The thumbnail image for this video has been updated - thumbnail added by mintbbb.

The thumbnail image for this video has been updated - thumbnail added by pumkinandstorm.

I'm not sure they need to do anything groundbreaking. Sometimes, it's just a combination of the right product, built from common components at the right price and put together with the right marketing.

Risking the ire of the apple haters here but look at the iPod. There wasn't really anything particularly special about it. There were plenty of other MP3 players around with similar (or better) specs at the time, but the iPod is the one that succeeded.

It could just be that the background level of technology has reached a place where it's now feasible to do decent HMDs.

Sorry, I was referring to the latency papers that @ChaosEngine posted. He mentions that their head-tracking sensor is capable of 2ms samples. That is indeed impressive for consumer level tracking. However he goes on to say that they use this "density of samples" to integrate a more accurate measure of position. In other words, they average over multiple temporal samples in order to achieve a more accurate spatial measure. However, if you do that, the samples you are using to produce a head position instantaneously is at least partially influenced by spatial information collected more than 2ms ago. When you do this too much you'll get sluggish head tracking which lags the head, and can often make people sick. However so can the alternative of low-latency though spatially noisy head tracking. That will just look jerky and spastic instead of sluggish. It's a balance, but his suggestion that they chose to average suggests that they needed to due to noise in their custom sensor. I have no doubt these guys have done a better job than previous attempts at low cost HMD systems. I just don't think they've done anything ground breaking, and even the cutting-edge technology I find unsatisfying.