Palmer Luckey and Nate Mitchell Interview: Low Persistence ‘Fundamentally Changes the Oculus Rift Experience’ [video]

At CES 2014, Road to VR had the opportunity to meet with Palmer Luckey, co-founder of Oculus and inventor of the Oculus Rift, and Nate Mitchell, VP of Product at Oculus. We got our hands on the much anticipated ‘Crystal Cove’ Oculus Rift prototype and sat down to have the dynamic duo answer our curious inquiries.

Yesterday Paul and I got our first chance to try the Oculus Rift ‘Crystal Cove’ prototype featuring positional tracking and low-persistence display technology. While much of the hubbub from the press has been about the (very impressive) positional tracking, Luckey and Mitchell told us that low-persistence turned out to be an even bigger addition to the Oculus Rift experience than they had anticipated, saying that the technique would be an important fixture for virtual reality going forward.

ces 2014 oculus rift crystal cove prototype interview palmer luckey nate mitchell low persistence positional tracking

Low-persistence takes advantage of the new 1080p OLED panel that’s built into the Oculus Rift Crystal Cove prototype. OLED technology enables extremely low response times and extremely high refresh rates. While the two obviously couldn’t go into detail, they went so far as to say that they had “eliminated pixel switching time,” with the current panel. They also mention that this is not something that can be done off the shelf; low persistence on the Oculus Rift Crystal Cove prototype requires custom implementations across the hardware.

A full-persistence display has its pixels lit all the time. As Luckey explains, a full-persistence display only shows the correct scene orientation for one point in time; assuming the user continues to move their head, the scene orientation is out of date until the next frame can be drawn. The low-persistence technique lights the pixels only when the scene orientation is correct and goes dark immediately thereafter. Thanks to a high refresh rate, this happens so quickly that the user sees one continuous image. The end result is significantly reduced motion blur and potentially less nausea.

For detail on low-persistence, smearing, and strobing, Valve’s Michael Abrash has some great introductions:

In our demo session, Oculus showed us a build of EVE:Valkyrie which was designed to work with both positional tracking and low-persistence. They were able to toggle the low-persistence on and off to show us the difference that the technique made on motion blur (it was very significant).

While Oculus still says that they’ll be releasing a second Oculus Rift developer kit (DK2), they have no  announcements currently about timing.  Crystal Cove is just a prototype for now, but they say the DK2 will be as close to feature-matching  he consumer version as possible. When asked if they were still aiming for a 2014 consumer launch, they didn’t seem so sure, saying that their latest round of funding enables them to do whatever they need to make the best virtual reality headset.

We’ll have an in-depth article about our hands-on time with the Oculus Rift Crystal Cove prototype in the near future, stay tuned!

Comments

  1. Druss says

    The longer they take to release a consumer version the longer I have to (very very slowly) save up enough money to build a computer to give me the most awesome experience day one. The money for the Rift itself will soon be put aside (400 dollars including international shipping and a little extra just to be sure).

    • Druss says

      Do hope they stay at 1080p with the new, higher refresh rate. Else the PC required will be too expensive for the wide audience they are shooting for. So put the money in optimization and better tracking, not harder to drive screens. If a single mid-range card can’t run it you lose a huge market in one go, if a single high-end GPU can’t run it they might as well make it 700 dollars because any chance of reaching a big market is lost.

      • eyeandeye says

        I’ve been wondering about that myself. If they were to put a 4k OLED screen in the consumer model (assuming such a screen even exists), that seems like a huge leap in demand for current and near-future graphics hardware to suddenly have to render four times the pixels, at a smooth 120 fps (minimum, if I recall). I’d like to know if something like AMD’s Mantle would help. Supposedly the performance gains are pretty huge. But probably still not enough.

      • Brad Hawthorne says

        At issue is always going to be which comes first, the chicken or the egg? Does the hardware to push 4k at a proper FPS need to come first or do they need to just not compromise and do UHD OLED anyways? Whatever they do it’s going to be popular. One thing I will say with my personal experience of trying Crystal Cove at CES is that I’d personally prefer they do UHD. I just don’t expect that sort of pixel density in specialized OLED at a consumer price point yet. 5 years from now will be a very exciting time in VR. More exciting than it is right now. Hardware power is growing into where it needs to be in order to make all these things viable over the next year or two. There is a very aggressive evolution going on with VR, even within Oculuis VR with it’s prototypes. The consumer market drives demands you’d never see happening in the industrial or military side of things.

        • Druss says

          Brad Hawthorne says “Whatever they do it’s going to be popular.”

          I’m afraid not. Fact is, only a percentile of all gamers got to try the Rift on for themselves. So there is a big group that doesn’t have first hand experience, of that group most are on the fence about VR (because lack of vision (not blindness, the other one) or whatever). If the hardware is too hard to drive they would scare away more people then they would if they doubled the price point. It is crucial for the quick adoption of VR that they walk a fine line between reaching the biggest market available and delivering the best experience possible.

          In fact, quad 780Ti/Titans would not run 4K at 120Hz let alone the duplicate images the Rift needs for each eye. So making it would be fine, but releasing it as the consumer version would be disastrous. With the 4 expensive video cards, the extreme CPU needed to drive them, the huge powersupply needed to power them and the large case to house them. Not mentioning the wasted potential if not water-cooled and overclocked, are talking about a 5000+ dollar PC here that right now could not give a good experience. Maybe in a year you would get a much more powerfull PC for that money, but then we are still talking about a PC which non but the most hardcore of PC builders or a very rich person could afford.

          So I really hope that, just like in real life, the egg comes first and Oculus sticks to the 1080p 120Hz panel.

Leave a Reply