It had been a while since I wore an Oculus Rift, and it was time to peek at the latest version of the hardware SDK, called Crystal Cove. I had missed out on the previous evolution, but it looks like this latest revision is much more interesting anyway. Oculus VR has worked hard to fix every pain points that it could, such as latency, response time and other critical aspects of its virtual reality (VR) headset. I’ve been following VR since it was using non-textured 3D graphics, so how far have we come? Read on.
When I tried the original Oculus Rift at CES 2013, it was very nice to see a company taking on this problem that was all abandoned. Sure, there had been progress since the nineties, but the typical problems were still there: the head tracking needed work, the displays were low-resolution and the latency was quite important. The big win was the immersion: with the first Oculus, there was virtually no black border and the immersion was great.
Move forward to GDC 2014. Oculus Crystal Cove now features high resolution AMOLED displays with a response time of 2ms (vs. 15ms) which eliminates the ghosting effect from past hardware.
In addition to this, Oculus had added head position tracking system which is based on a camera mounted on the screen or anywhere in front of the player and LED lights on the headset itself to make sure that it can be tracked as accurately as possible. These changes are the base of all the progress that was made since the last revision of Oculus Rift.
The head position tracking is completely critical because it helps the rendering engine match what you see with what your brain expects. For example, as I move forward, things get closer, and I can also lean and look at what’s around my virtual chair. This is a big deal and it was missing from earlier Oculus hardware. I love it.
At the moment there is no body tracking, so the system is unaware of what your body is doing. The camera can still track the head motion, so if you bend your knees, the camera will go lower, but the app doesn’t know “why” it goes lower. Depending on the context, this may or may not be an issue.
The industrial design is still a work in progress. It surely looks more polished than the initial one I’ve tried last year, but this is probably not how it would end up as a consumer device so I would advise not to judge the device by its looks.
I think that things are small enough to allow for a nice “final” design, but with budget restrictions, we will see how creative Oculus gets when the time comes. I’m confident that this will not be an issue. At the moment, the headset is easy to put on, and I didn’t need any supervision from the Oculus staff to get going.
With a weight of 1lbs, the Oculus Rift Crystal Cove is not too heavy on the head, but it was substantial enough to make me want to avoid moving my head too fast. With VR games, it’s usually not a good thing to move your head too fast anyway (that’s how you get motion sickness), so this wasn’t really a problem, at least with the demos that were playable at GDC14.
The whole setup seemed comfortable enough, and given that this is a prototype, I expect a retail version to be better. We’ll see. Since I have used it for 10mns or so, it’s hard to tell how it would feel if used for a couple of hours, but right there, it felt very good. I’ve mostly not been sick with VR headsets, so I would say that I’m one of the better potential customers. Your luck may vary, but you will never know before you try. At GDC, some people loved it and others complained of being sick.
The AMOLED display makes a huge difference when it comes to image quality. The first Oculus Rift was so low-resolution that everything seemed a little blurry (like a Nintendo GameCube).
The image is now clearer, but in the demo that I was playing, it felt like I was looking at a 3D lenticular display in which the image had this typical glasses-less 3D TV feel – very much like a Nintendo 3DS. I was surprised, and this makes the overall image quality seem less than the 960×1080 pixel per eye should provide.
There are three components in perceived latency when using VR headsets:
The display seemed to be very responsive and there was no motion-blur during normal play time. I did not refrain from moving my head to look around me, but I didn’t go into full shaky mode on purpose either. To me, the display responsiveness seemed adequate for the task at hand.
The head tracking was quite good. There was a lag, but it was quite low, and overall, the image matched my head motion pretty well. I have to say that this was pretty fun and when things work well enough – only then you can be comfortable and be immersed into the application/game itself.
The rendering speed is highly dependent on the game content, so in my case, the demo was simple enough to allow for maximum speed rendering I think. Interestingly, even if the game was being rendered at 60FPS (I’m not sure it was), I had a feeling that things were running at 30FPS. It may have been that the head tracking induces some lag, but thing felt slower than if I had a mouse to control my virtual head rotation.
At the moment, games are still being built, but at the moment, I can see a few categories of games that would benefit from something like Oculus VR. They have these attributes:
High framerate is not really a problem, as long as you don’t have to move your head too fast. Since the system and display can now handle high-resolution and high FPS content, I suspect that the head-tracking would be the main source of lag.
I can easily imagine playing simulators of all kinds (flight sim, car race, space combat…), or adventure games like Skyrim, WoW and others that don’t require fast action like a Battlefield of CoD would.
Most top-view games like real-time strategy (RTS) games should look quite fabulous as well. Most likely, games will be tuned for this particular experience, and I also hope to see exploration games, education games or apps whose gameplay is based on the fact that you have to look around.
Oculus has made great progress with the Oculus Rift Crystal Cove, and nothing superfluous was added in this new version. Higher display resolution, fast screen response and head position tracking are all critical pieces of the VR puzzle. Overall, I was very satisfied with the head tracking and screen response. However, I was also a bit surprised that the screen didn’t look as “high-resolution” as I would have thought – this is significant.
Although the overall experience was much better, I felt that the field of view had been reduced and I hope that it will go back to a border-less one when the product comes out. With the first Oculus Rift, I could virtually not see any black border around the image. With this version, I could clearly see it.
I’m not sure how much Oculus can tweak this before this becomes a consumer product, but I’ve been assured by the staff that everything would be made to have this as good as possible. I’m looking forward to seeing what the next step will be like.
|Display||960×1080, 75Hz, 2ms|
|Field of View||100 degrees|
|Headset USB||USB 2.0|
|Camera USB||USB 2.0|
|HDMI Input||HDMI 1.4b|
|Motion Sensors||Gyroscope, Accelerometer, Magnetometer|
|Sensors Refresh Rate||1000 Hz|
|Tracking Sensor||IR CMOS Sensor|
|Tracking Refresh Rate||60 Hz|
|Weight||0.97lbs (without cable)|