doing there. We're working with them, Epic, hopefully Project Cars, and Hawken. Overall we've got a ton developer support.(opens in new tab)
You were saying in the conference that you want games to be made with the Rift in mind?
Not just with the Rift in mind, but with virtual reality in mind. These aren't things I'm telling you for our device. It's true, any VR rig needs a high frame rate, you need it to be 3D, and you need to not take camera control away from the user. Games that are made with the Rift in mind are also going to work with other headsets.
I saw Sony's Mick Hocking give a talk about what to do for 3D, along similar lines.
Yes, it's not just for Sony games or the PS3: this is what you do for all 3D games, period, and that's true. The best games will be developed with VR in mind, but there will also be a lot of ports. Games that are designed for traditional mediums, but they just add head-tracking and they're away.
You also talked about how you'd like to have games hardware built into the headset.
It's not even about having an established user base. I figure if I built the best headset possible, people are going to want it. Maybe that's not true...
I'm not a huge fan of Steve Jobs, but that's what he did. He didn't go round asking everyone, “Well, what do you want?” He built the best possible device that he could and said, “This is why it's the best and you're going to buy it or not.” He did pretty well with that. I don't think you do that with VR, as it's a much different community. But it's the same that if you make the best thing possible and show them, and they believe you...
How many people are there at Oculus now?
Sixteen full-time employees and a couple of part-timers, and we're making more hires. Probably twenty by the end of the year. Almost all technical – even our CEO is a software engineer and computer programmer. Our office manager is the only non-technical person. That's a good way to be for a small company. You don't want to clog everything up with project managers and middle managers. It's just everyone working on the same thing.
Michael Abrash made comments saying that even 1920 by 1080 isn't enough pixel density to do VR at a 90 degree viewing angle. Do you agree?
I'm not going to necessarily disagree with Abrash. He's right that resolution is important for 3D and good VR. One thing that could have been pushed more in that blogpost is that when you perceive 3D there are a lot of different factors; where are your eyes are focussing, where are they converging, how your eye turns dual edge data into very precise lines. With a low resolution, you don't get that much data off the edge, so you don't get as good 3D. The example he used was that having a virtual spaceship appears very real in a 1080P field-of-view headset and not so real in a wider field. They're working on putting virtual objects into the real environment; that's very different from the full virtual reality Oculus is going for.
How does it work with things like motion blur and depth of field?
Motion blur is an effect to try to simulate what we see in real life, so really if you turn your head very fast, your brain can't process the image moving that quickly anyway. Weaving down a corridor is really bad on this [the old Oculus display panel]. If you read the Kickstarter update, one of the changes with the new panel is that it has a much higher contrast ratio and a much faster switching time; almost twice as fast, so you have much less blur in the new panel.
And depth of field, which is an artificial effect as well?
Depth of field works better in a small field of view than a large field of view. In a large field of view, things out here aren't in focus because you're not looking at them in VR. It doesn't make sense to waste time on depth of field effects when things are like that.
I do think resolution needs to be better for the best VR, but even at low resolution you can get a good experience. I'm not going to say that this is the best experience, or that resolution isn't important; it's a really critical factor and this is good as it gets today. We're going to be building one at double this resolution within a year and even more after that, so there's a lot of other things to work on besides resolution. We can't just sit around and wait for 4K displays to hop around
Thank you, Palmer Luckey.