Later this month, Valve will update Team Fortress 2 to make it compatible with the Oculus Rift virtual reality headset (giving Rift users a fancy new Oculus TF2 hat in the bargain), and TF2 will once again find itself in the familiar role of being a test bed for Valve’s foray into a new kind of gaming.
I had a chance to spend an afternoon at Valve playing TF2 with the Oculus, and learning what Valve have planned for the VR platform. Over the course of a half dozen bot matches, I discovered that the Oculus absolutely works as advertised, providing a natural 3D immersive experience. But that turns out to be the easy part.
Of all things, it’s the fencing that gets me the most. Right there in the Blue team base at the start of Badwater Basin, waiting for the Payload match to start, that’s when I start to see what the Oculus Rift could mean for immersion. The sense of presence it can evoke is remarkable. Something as common and unremarkable as the chainlink gates, which I’ve seen a thousand times before, become something almost tangible. I want to hook my fingers around the links and rattle the fence a bit. Beyond it, the familiar sights of Badwater rising in into the distance: red-brown dust, the hill dotted with rock chimneys, and beyond it, the rooftops and sky. Familiar sights, but this time it felt like a space I could get out of my chair and walk through.
“It’s the parallax effect. You notice it a lot more with the combination of 3D and head-tracking,” explains Joe Ludwig, the Valve programmer leading VR development. “There is so much more motion coming from your head that you’re getting from the game, plus the 3D, that your brain just gets a ton of feedback that it’s in a 3D space and start interpreting it that way.”
And once you have made that leap, TF2 in the Oculus becomes something not quite new, but a vision of what is to come. It is disorienting and engrossing, at once both more natural and more contrived, and above all it is more convincing. In some ways I felt like I was seeing it for the first time. The strength of the experience has me looking forward to the day when developers can create games intended for 3D VR headsets like the Oculus.
We’re not there yet, but when Valve release the VR update for TF2 later this month, we’ll be a big step closer.
Time to build railroads
VR is an idea whose moment has never quite arrived, always running aground on a combination of technological limitations and immaturity in its potential marketplace. “VR has been ‘in the next five years’ for fifteen years,” Ludwig says. “There are three big reasons why that was true. The old headsets themselves were terrible. They were heavy and expensive, and they had to put a bunch of glas between the display and the eye in an attempt to correct the distortion that you get just from magnifying something that’s so close to your eye. Sometimes they even used CRTs for displays.”
Ludwig also recalls that during the last great burst of VR enthusiasm in the nineties, there weren’t many 3D games around that would take advantage of it. 3D gaming was in its relative infancy, so VR gaming consisted of glorified tech demos.
“So the content that you got when you put on these terrible displays was not very compelling,” he says.
The third issues was expense. “A VR system would cost you tens of thousands in the early nineties, and that’s way beyond what an individual could buy. So they ended up being used in these location-based entertainment service scenarios, and a few years later they all went away.”
Several important changes have occurred in the last several years that make VR much more practical. For one thing, improvements in video cards mean that VR no longer requires expensive, delicate optics to display correctly to users.
With no correction, VR headsets appear to have bad pincushion distortion, where everything appears tapered inward toward the center of the image (you may remember a pincushion setting on your old CRT monitors). What video cards can do, however, is render images with an offsetting amount of barrel distortion, giving everything a bulbous appearance that will display correctly when viewed through the pincushion distortion of the VR headset.
The advent of smartphones led to several initiatives in miniaturization that make high-res VR displays practical as well, but perhaps more importantly they’ve also led to the proliferation of MEMS (microelectromechanical systems) gyroscopes and accelerometers. That has greatly simplified motion-tracking and responsiveness.
With all these pieces in place, Ludwig says, another major foray into VR was inevitable, and 2012 saw a few different groups taking a serious look at the technology. Valve were interested in VR as a part of its investigation into wearable computing and Augmented Reality. John Carmack started looking into VR and seeing how he might adapt his games to it, and the Oculus Rift raised $2.4 million via Kickstarter. Ludwig explains the sudden convergence with a quote from technology writer Cory Doctorow: “When it’s time for railroading you get railroads.”
Obstacles on the track
Ludwig warns me before we started my session that a large majority of players suffer from motion sickness in VR. Even Ludwig, no stranger to FPS games, admits that he struggled with adjusting to Team Fortress in the Oculus Rift.
Valve’s early tests involved just nine people, and seven of them wanted to stop after just 20 minutes of play. One thing that Valve have found that helps is a process of progressive acclimation. Starting a new group of players on a week-long regime of increasingly long play sessions got most of them over the worst reactions, but Ludwig is still concerned that it’s going to be an obstacle.
More immediately, he’s concerned I might get sick all over his group’s workspace. He sets me up at a folding table a few feet away from the actual PC running TF2. It feels a bit like the kids’ table at Thanksgiving. The old Oculus Rift Valve used for the early stages of development is sitting atop it, sporting googly eyes and a fake mustache, but I’ll be wearing a much newer, nicer, and clean-shaven rig for today’s session.
The Oculus is surprisingly comfortable and balanced. After I put it on and start playing, I immediately forget the device strapped to the front of my face. However, a couple drawbacks are also apparent. First, it’s very low-res. Without much concentration, I can spot the individual square pixels that make up the image, and trace the pitch-black street grid between them. Second, it seems to be calling attention to the fact that my right eye’s contact prescription is badly out of date. WIth both my contacts, I usually have no problem seeing or reading, but inside the Oculus Rift I’m struggling with blurriness.
But even before the round starts, I’ve largely acclimated to my new perspective and started to lose myself in the feeling of presence the Oculus Rift can grant. The round begins and as the heavy, I make a methodical advance down the tunnel at Badwater, filling it with chaingun rounds like I have a hundred times before. But the difference here is that while I’m engaging soldiers and medics up ahead, I’m scanning my surroundings looking for new threats. I narrowly sidestep a rocket and track it as it flies over my right shoulder, feeling like it’s only inches from my face.
The controls are a double-edged sword. On the one hand, independent looking and aiming seem like a huge tactical advantage. I’m able to cover multiple angles of approach at once, head on a swivel as I glance at potential danger areas while my aim remains steady on another target. On the other hand, being able to aim independent of my character’s movement feels completely alien and at odds with twenty years of shooters. I struggle with it throughout my session, and it’s made a bit worse by the way the reticle still turns my character once it hits the edge of the display. Mouselook no longer turns your character, except until it does.
Ludwig admits this is a problem, and it may not be solvable for a game like TF2, whose design is so rooted in classic shooter traditions. But more broadly, it points to how early in its development VR truly is.
“For input mechanisms, we’re in the same state the FPS was in before mouselook,” he says. “This is super early. Nobody knows what the right input scheme is. And this is something we’re going to have to figure out over the next several years.”
Oddly enough, most users have a suggestion for their ideal control mechanism: some kind of gun.
Ludwig tells me, “A lot of people seem to think [the controls] will be motion control. The second thing people say after, ‘Wow!’ is ‘I want a gun!’ We’re not sure that’s the way it’s going to go. Because once you get your shoulders and elbows involved, it tires you out. And it’s a lot less precise than wrists and fingers. But we’re willing to be proven wrong.”
Mouse and keyboard works fine for me during my session, but while WASD is no trouble, trying to put my hand on other keys is a much trickier proposition. I nearly send a cup of coffee flying across the room during one ill-advised flail for the function keys. The Oculus, which in some ways seems like it could be a godsend to sim gamers in particular, creates some obstacles for games with more complicated controls.
But for simpler fare, like TF2, it’s really impressive and easily the best use of 3D technology I’ve ever seen, in part because it does not at all look or feel contrived. I never ran into the motion sickness issues Ludwig warned me about, even after I start bouncing all around the level with the scout. By the end I’ve even adjusted enough to the controls that I start to benefit from the VR’s fludiity. At one point I shotgun a demoman in the motor pool of Badwater, then switch to my pistol whip a volley of shots over my left shoulder (yes, it really felt that way) to drop a soldier coming down the ramp.
Watching him fall, I start to feel like VR is something I could get used to.
I hope I have that chance, both in TF2 and other games. But Ludwig and Valve are vague when it comes to their intentions and hopes for VR gaming. That’s understandable, of course: TF2 is their first foray into VR gaming and nobody really knows how it will be received by its community, or what will end up happening with the Oculus Rift.
“It’s an experiment for us,” Ludwig admits. “Most of those people have never tried VR. It’s really going to be telling what happens the next few weeks as people actually get them in their hands. In theory at least, [Oculus headsets] are dev kits and only developers are buying them. We don’t know what’s going to happen. Oculus doesn’t know. It’s a big mystery. But it depends on how actual gamers respond to this stuff when they actually get their hands on it.”
Valve have come to VR in a bit of a roundabout way, treating it as a laboratory for their own investigations into wearable computing and augmented reality. Ludwig explains that “AR is a ton of hard problems to solve before it works. VR is some of those problems. And once we get VR up and running, all that work will apply to AR.”
For now, Valve are mostly focused on the VR update to TF2, and how that community reacts. The reception VR gets from TF2 players will both inform Valve’s plans for VR in the future, and influence how they address the challenges it poses. But Ludwig admits that they’re flying blind here, and it’s hard to predict anything about how people will take to the new technology.
“This isn’t something you can read about in a press release and understand,” he tells me. “It’s something you need to see and experience.”