id Software's John Carmack has been nursing a hobbyist's obsession with virtual reality.
id Software's John Carmack has been nursing a hobbyist's obsession with virtual reality.
Earlier in the month Polygon was shown his home-brewed virtual reality simulator, a head-mount display made up of recycled work he had done for the aerospace industry. But beyond the technical labour involved in creating gameplay innovations, Carmack says the future of all successful games technology comes from a simple and animal basis: Not technical hard-power but "a wide-eyed, grinning response" from the user that what the developer has made has really "got something."
The future of innovation will read as Greek to anyone without an understanding of the engineering concepts that are at the root of products to come, he says. Even now there is a substantial market of technology users to whom technology is an alien landscape. Carmack frets about the idea that people may start feeling the largely low-level experiences they're having in mobile gaming could be seen as "good enough," undercutting the big mega-title side of things in the realm of consoles, and following on from that worry he believes the next generation of consoles will largely only whet the palate of "extreme aficionados."
"In many ways the next generation will be a lot of what we do right now with all the nobs turned up. We'll be able to go ahead and render at 1080p, running in high dynamic range and oddly enough you probably won't bust up all the additional power when you do that and then the game development is easier because there will be faster individual processors. But a lot of that will be for the extreme aficionados out there," he says.
"The biggest frustration to me is Kinect has a huge amount of waste to it."
"I fret a lot about the idea that people may start feeling that these mobile experiences are good enough that they could undercut this big mega-title side of things on console."
Carmack highlights the fact that there are 10 times as many people playing games on their cell phones as there are hardcore gamers, suggesting this might harm the growth of more complicated technical developments when so many potential users are content with low quality tech.
Peripheral developers and researchers like Microsoft have catered to this fact, Carmack says, explaining, "they're gotten away right on the border of acceptability with games on Kinect where the controls are intrinsically two steps removed because you're moving your body and you're integrating time over it.
"The biggest frustration to me is Kinect has a huge amount of waste to it and for years I've been all about the latency and responsiveness. 60 frames per second games on the console is unusual, most people don't go for that challenge. But Kinect went in the opposite direction where you've got 70 to 100 milliseconds where over the course of a second if you wave your arms and you see your character wave his arms that much time later. A third of that is down to the architecture of the way it's designed.
"But the other two thirds really could have been avoided. I've complained to Microsoft about that because it is avoidable. They actually have two cameras -- a color camera and a depth camera -- so they capture the data from one and then they compress it, then they decompress it, and then through a really complicated process we do the object identification on there. Now each of those is part of a pipeline and you've got a buffer system which means you've got an X millisecond gap in between it. This takes 50 milliseconds to come out of the back of the computer and when you think about it you could send a packet across the Atlantic to England in less than 50 milliseconds because router manufacturers give a damn about latency."
"Everybody wants this Terminator vision."
Similarly Google is already working on its own vision of virtual reality with its Project Glass, a research and development project designed to create an augmented reality head-mount display. It's a concept close to Carmack's heart but he believes they are thinking small in comparison to what is technically possible.
"The Google glasses vision of virtual reality is possible but many people in that industry are somewhat irritated at Google for projecting a vision that's sort of like VR but people for 20 years thought that you would be VR. But it's not like that, it hasn't been anywhere near that cool. And Google sort of chose that same route with the augmented reality, where everybody wants this Terminator vision but what they got with augmented reality was possibly just a postage stamp size thing on your field of view.
He admits a lot of optics are still problematic, stating that while a number of interesting innovations have come out in the last six months none of them are consumer friendly, and none, certainly, are cheap. More importantly most of these advances have been, themselves, home-brewed technical cocktails. The professional industry itself has yet to touch the boundaries of even this current generation, having failed to fully innovate with what is possible using the basic tech available.
"With consoles we have not reached the limits of what we can do this generation."
"It's interesting in that even though PCs are 10 times or more powerful than the first generation of consoles, we could go quite a bit further with this generation. With consoles we have not reached the limits of what we can do on this generation. I mean obviously the next generation is right around the corner, it's coming, but we could go a lot longer with this generation before we say this is tapped out. We haven't been technically provocative, really, this generation.
"We only know what we want to do and that we want to do it better, and really any vision the designer comes up with we can do a pretty good job on this generation. We may not have full resolution with anti-aliasing, with high dynamic range, but with that we're getting into things that a lot of people just don't care about, it's the audiophile kind of thing where they're talking about their gold plated connectors. And it's getting a little bit niche-y when you have so many people just happy to play their cell phone games. Low resolution, no anti-aliasing."
This is the reason that despite his background in hardware, his metric system for deciding what matters in the development of a game is whether someone walks by his office, glances in, and is able to tell whether it's on a new generation. While some players simply can't tell the difference between 60 and 120 hertz on a PC, he says, all users can tell the difference using a head-mount display. A mouse response on a PC is based on your hand's relation to your head position, but a head-mount display is "zero degrees of separation. Your head knows what the world is supposed to look like when you turn your head. Everyone can tell the difference between 60 and 120 hertz there."
His current model for the simulator is still rough, built from material put together in a workshop then tinkered by Carmack and colleagues. It's hardly a beauty with its ski mask-on-duct tape chic but it speaks to how technical leaps can be made simply by harnessing an armchair engineer's work ethic.
"All the pieces became available and right now we could be integrating them."
"All the sudden that this grand technical vision of decades ago of virtual reality and virtual worlds happened without anybody noticing. All the pieces became available and right now we could be integrating them and tie them together."
He maintains it's possible to put together a working model right now built almost entirely from off the shelf parts, acknowledging the resolution may be worse for wear. However he speculates this is will be a hacker's market, with budding programmers and engineers finding the way to move the tech forward.
"I'm so excited to see this become a hacker homebrew-maker product," he says. "This is not the final product, it's a kit, you can build it. And they're actually going to be at the cutting edge of research here. Figuring out what's the best sort of ergonomics, these are things that individuals in their workshops can make real, legitimate contributions in."
Carmack isn't leaving it entirely in the hands of individuals. ZeniMax itself has no intention on making head-mount displays, but instead Carmack has been briefing Sony, amongst others, stating he has no preference who runs with the "hardware ball" whether it's Apple, Sony, Microsoft or Google. He is currently in talks with a number of companies about the future of this tech, despite it still needing some work.
"It's a systems integration challenge because VRs have a long pipeline, and if you do a botched job on any part of the pipeline you can ruin the entire thing," he says."We're still working on the hardware side of things, we need 120 hertz updates to really get it right, we need better vision oriented accuracy, but it's all very very close, it's all in play."
"It's all very very close, it's all in play."
Referring back to his metric for "what matters" in tech development, Carmack speculates the next generation will be even more difficult for the average user to gauge. "We were able to judge whether we're dealing with new tech this generation," he says, "and the last generation, all the way up to now. Although it was a little bit harder on this generation but it's going to be even harder the next time."
He imagines the future generation will include a focus on game design auteurs, in which large-scale console games will be headed by creative geniuses, and where E3s in the next coming years will feature trailers that will have "directorial, cinematic visions" that wouldn't feel out of place in a movie theater.
"But I'm on the technology side, all of art can be decomposed into statistics. When you set down the controller at whatever point you were in the game and take a look at the game it's going to be better, it's going to be awesome, but it's not going to be this completely different experience."