Oculus and Alienware cofounders geek out about taking VR to the masses

Azor: What’s going to change—Even the way we speak. We’re not going to say that we took a picture. Maybe it’s a scene that I captured. What else is going to emerge from this? It’s fascinating. We won’t just look at a scene from a distance. This will allow people to be in this place seeing us talk. It’s not just a memory. It’s an entire experience.

A lot of times we try to re-create something by editing a video as well as we can, or even making the game or the 3D experience. But we can put someone in this room right now. We can capture it in its full form. You guys have a perspective right now that’s completely different from Palmer’s and mine. Fast forward three years and this scene will be completely different. The cameras and the videos coming out of it will be completely different.

Luckey: There have been people working on this specifically for football. They’re experimenting with capturing playing fields from all around the field so you can generate a realtime synthetic view. You can move a virtual camera around inside. The cameras are going to look very different. It’s going to be captured from a bunch of angles, from places we don’t notice, and you’ll be able to place virtual cameras and virtual viewpoints anywhere inside the field.

Question: Have you committed how long you’ll keep each min spec locked in? Is there a set time you’ll keep that, so when consumers buy certain hardware they’ll have a length of time that will be the best experience?

Azor: The min spec is a result of the processing requirements that the experience has. Today that requires a certain set of hardware. Fortunately, that hardware gets faster at a very rapid rate in our industry. We’ve seen that for a long time. I expect that min spec to get more affordable as every generation of new parts comes out. We’re in a fortunate place where that happens every nine to 12 months. I expect specs are going to come down as lower price points meet performance requirements.

Luckey: The other thing you can look at is, the experience you have with the Rift and a recommended spec PC is the bare bones requirements for running a headset that can deliver that oft-used buzzword, “presence.” If you have something that can deliver presence – 90 frames per second at a high enough resolution — and not fail, that’s not going to change in the future. Even if you go five years out, I don’t think that minimum specification will change. That will be the minimum bar for quality virtual reality.

The cost for that min spec is coming down. Eventually people won’t be buying special PCs for VR. They’ll buy PCs that simply have that power integrated into them. That’s going to be huge, when PCs are being built to the power where they’re able to meet this specification without having to be a specialized VR PC.

Azor: Our design considerations now have VR in every single one of them. In the future, hopefully every Alienware system, and every Dell XPS system eventually as well, will be an Oculus certified experience.

Luckey: The more they’re optimized for VR, the fewer hours it takes to make it work well for VR, as opposed to just building it for gaming or word processing or whatever a computer is for. The deeper the optimization goes, the lower the price can be.

Azor: Direct connections to the graphics hardware are one issue. There are others that are going to be dealt with.

Luckey: The funny thing is, that layout isn’t necessarily bad for traditional gaming. Especially for a laptop, where most of the work is taking place on the integrated display. That’s actually a great architecture. It’s not a flawed architecture. The problem is when you start to stack VR on top of something that was never built for VR in the first place. By building specifically for VR, or at least keeping it in mind, you make sure you don’t end up in scenarios where you build architectures that have critical flaws for VR.

Question: I wonder whether the 970 min spec is too low. I imagine a 980 Ti would really be optimal for VR. Are we aiming too low with the 970?

Luckey: There’s a reason we call the 970 a recommended spec and not the minimum spec. Right now I play a lot of games on a 980 or higher because they’re internal builds that haven’t been optimized yet. But when I’m playing final builds, I use a 970 rig, because that’s what the majority of people will be using. Once you go into the 980 world, it’s rarefied air. Not many people are up there.

We’ve been working on virtual reality for a few years now. Imagine going back two or three years and saying that the 970 isn’t going to be enough for VR. It’d blow your mind. There will always be cards coming in the future, but at some point you have to say, “What’s something I can make a good experience with?” You’re not going to be able to make every game in the world that you can imagine work on a 970. But there’ve always been limits on PC games. How many objects can you render? How many textures can you stream? A 970 is going to allow most games to look good because developers are going to be optimizing around it.

If a developer says they want to make VR games with 5000 spaceships around this giant Death Star, you’re not going to necessarily do that on a 970 without cutting graphical fidelity to match. A 970 is not the minimum requirement. It’s plenty. It will render a lot of great-looking games.

Azor: The way we’ve played games up to now, one of the principal ways we’ve improved immersion has been through resolution and improved textures, shadows, anti-aliasing, all these little things that have made evolutionary changes to create a more immersive experience. Larger screens, all that stuff. The way you immerse yourself in VR is less dependent on those elements, the typical performance markers we’re used to now, and it’s a lot of other factors.

There’s a different balance for immersion now. It doesn’t necessarily have to be crazy resolution or crazy textures. We have other tools to play with that aren’t as burdensome on a graphics card as some of the more traditional things we’re used to. On top of that, you’re designing the content to a very specific set of hardware on all fronts, both on the PC side and on the headset side. If it doesn’t play well on 970, you can make it play well by turning things down a bit and using some other immersion tools that you haven’t had at your disposal before, or you can say, “This is going to run on a 980.”

Above: There was a long line to try out the Oculus Rift at CES 2016.

Image Credit: Dean Takahashi

Question: You mentioned optimizing for VR. To my understanding, shouldn’t that be higher GPU power and compatible drivers? What else is there to optimize?

Luckey: There are a lot of other things going on. The GPU and CPU can be optimized in a lot of ways for virtual reality. A lot of it also comes down to the system software stack, and even fundamental things like—You asked about laptops and how that will hook up. It’s not just about putting in a better GPU and CPU. It’s about architecting the entire PC in a way where the system is working better for VR.

You guys have built a lot of custom motherboards. There are lots of things you can do beyond putting in a good graphics card and CPU to make a system work well for a particular application.

Azor: I have to be a little cryptic in the answer because VR has introduced a bunch of challenges and opportunities that result from those challenges. We have a rich history of solving problems for gamers – a lot of firsts, a lot of innovations, a lot of patents, a lot of IP – in the spirit of building this evolutionary experience that we’ve seen over the last 20 years. We have a laundry list of opportunities now with VR. If I told you all of them then my competitors would have that list as well.

Question: So it’s not just a performance challenge. It’s architecturally different?

Azor: There are architectural considerations. There are form factor considerations. I wouldn’t say everything changes, but the considerations that go into our designs now have an additional layer that we have to think through that we never had to think through before. Maybe I have to sacrifice a little bit of battery life in order to make this a VR compatible configuration. That’s probably worth it. There are other issues around simple things like cables.

Luckey: One really simple issue is just having USB 3.0 controllers that are able to work properly. I don’t know if everyone remembers the early days of USB 3.0, but there were some pretty bad control chips back then. They didn’t exactly follow spec. That can be a problem when you’re relying on everything working to maximum performance on the USB 3.0 spec. Luckily it’s not really an issue now. But that’s the type of thing that–For a lot of people it doesn’t matter that you’re able to deliver maximum throughput, maximum speed, and maximum power on a single port. It starts to matter when you’re making high-end gear that does rely on that.

Azor: In the past, if there was any lag introduced in playing a game on a flat screen, it was inconvenient. Now it’s the difference between getting sick and not getting sick. What are the opportunities there, to reduce the lag across the entire configuration? It’s not necessarily just on the graphics card or the display. There are so many different challenges and opportunities that this has presented.

We’re at day one. Where was PC gaming 20 or 30 years ago? It was well established before we even got into the game. Where are we now? If I knew then what I know now—We know a lot now. We’re going to be able to innovate and accelerate and do some things that are pretty miraculous around VR that took us a lot longer when we had less resources and less knowledge.

This is revolutionary. There will be revolutionary problems and opportunities for us. We don’t have all the answers yet, but we’re invested. We’re committed. You can expect to see solutions coming very rapidly.