'We are coasting on novelty': Carmack advises VR devs to build better mobile VR apps

Today Oculus CTO John Carmack shared his thoughts on the state of VR development in his closing keynote at the company’s Connect dev conference.

Carmack, of course, is a veteran game developer and engineer, so it’s always interesting to hear his perspective on...well, just about anything, but especially game design and rocketry. Both topics came up during his keynote, but devs should know that Carmack’s primary focus this year, as it has been almost since he joined Oculus in 2014, was mobile VR.

He spoke for over an hour and covered a variety of technical details (you can watch the full keynote on Oculus' Twitch channel), but one of the bits that jumped out was his emphasis on how VR developers should be looking towards mobile, rather than PC-powered VR platforms.

"The future is mobile"

“It’s still not easy to make a Gear VR app,” Carmack said at one point. “But I come up here and I say, the future is mobile. I think as we look towards a billion people, it’s not going to be on PC.”

To be clear, Carmack didn’t dismiss the value of PC- or console-driven VR headsets. But he did point out that there are far more VR-ready phones in the world than any other VR platform, and that's not likely to change in the near future.

"There might be a hundred million PCs that can do this, but I believe in the mission that Facebook had when it bought into Oculus, of having a billion people in VR. So it’s not going to be a higher and higher bar for performance; it’s going to be a lower and lower bar for adoption.”

It's clear that Oculus as a whole has adopted not just Facebook's mission, but its way of doing business -- this year's Connect painted Oculus as more of a platform company than a hardware company.  Carmack's advice to devs, then, was to look at mobile VR headsets (like the Oculus-powered Gear VR) not as low-powered, inferior VR hardware but as entry-level VR -- the place where you can potentially reach the most players.

“This level of perforrnace, I don’t think is going away. Things will get a little easier, but it’s still going to be hard, and [mobile] is where I think the most people you can impact are,” he said. “The PC is going to be the creative class. It’s going to be the laboratory [...] where we mine all that and say 'Okay, what do we want to push out to the lower-end systems.'”

And indeed, much of Carmack's talk was about how developers can tweak their VR games and apps in ways to make them perform effectively on mobile VR hardware.

“Sometimes you marvel at what some artists can do with just a few lines,” he said, by way of analogy. “You can’t change the tool right now, or the hardware, so that means your design needs to change. You don’t make [a game] in a vacuum.”

While he sympathized with many devs' complaints about having to build great games for smartphones that are prone to heat issues and often have tons of things running in the background ("if we could just push everything that’s not VR-related -- not kill it, but push it to one core [...] I have some hope we could get ot that point eventually,") he reminded devs that while modern smartphones are underpowered compared to top-of-the-line PCs, they're actually overpowered compared to the machines people were making great games for just a few years ago.

“Find old-timers, anybody that worked on an Xbox or an original Gamecube or something like that, and tell them your minimum clock speed is 800 megahertz or something," he said. "They’ll say, 'megahertz?!' It’s aboslutely possbile to still do great things with that.”

He wasn't joking, either -- at one point Carmack suggested some older game makers should come into mobile VR development, since they're potentially well-versed in wringing meainingful experiences out of tricky, underpowered hardware.

"There are people out there who I think should be recruited into the VR space, who have been there and done that 15 years ago,” he said. Beyond that, he noted that “I hope to inspire some people to buckle down, learn the right techniques, learn what you need to do differently to actually make [great mobile VR experiences] happen, so we can get richer sets of content on this.”

A big theme of Carmack's presentation was enriching the overall level of VR software, on mobile and in the VR market as a whole. He cautioned devs that VR has a patina of novelty, one that quickly gives way to disinterest if someone finds themself using VR apps and games that are inferior to non-VR variants.

"We are coasting on novelty"

"We are coasting on novelty, and the initial wonder of being something people have never seen before. But we need to start judging ourselves, not on a curve, but in an absolutely sense," he said. "Can you do something in VR that has the same value, or more value, than what these other [non-VR] things have done?"

If not, maybe it's time to carefully evaluate whether your project needs to be in VR -- or even needs to exist at all. "We need to be harder on ourselves," said Carmack, before going on to outline some concrete guideliens to making higher-quality VR experiences.

“Now there are objective measurements of quality," he said. "Load time is one of them. In the game industry [...] most teams are fighting and struggling to get load times down to 29 seconds so they can call it a day. That’s acceptable if you’re going to sit down and play for an hour….but [in VR] initial startup time really is poisonous An analogy I like to say is, imagine if your phone took 30 seconds to unlock every time you wanted to use it. You’d use it a lot less.”

He went on to outline more technical ways to improve VR games, like improving the user interface ("you do not want to have text floating across empty space")  and using voice in games ("voice is still underutilized in VR apps. It’s one of the best things in adding quality, and [...] is relatively cheap"), but he returned multiple times to the need for minimal loading times in VR.

“There are apps that I wanted to play, that I thought looked great, that I stopped playing because they had too long of a load time," he said.." I would say 20 seconds should be an absolute limit on load times, and even then I’m pushing people to get it much, much lower.”

In some ways it’s worse in VR, Carmack says, because unlike with PC or console games people can’t pull out their phones during a load screen to check Twitter. They’re trapped inside the headset, with the loading screen plastered on their face.

And when he began running out of time to speak, Carmack cicled back around to his central piece of advice for devs:

"The biggest thing I want to impress on people is that you should not design VR apps just to be novel. This is a mistake almost everyone makes, to some degree -- ‘this is VR, I want this to be novel.’” he said. “This is misguided. It’s not just that it hurts your performance, or the visual quality isn’t as good; it’s actually the wrong thing to do."

Instead, he exhorted devs to design VR apps that are good -- that, as mentioned earlier, are as good or better at what they aim to do than their non-VR equivalents.