Intel’s VR chief expects virtual reality to divide into multiple gadgets and markets
摘要：Above: Mark Hachman of PC World at the Intel CES 2017 press event. Everyone wore VR headsets.Virtual reality hasn’t taken off in a big way with consumers, but Intel believes in its potential. It’s not clear just yet what will be VR’s biggest hit, but the platform will produce a huge amount of data that will have to be processed in the cloud using Intel infrastructure, said Frank Soqui, general manager of virtual reality at Intel.
Above: Mark Hachman of PC World at the Intel CES 2017 press event. Everyone wore VR headsets.
Virtual reality hasn’t taken off in a big way with consumers, but Intel believes in its potential. It’s not clear just yet what will be VR’s biggest hit, but the platform will produce a huge amount of data that will have to be processed in the cloud using Intel infrastructure, said Frank Soqui, general manager of virtual reality at Intel.
So Intel is working on improving VR on a variety of fronts, including more affordable and better headsets, impressive applications on the consumer side, and, for enterprises, ways to process all of the data in the cloud and deliver it efficiently to users. The to-do list includes improving WiGig, a wireless transfer technology that has a short range but delivers lots of bandwidth in the 60-gigahertz spectrum. And Intel is working on RealSense sensing cameras to better detect where you are in a virtual environment.
Soqui believes that the market for headsets will separate into distinct segments, including mobile, standalone headsets; PC-based systems; and heavy-duty location-based entertainment.
“It’s a compute continuum,” he said in our interview at the VRX event in San Francisco last week. “What you realize is, one size doesn’t fit all, because each one has its use profile, performance profile, whether you need it to be portable, mobile, or completely immersive in a sit-down perspective.”
Here’s an edited transcript of our interview.
Above: Frank Soqui, general manager of the virtual reality group at Intel.
VB: What was some of your talk’s focus here? You’re talking about a lot of different partners?
Frank Soqui: I talked about a technology point of view that was more than just the end point. I talked about a technology point of view that was about these highly immersive work flows, the terabytes of data they’re throwing off, the ability to manage those terabytes from cloud to edge to client — that’s becoming increasingly important. For those to be really interactive, low-latency, we believe we’re going to have to layer on top of existing workflows, workflow definitions, to make sure that they can be highly interactive and immersive, not make people motion sick.
We’ve started in several technology areas. On the client side, applications, road map improvements. On the headset device side, we have computer vision and RealSense technologies, as well as WiGig. We announced a partnership with HTC on WiGig. From a broadcast perspective, making sports more interactive in a VR-type capability, letting end users control what they want to control from any angle, replay zoom in, zoom out — just being able to view sports any way you want to.
I talked about the business value. We actually showed it in context. We have quoted improvements in efficiency and cost savings because of what they’re able to do with VR and Intel technology. I talked about car manufacturing and customer acquisition, and I talked about construction environments and customer acquisition. I talked about surgical theater — not only surgical outcomes, but the customer connection. The constant theme throughout this is helping businesses do what they need to do more efficiently, with better results, as well as extending to better customer connections in almost every one of those cases.
I could have done the Smithsonian as another example. We’re helping the Smithsonian as well. We’re learning very early on about enterprise-level deployments, not only from the PC endpoint perspective, but about how it connects to the rest of the workload and work flow, so we can improve that pipeline.
Above: Intel CEO Brian Krzanich shows off Project Alloy at CES.
VB: I talked to Kim Pallister at Intel a couple of months ago. He was saying that you had an idea for Project Alloy, bounced it off a lot of people, and nobody really went for it. This idea of doing WiGig from a desktop PC, getting more processing power to work on the VR imagery and then getting that through WiGig into the headset, looks like it would work really well. But then I also saw the WiGig announcement where it wasn’t clear how much Intel still wants to do — you’re not going to use it for wireless charging, right?
Soqui: There’s WiGig for backhaul and WiGig for VR. That should be how you think about what we’re looking at with WiGig. And then WiGig as part of our 5G evolution. As we get to 5G WiGig will come along for that as well.
VB: Investment in WiGig, or the confidence in WiGig, is still pretty high?
Soqui: Yeah, confidence in WiGig is high. That 60Ghz unlicensed band is still pretty valuable. We have made it part of our 5G investment portfolio and our road map plans.
Above: Intel’s demo of WiGig-based VR.
VB: WiGig does get to some kind of cost curve, then? It gets into production and cost comes down.
Soqui: As we get into integration of 5G and calling out the 60Ghz band, yeah, more integration means it’ll get less expensive.
VB: At that point, VR then makes good use of it?
Soqui: We’re going to make good use of VR now, while it’s a discrete accessory. Because of the market’s nascence, an accessory market looks good right now. And then as we experience the success of that market and broader adoption, then we get to more integrated solutions. Not only in our chipset, through our 5G investments, but integrated in the headsets. But right now? In the nascent market? Everyone’s gut reaction is, boy, this is really important, let’s test it as an accessory market first. Some people will try integration right away, but I think people are still cautiously optimistic about VR in general, let alone how much of that market can be taken to wireless.
VB: Kind of like how HTC is selling these trackers now.
Soqui: Right, to test the value and the use. Price is whatever price needs to be.
VB: That still makes the PC pretty important, right? The WiGig solution as opposed to the standalone headset.
Soqui: Here’s how I view it. It’s a VR-compute continuum, the way I even view PCs as a compute continuum. There’s mobile. There’s desktop. There’s laptops. There’s two-in-ones. There’s all-in-one. It’s a compute continuum. What you realize is, one size doesn’t fit all, because each one has its use profile, performance profile, whether you need it to be portable, mobile, or completely immersive in a sit-down perspective. From a PC perspective I include laptops, notebooks, and two-in-ones, even down to 15-watt devices, which we’re introducing this year. People want the ability to have those be VR-ready, VR-capable. That’s what the client group is focusing on.
In addition, we’re partnering with the Movidius and RealSense guys to help them, and help ourselves, make PC experiences more valuable. With the sports group we’re making sure that the application that runs on a phone runs better on a PC. PC becomes central to a lot of those types of extended business models as well. You add in the data center — you’ve heard Brian talk about this. The virtuous cycle, cloud to client, that’s exactly what’s happening. Now I’m trying to add in the cross end-to-end platform elements because I want to optimize that pipeline, hardware-wise and software-wise, so we can deliver really compelling interactive experiences.
Above: Intel’s WiGig-based VR demo.
VB: I see that the solution that might win is the one with a desktop PC to WiGig to VR headset. You take the wire out of that headset and it’s much better.
Soqui: The reason I call that a compute continuum — I think it’s a false assumption to believe that there’s one winner here, however you describe winning. One of the analysts was showing me today — more users might be using phones to do VR, but people might be making more money off a PC, even with fewer units. I don’t think anybody’s going to bet on just one. A lot of software developers want to make sure they can scale their experiences as much as they can across the available compute. The ones that can do that, people that enjoy their experiences and own multiple devices can have the experience on anything, and it scales with the capability.
That’s how I see the market evolving. This idea of one size fits all — how many devices do you have at home? I know I have a tower, a notebook, a phone. My TV is smart. I have compute everywhere. I don’t rely on one device. It’s situational, depending on what I need to get done and how I want to experience something. I have different tools to do that.
VB: Do you have a good feel yet for what you’d consider to be VR 2.0? VR’s second generation? Or the generation of VR that works?
Soqui: I’ll look at it like, how do I see VR evolving? There’s a number of vectors. The market is going to segment some more. Right now people are trying to produce one headset and make it span everything. It can’t. That price point and the cost, it just can’t happen. We’re going to start seeing a more segmented market instead of pure price competition. All you’re doing is giving away margin when you do that. To maximize the return to your company and maximize the customer experience, you have to have this segmented model.
My prediction, in this coming year, we start to see segmentation. You see higher-end headsets, you see medium range, and you see a low-end range. The challenge for the user will be, “How does my content run on these?” Setting that expectation. I think there will be some confusion, some education we have to do with end users as that segments. It’s partially the responsibility of the hardware providers and partially the responsibility of the application providers.
If I were an application provider, I’d make sure that my software scaled according to available compute, and so my customer experience wouldn’t be bad at all. It would be like, “Hey, you bought this platform, and this is the experience.”
VB: How much of this VR continuum would you think Intel is going to target and serve?
Soqui: We’ll target them all. Even if you say, “Well, Frank, you don’t make phones,” all of that data goes back to the data center. We have to take into account those work flows and workloads as well. At every layer, Intel needs to be cognizant of the value proposition and how we’re going to improve that.
I’ll give you the Intel support group as an example. They started out bracketing both ends. They had these amazing Xeon servers, and even though we don’t have silicon in the phone side, they were targeting phones as a consumption model. One feeds the other. Now we’re including PCs. We make strategic bets along that entire compute continuum, cloud to end to client.
Above: SK Telecom and BMW have developed a connected car running on a 5G network.
VB: The Olympics seem like a big thing for you guys. You’re going to have this 5G demo there. I don’t know what else goes with that.
Soqui: I don’t think we’re disclosing a lot right now about what we’re doing with the Olympics. Sorry. We announced categories, yes, of things we want to go do.
VB: Where do you think VR is going to make a splash, then, as far as the coming months?
Soqui: Sprint Vector looks like it’s getting a lot of interesting attention now, because it’s also competitive, and it includes motion. We have some high expectations for that. I’d like to see Fallout 4 come out, with how well they’re doing in the VR space. I don’t know this for a fact, but I’d like to imagine that games like PUBG and H1Z1 — they’re ideal for VR, just ideal. I’d love to see them extend to that. I know there’s a development cost, but I’d love to see them do that.
The Windows Mixed Reality headsets, it’s a tough start. They’re new and they need more content. But I’m very optimistic, because Microsoft has done two things that are very encouraging. They’ve adopted the Steam client, the Steam applications, and they’re part of OpenVR. They’re starting to move in a direction that’s going to foster a lot more innovation. You’ll see more innovation and experimentation. You’ll see a lot more proof points happening on the commercial side and the enterprise side.
VB: Sony is out there, but do you think it’s inevitable that all the consoles will have VR at some point?
Soqui: I would say it’s hard to ignore competitive pressures. It’s hard to ignore a compelling use experience. I would imagine that you have to enable that on your console. That would be my point of view. It’s too hard to ignore. One supplier, one OEM, is already proving it has value. When there’s value, everyone is going to want a piece of that. They can’t be left out of it. Sony’s definitely proven that.
Above: Lyle Becker, 100 years old, tries out VR for the first time.
VB: What’s the thing that people are going to go to Intel for? What expertise are you guys going to drive?
Soqui: One is our developer expertise. The Software Solutions Group is an amazing resource. We have developer relations. We have partnerships and engagements, engineering engagements, with all the engine guys, all the ISVs that are important to this market, both in the commercial and consumer spaces. The strength of our data center, thinking about this end to end, is another strength. The fact that we focus on open standards.
We’re a trusted advisor. We want a lot of innovation on top of that open platform. People come to us for consulting because they’ve come to us for consulting before on cloud, on edge, managing secure PC clients. Especially as VR is intersecting enterprise, they’re going to come to us for that as well. “How do we manage the headset? I’m using vPro and it’e entering my environment. What’s your advice?” They come to us for our point of view on value propositions that we can help them either test or implement. We’ve pulled partners along in with that – ISVs, outsourcers. We do this all the time, all day long.
VB: On those partners, are there some you would name that are key partners?
Soqui: No, I wouldn’t say that there’s a “key partner.” We tend to focus in a couple of areas. Who are the market leaders, the market makers that we want to go partner with? We can move the market faster. Where are the independents that are more innovative, that we want to take some risks on? Really cool ideas come out of those guys. We kind of bracket it, instead of saying there’s one we like.