Hands-on: All Four Project Morpheus Demos
Posted on Monday, March 24 @ 12:00:00 Eastern by blake_peterson
The most successful demo, which GameRevolution's Nick Tan reviewed in a manifesto to glowing acclaim, is CCP’s EVE: Valkyrie. Valkyrie is an in-development space dogfighter game that is, in my assessment, the first commercially viable game built from the ground up specifically for VR (another, earlier version of the demo was shown at the Oculus booth), and it shows. Though it lacks the immersion of The Castle, it doesn't have any of the glitches and is actually a good game. Though it has a low-res arcady look, it is still very immersive; and I leaned back into the seat as if feeling the G-forces as I was catapulted out into space, Battlestar Galactica-style.
Valkyrie’s amazing gameplay mechanic is that anything you can see in the center of your vision can be targeted with missiles. Machine guns only fire in the front, but having dispatched one enemy and flown through the center of its fiery explosive death (YEEHAH!), I craned my neck to find the next one by looking around and locking on, firing missiles while I pitched the fighter up to add machine gun fire to the mix. Whereas the previous tech demos are very clearly demonstrations, Valkyrie shows the potential of what a game can and will be in VR, and does so with a great deal of polish.
Valkyrie is also multiplayer, with two competitive head-to-head players surrounded by a number of bots, showing that the tech doesn’t impede multiplayer experiences. Where Valkyrie is less immersive than The Castle since there is no haptic interface (a game like this could bring a resurgence of old-school flight-stick controllers), it isn't a problem since the player and their avatar are seated in the cockpit the entire time, making it easy to sink into the reality.
More problematic, and practically a treatise on how conventional game development doesn’t work for VR, is Eidos/Square-Enix’s Thief demo. During Sony’s talk that revealed Project Morpheus, the speakers laid out a roadmap to successful VR experiences. Some of these are that “headtracking is king,” and that taking away control of the player view makes them sick; another is that high frame rate and low latency are priorities. Thief had problems with all three, sacrificing frame rate, low latency, and headtracking for higher-resolution graphics and visual fidelity. The result is incredibly mixed.
Standing still, Thief looks wonderfully moody, with a high degree of detail that every other demo lacks. It is the only game that looked like it belongs on a current-gen roster in terms of its visual style and level of detail. The attendant warned me to try to use the movement controls as much as possible, only to use my right stick for turning and not looking, but even doing this made me nauseous after a short while.
Thief’s low framerate—I heard offhand that it was locked at 30fps, and that VR needs the benchmark of 60fps in order to work properly—makes it look like watching someone play the game through a crappy webcam capture that's sickening the more motion there is in frame. Even using the controller to turn creates a slight sense of vertigo, and there seems to be a slight delay between when I try to move or look and the action, increasing the problems.
Visually, Thief looks great-ish, though it has the exact texture/shader problem that The Castle developers works to avoid. The stony environments have an odd particle-board look to them, of a plastic un-reality, like looking up-close at a plastic bench painted to look like stone. The flame effects also has a very two-dimensional look, like someone had a moving cardboard cutout of flames that lacks any dimensionality to them (though it did have some nice smoke and spark particle effects).
Thief, in trying to create a higher-definition look, succeeds in the appearance of greater fidelity in its 2D textures. But everything looks like a plastic toy-set painted to look real if you squint or shut one eye. It highlights a number of problems that VR presents the industry, since rendering the images separately for both eyes requires a great deal of processing power.
Essentially, VR appears to have an issue where it’s necessary to scale back the visuals to what can be ray-traced and lit dynamically in order for environments to look natural. The scaling back is a necessary expense due to the processor costs involved with rendering out frames for both eyes at the same time from the slightly different angles in order to look properly “real.” In this case, lower resolution but more realistic rendering clearly trumps detail. The “arcady” look of Valkyrie and the more basic environment of The Castle performs better than the artificial murk of The Deep and the hi-def-aiming Thief, due to scaled-back but more accurately rendered visuals.
This loss of detail might sound like a bad thing, but the experience of being in a virtual environment that works perfectly is so much more immersive than being in one that doesn’t but has higher detail. Ultimately, it comes down to what has the most fidelity of experience, and in a virtual experience the space feeling real is clearly much, much more important than a higher degree of detail. Additionally, the 1080p screen is split between both eyes, halving the horizontal resolution (so each eye gets 960x1080 pixels), which has less clarity than the average user might expect, allowing for less detail.
To give an idea about how this plays out on an individual level, I would probably buy Project Morpheus if it had low, coke-bottle-lens resolution in order to get to play a full version of EVE: Valkyrie. That the tech is only going to get better is just icing on that virtually real cake.
FOLLOW ME ON TWITTER. YOU KNOW YOU WANT TO.