Importance of 6 DoF Tracking for VR
I had some discussions in the past about how important 6 DoF tracking is for virtual reality. My last post about the FullHD version of the Oculus Rift started some more, both on the net as well as in person here at Siggraph. So let me write down some of my thoughts that don’t fit into a 140 character tweet.
What does 6 DoF mean? The current Rift devkit can track the rotation of the device by using various sensors. So there are three rotations that can be measured: yaw, pitch and roll. This are three degrees of freedom (DoF). If you would also track the translations or movements of the device, this would add another three (along the X, Y and Z axis): 6 DoF.
So how could this be done? In the past I have seen this kind of tracking using multiple Kinects, optical cameras and infrared cameras. Optical tracking can be done inside-out (a camera is attached to the Rift) or outside-in (an external, stationary camera looks at the Rift). In both cases you either try to get the movement from just what you see or you help the algorithms by providing well known optical markers (google ARToolkit to get an idea). Magnetic tracking (like it’s been used for the Razor Hydra (afaik) is another option.
So why do we want to track the translations? In my opinion, we have these motivations:
- Better immersion – the less constrained your movements are, the more you feel like you are actually in that virtual world.
- New player-world interactions – Think of a shooter where you can look around a buildings corner by just leaning to that side. No extra buttons needed…
- Less motions sickness.
While the first two points are nice and get us closer to perfect VR, we would still be not there with them. There are still other open problems like a still too low field of view, a too low resolution, no full body tracking etc. So here everyone can define for him/herself what is more important and should get a higher priority to be solved first.
The third point however is more interesting. In my opinion and both personal experience as well as experiences of people who I talked to, moving your head without getting the correct visual feedback can introduce or intensify motion sickness. One extreme example is Half-Life 2: I have no real problems playing this even when I strafe a lot. But as soon as a new level segment gets loaded and the screen freezes with the last rendered frame, I do get dizzy. Normally I am in a motion and suddenly the image I see does not match this motion anymore. In case you are more sensitive to this, even moving your head slightly (and unintentionally) without getting the correct feedback will increase your motion sickness.
This would suggest that solving the 6DoF problem should get a high priority right? Not only increases it the immersion and makes new interactions possible, it is needed to allow more people to experience VR in the first place. Well getting this right is not an easy task. As I already mentioned, I had the chance to test some VR solutions with full 6DoF tracking. The problem with all of these solutions was the precision of the tracking: if the motion detection is not perfect, it will introduce virtual movements that you see but that you never actually did! Even with this problem, it can and does increase the immersion and it does create new possibilities of interaction (up to actually walking around in your VR world). But: It does not solve the motion sickness problem when in fact it can increase the motion sickness. I can force myself (or learn to) not to move my head and only to rotate it but I can’t ignore wrong phantom movements that originate from bad tracking. Note that a too slow update rate of the translational tracking can create a noticeable lag that leads to similar problems even if the tracking itself is solid (video based tracking might not get faster than 30Hz depending on the cameras used for example, the Kinect is also limited to 30Hz).
So sum it up: adding 6 DoF to get a better immersion is nice, but IMHO, a higher resolution is more important to me – feel free to have a different priority. Adding 6 DoF on the other hand to reduce motion sickness is more important, but for that to work it has to be solid and I wouldn’t be surprised if we would have to wait for that some more time and not see this in the first Rift releases.
Update: Tom Forsyth from OculusVR noted a fourth reason: “Fishtank games” (think of Tycoon games, Sims, god games etc.). I agree, this is a good reason for 6 DoF tracking as being more free in your movements will make these games possible in the first place. I am not sure how well this will work but it it is definitely something we should try out. The tracking does not have to be perfect to make these kinds of applications possible (as long as you can handle a bit of wrong motion without getting sick).
In the end we have different motivations for adding full 6 DoF tracking each with different quality requirements. Eventually we will get high quality tracking and the other features required for good VR (like higher resolution, higher FoV, maybe light field displays in the (very) long run?) and everyone is free to set own priorities of what he/she is wanting more. I’m sure (not only) Oculus is working on all of there topics :-)
 read: I didn’t make a solid study not can I point you to one which tested this thesis