Digital Foundry: hands-on with Project Morpheus
In-depth analysis of Sony's VR headset - and why PS Move is the real game-changer.
It's been almost a year since we first heard about the existence of a Sony virtual reality headset. What began as a vague rumour about a potential Oculus Rift PS4 hook-up gradually shifted into something much bigger - and infinitely more exciting. The hardware we've tested and the demos we've played confirm that this is no mere Oculus copy: there are commonalities in the technology, but Sony has its own distinct vision for VR, and a console-specific strategy for getting the most out of the hardware.
It all starts with the behind-the-scenes organisation - the foundation on which Sony's VR development is built. At the GDC reveal earlier this year, we saw two familiar faces - Dr Richard Marks and Anton Mikhailov, key R&D staff behind PlayStation add-ons including EyeToy and the brilliant but under-utilised PlayStation Move. They are joined by the newly-formed Sony Immersive Technology Group based in Liverpool in the UK - effectively an evolution of the stereoscopic team that spearheaded Sony's PS3 experimentation with 3DTVs. Finally, in Japan we have the hardware team, whose contributions include sculpting the Project Morpheus prototype itself.
This international collaboration has produced a headset that isn't quite ready for the consumer, but clearly demonstrates that Sony is on the right path. As discussed in our recent Project Morpheus Spec Analysis, the parallels with the second-gen Oculus Rift dev kit are numerous - suggesting that two sets of gifted engineers faced with the same problems came up with very similar solutions. But having sampled the Morpheus prototype, the differences - though subtle - are actually quite profound. Sony is behind Oculus in one key respect, the display, but substantially ahead elsewhere.
In PlayStation Move, it has a 3D controller that's already part of the console's infrastructure, and has no real equivalent in the PC space that Oculus can tie into so directly. And then there are the ergonomics of the head-mounted display itself: Sony has always made exceptional hardware, and with the Morpheus prototype it has done it again. The visor itself is relatively svelte and light but can easily accommodate those who wear glasses, meaning that the multiple lenses of the original Oculus Rift aren't really needed. The headset straps on using a Rift-style band, but this is supplemented by an additional rear plastic "halo" that secures itself onto the base of your skull, clicking into place. It helps in distributing the weight of the headset, but it also houses additional LEDs - meaning that if you turn away from the PlayStation Camera, it can still track your head movements accurately.
An unwieldy head-mounted display (HMD) is a key immersion killer, but the ergonomics of Project Morpheus are an understated but hugely important feature. The headset locks in place while remaining very comfortable, so swift movements don't cause any nausea-inducing wobble in the display right in front of your face. Being designed by an international team also helps - the hardware factors in multi-racial physical characteristics (like the European Roman nose for example), with all Sony territories roadtesting the design to ensure a common fit.
However, while highly impressive for a prototype, it's clear that there is still work left to do. Bizarrely, the headset jack is currently positioned inside the visor, which doesn't seem like the best place to put it, while there's both good and bad news about the current development kit's LCD screen.
To address the main problem first, it's clear that the display suffers from motion blur issues, manifesting profoundly with any kind of fast motion - an inherent problem with LCD technology, exacerbated by the fact that your eyes are mere centimetres away from the screen. We're also fairly sure that the field of view is indeed narrower than Oculus Rift's, manifesting as black rounded edges on the left and right of your vision, intruding slightly into the field of view - but easily 'filterable' by the brain, and not unduly impactful to the experience.
The good news is that the display is no way final - Sony is actively investigating alternative technologies to address the blurring problem with OLED first and foremost in its deliberations. Elsewhere, we found that the 960x1080 per eye resolution is actually far more impressive than we thought it would be. Remember that those pixels need to accommodate both your focus and peripheral vision. On the 640x768 per eye first-gen Rift, the result was the perception of a disappointingly miniscule resolution, with a highly distracting "screen door" effect where you could see between the pixels.
"While the PS4 may lack the raw rendering power of a high-end gaming PC, the fixed platform means that developers optimise to the console spec and in Move they have a superb 3D motion controller."
This is far less of an issue with Morpheus, and we were pleasantly surprised by how good image quality is in an environment where resolution remains at a premium. In discussing the situation with Sony, it's clear that some effort has gone into judging how to best apply the fisheye lens effect that distorts the image, with a stronger focus on retaining resolution in the key focus area. Over and above that, we wouldn't be surprised if the narrower field of view also contributes to improving image integrity.
Sony hasn't ruled out a higher-resolution screen either: PS4's HDMI 1.4a support should - in theory - accommodate a 2560x1440 output at 60Hz, and assuming a full-blooded Morpheus release sometime next year, there should be plenty of mobile screens for the company to choose from. However, our money is still on a 1080p panel for the retail headset, and while higher resolution would be preferable for image quality (albeit much more challenging for developers), this is still good enough to produce a genuinely immersive experience.
The proof of the pudding can be seen in the 'social screen' - a bespoke video output that emanates from the Morpheus breakout box that's connected to the PS4's HDMI and USB ports. The social screen allows other gamers in the room to connect with the HMD wearer's VR experience by processing one of the 960x1080 eye views, removing the fisheye effect, zooming in on the focus view and outputting that to the HDTV. The result is a little strange: as it's 'undistorting' the fisheye view, the edges of the image run at a progressively lower resolution the further out you go, but it does allow you to appreciate that although VR is necessarily reduced in terms of pure pixel-count, it gets the job done.
However, in comparing Morpheus to what we've seen from Oculus VR, it's perhaps surprising to discover that a truly transformative element of the proposition comes from a piece of hardware that you might already own: PlayStation Move. Our aspirations for the hardware were never fully realised, but the hook-up with Morpheus is a match made in heaven - in fact, if there is to be a struggle for market leadership with Oculus (and potentially Microsoft), the existing motion controller is undoubtedly one of the strongest weapons in Sony's arsenal. In our look at the first Oculus Rift dev kit, one of our key conclusions was that the headset demanded an improved interface compared to keyboard and mouse or a console gamepad. Move isn't the perfect solution, but it's a fundamental step closer to a more deeply immersive form of gameplay.
Sony's Castle demo was first seen at GDC and illustrates this quite nicely with a spectacularly successful integration of Morpheus and Move. Look down and you can see your hands. By holding two Move controllers, the motion of your arms is perfectly mapped in-game to the on-screen appendages. Press the main triggers and your fingers flex. Reach down towards a sword, press the trigger and you pick it up. Turn your head towards a second sword in its stand, repeat the process with your second Move and you're dual-wielding both blades that clash as you strike one against the other.
In front of you is a vacant suit of armour to interact with using your swords. Here's where the next of Move's arsenal of technology comes into play: swiping into the air with the motion controller sees all of that kinetic energy accurately mapped in-game. Lop off the head with a swift lunge, tap the armour gently with the sword - everything reacts precisely as it should. It's a simple demo, but as a proof of concept it is compelling: Morpheus is all about immersion, but it is Move that supplies a crucial component that felt missing in our initial Oculus testing - the means to physically reach out and interact with the game world.
The castle demo concludes with a spot of archery-based target practice (think Sports Champions 2), again with Move taking centre-stage, before the demo ends with the arrival of a screen-filling dragon, demonstrating the immense sense of scale Morpheus creates - to the point where we actually backed up into a real-life wall to get a good look at the beast. But it's Move that is the star here, and at least one of the controllers - preferably two - should be bundled with the headset when it eventually hits retail: it's that crucial to the experience, and a defining factor in separating PlayStation VR from its competitors.
Sony's second demo - the Deep - is far less interactive, and more of a rollercoaster ride of sorts, designed to engender emotion via immersion. Descending under the sea in a metal cage, you find yourself checking out benign marine life before suddenly coming under attack from a shark, with only an ineffectual flare gun for defence. Controlled by the DualShock 4, the process of pointing and shooting is decidedly two-dimensional in nature compared to the castle demo, and doesn't feel quite right in comparison. You can point with PlayStation Move, and there's even a pistol-style mount for the controller - the implications here for first-person shooters are obvious. The Deep is an interesting example in emulating the fear of the unknown (you are constantly looking around in an attempt to see where the shark will strike next), but the lack of interaction is palpable. You are along for the ride, but you are not really in control of it.
"Sensory immersion is one thing - PlayStation Move is the key to introducing meaningful interaction with the in-game world."
We'll be returning to Morpheus soon with an account of the in-depth talk we had with the Immersive Technology Group, but several highlights from the conversation spring to mind. Firstly, Sony is treating VR almost like an offshoot platform, as opposed to a simple PS4 peripheral. The team is talking to developers, implementing recommendations for best practices, and defining a series of technical requirements that ensure a good experience. A key element is smooth frame-rate: the team defines 60fps as the bare minimum for an immersive VR experience (actual specs for the final display and its highest refresh rate are unknown right now, but there may be a higher frequency option).
The sense we get is that rather than porting triple-A console titles to Morpheus, Sony recognises that this isn't completely compatible with the best VR gameplay and is actively pushing for original content. Bearing in mind the minimum 60fps requirement, the advice the Immersive Technology Group offers is that games should be built around PS3-quality visuals as a foundation, before adding further embellishments, but always with 60fps as a base-level target. Frame-rate is the primary foundation on which immersive VR is built, and while PS4 may lose out in horsepower compared to PC, the notion of developers targeting a fixed platform can't be understated - the VR experience on console will be validated by the developer on the exact same hardware you run at home, and indeed by Sony itself. In hitting and sustaining a performance target, that kind of advantage is priceless.
Also of interest were the team's thoughts on game integration. While first-person perspective is the obvious choice, the ITG has seen great success - and much easier implementation - through relatively simple ports of third-person titles. In effect, Morpheus could bring an IMAX-like impact to video gaming. Judicious use of third-person could also play a part in addressing the VR nausea issue too. Take, for example, a first-person racing game where you crash, sending your car spinning in the air. Keeping the in-car is guaranteed to be quite unpleasant, but adjusting the camera outside to a Burnout Takedown-style viewpoint could work really well. The castle demo also introduces a new challenge to developers - if there are objects in the world, the VR viewpoint practically demands the ability to reach out and interact with them. It's not just about convincing rendering, there are serious simulation challenges ahead here too, and providing guidance to game-makers is a key element of the ITG's remit.
Oculus VR vs Sony. It's a mouth-watering proposition: two entirely independent takes on virtual reality, designed with two different platforms in mind, but with a combined audience reach that could - hopefully - make VR a financially viable prospect for games developers. Oculus has the momentum, the bleeding-edge tech and the community backing, while Morpheus has a ready-made ecosystem, VR-friendly interface hardware, an international team of skilled hardware developers and proven first-party game-makers. After our hands-on experience, the potential here is intoxicating: virtual reality could well become the true next-generation in interactive entertainment that we've been waiting for.