The Oculus Rift Review

Introduction & Overview

If the next chapter in PC gaming is about to be written, that future begins in earnest today with the first commercial shipment of the Oculus Rift arriving in homes starting now. Four years of investment, tinkering, development kits, showcases and hype have trampled a path to the moment of truth, and while there is surely much more work to be done, it's time to see where we are, and whether it was worth the wait.

Several of us spent the past week with the official Oculus Rift, and we're anxious to tell you what we learned and how it performed.

A Brief History Of The Rift

The Rift and its inventor, Palmer Luckey, first entered the public spotlight in 2012 when Luckey launched a Kickstarter campaign for the first-generation Oculus Rift Developer Kit (DK1.) The campaign had endorsements from gaming industry heavyweights such as John Carmack, the founder of id Software, who later joined the company as chief technical officer, and Gabe Newell, co-founder of Valve Corporation, which later formed a partnership with HTC to launch the competing Vive VR system.

That campaign sold $2.4 million worth of developer kits. In the nearly four years since the Rift Kickstarter campaign, Oculus has shipped approximately 200,000 developer kits, allowing many indie and established development houses to dabble in VR content creation. Curious consumers also got an early look, either through dev kit purchases or the many content showcases produced along the way.

Soon after the Oculus Kickstarter, Michael Abrash (then with Valve, but now the Oculus chief scientist) put down essentially a playbook of what needed to happen in order to make virtual reality--from a hardware and comfort standpoint--viable in consumer gear. In one quintessential piece, Abrash outlined the hefty work needed on the latency side of the performance equation, identifying challenges in tracking latency, application rendering latency and display latency. At that time, the cumulative latency of those three elements (to oversimplify) was far too great. The VR industry seemed to settle on an 11ms target, but in his piece Abrash hovered on 15ms as the absolute maximum, speculating that we'd need to get closer to 7ms.

On the tracking side, the challenges he outlined were balancing IMU sensors, which have incredibly low latency but allow too much drift, and camera sensors, which are more accurate but introduce too much latency. His answer at the time was sensor fusion, the combination of some IMU elements and optical tracking.

On the application rendering side, his answer involved simplifying scenes, essentially designing games that were, from a design and quality standpoint, equivalent to titles designed five or six years ago. And finally, he spent considerable time discussing display latency, which he figured would require some fundamental changes in hardware, including higher display refresh rates. At the time, we were dealing with 60Hz panels, and Abrash thought we needed to get to 120Hz to solve the problem.

And so the hard work began.

Fast forward to today, and much has changed since Abrash penned those thoughts. We now have low-persistence displays with high refresh rates, powerful and accurate sensors derived from the mobile phone industry, significantly more accurate cameras and considerably faster graphics processors. The upshot of four years of development kits and consumer feedback from demonstration showcases is an ecosystem of industry partners, from GPU manufacturers like AMD and Nvidia, to gaming engine providers like Epic and Unity, to infusions of cash and intelligence from the likes of Oculus and Valve.

All of that energy helped propel the industry forward at a more rapid pace. Instead of simplifying content to match what we were seeing five or six years ago, VR-based games are coming closer to current game development, according to Nick Whiting, technical director at Epic Games. "You see a big push to use less dynamic lighting and other bells and whistles, so I would say we're looking at game quality from two or three years ago, but with a lot twists. We've come a long way and learned a lot more tricks, but we're having to be crafty in ways we haven't had to since PCs started getting way more powerful than consoles."

If you consider the time Palmer Luckey spent tinkering at home, the Rift has been in development for over four years now. The hardware has gone through many revisions. When the Rift DK1 first launched, it provided the ability to look around in VR, but you could not lean in, back or side to side.

The second developer kit (DK2) was released a year later and it included an IR-based system that allowed the HMD to be tracked in 3D space. This allowed you to move around a little bit. You could duck, move forward or peek around a corner. Suddenly a greater immersion was possible, and comfort improved significantly, reducing the feeling of nausea (motion sickness) that many people experienced wearing the first Rift headset. 

Oculus considered releasing an HMD similar to the DK2 as its first consumer model, but Facebook's acquisition of the company gave it the resources to build an even better product.

The Reality Of Virtual Reality

Recent history is littered with failures in adjacent technologies, including early attempts at virtual reality and 3D TV (along with other 3D initiatives). The modern era of VR has drawn enormous skepticism from the enthusiast community, which is arguably its primary audience for now, and especially those who haven't tried it or who had poor experiences with earlier hardware. That's the downside of some of those early showcases.

In addition, the minimum specifications to power modern VR aren't easily within reach, even for enthusiasts, as some of our own research data shows. Even if they were, executives from both Oculus and Valve have said that today's GPU power just isn't enough (put more positively, they've asked for more). Game engines have VR features in their SDKs, AMD and Nvidia expose special hooks into their hardware to help keep latency down and developers use carefully crafted tricks to maintain the performance they need.

All of that is done, at least for this first round, and it's time to find out where we stand. The Rift is sold out through July and the Vive through May, but the industry's future success now rests on the performance of these first-gen devices, the content and the experiences of early adopters.

There's a long way to go before we can truly call virtual reality mainstream. For one thing, the results of our VR Readiness survey suggest that a large percentage of enthusiasts reading Tom's Hardware have yet to try even the most basic form of modern VR (Google Cardboard.) If those folks haven't been exposed yet, Oculus and its competition have a ton of people to reach still. Our survey also suggests that most folks simply don't have the hardware necessary to run VR games. Chances are good that you're looking at hundreds of dollars in upgrades, and our respondents said they probably won't spend their money just yet.

Over time, the cost of VR-capable graphics cards will come down, as will the hardware that goes into the VR headset. Until then, content creators and hardware manufacturers need to build compelling experiences that get people anticipating the day they can afford VR. Virtual reality has the potential to change the world in ways we can't yet imagine. Its promise goes beyond gaming and into fields like medicine and therapy, travel and tourism, engineering, education, architecture and yes, even pornography. Oculus may bear the biggest responsibility for affecting first impressions; it has more eyes on it than any other player, including ours. Let's get started.

MORE: The History of Virtual RealityMORE:
AMD Liquid VR ExplainedMORE:
Nvidia GameWorks VR ExplainedMORE:
Oculus' Uncompromising Obsession With Hardware
MORE: The First 30 Oculus Rift Launch Titles
MORE: The Past, Present, And Future Of VR And AR: The Pioneers SpeakMORE:
Audio, Audio, Audio: The Key To Virtual Reality Immersion Is The Audio