GDC 2013: Michael Abrash on “Why VR is Hard (and where it might be going)” | Hacker News Reader

GDC 2013: Michael Abrash on “Why VR is Hard (and where it might be going)”

roadtovr.com

66 pointsstaunch17 comments

Excellent talk. I got my rift dev kit on Friday and I can easily see this being the future of gaming! I haven't had time to really dive in to the SDK yet, but simply watching a few videos tailored for the rift was an incredible, highly immersive experience.

That said, one of the videos makes me terribly motion sick, so... I'm hoping my body can get used to this.

There was a lot of exciting VR news out of GDC 2013 if you were paying attention and knew where to look, especially this:

http://www.roadtovr.com/2013/03/30/gdc-2013-oculus-rift-raze...

For anybody interested, I've got a browser plugin ready for streaming the sensor data, and will finish the implementation as soon as my kit arrives

https://github.com/grimwire/ohmd-plugin

Could you try out something I've modified to work on canvas on Firefox Mobile? Are you driving the standard HTML orientation variables? http://ljd2.com/nemesis/indexcanvas.html

That's interesting; I'll put it on the roadmap if it maps well enough.

I think the idea of VR is beginning to shift: head-mounted displays are clunky, a much more immersive solution is to project the virtual world on a large curved screen or a wall, and then use array microphones, and a set of cameras for tracking user motion.

I worked on a project like this a while ago: https://www.youtube.com/watch?v=lVVKNTP0lfo

There are a few places where this research is going on, USC's ICT also has a really nifty setup, a lot like HoloDeck v0.0001 from Star Trek: https://www.youtube.com/watch?v=tcTswP0sF9w

Taking you literally, I'd say that a wall is clunkier than a headset. But being reasonable, I'd say that those approaches are pretty different and neither is obviously better. Why do you favor projections?

The a headset and projecting on a wall are fundamentally different experiences and are each valuable for different use cases.

If I were to plan the battle on Endor or Hoth, yeah a wall project makes much more sense, but if I want to feel like I'm actually on Endor or Hoth, then a headset makes more sense.

The only way a projection is going to make sense is if people have an entire IMAX shaped room in their own home. And even then you still are going to have gaps in the experience like looking at the floor or the chair you're sitting in. The headset divorces you from your body and actual surroundings in a way that no projection can.

This is great and really forward looking. I have been working on an AR project on the side for almost two years now and the things Abrash points out as problems are core issues that we are trying to fix (without using real world markers).

Of the three core problems that he points out, tracking really stands out as the key differentiation for us. Adding to that, how the objects must remain fixed in space is really the factor that if solved would plunge us right into next generation AR. Improved latency is almost a nice to have at this point.

Awesome talk. I was just thinking about AR and it's possibilities from this invasive code post ( http://www.verious.com/board/iNVASIVECODE/augmented-reality/). It seemed like the first salvo towards VR, but it's good to hear they're progressing in unison. Will Doom 4 impress us to the next level?

I got my Rift dev kit on Friday, and my weekend was expected occupied.

I"m not sure if I want Doom 4 in VR. Even in the very rough, early, low-resolution state of the Rift there is an uncanny realism to it - I can easily imagine people literally getting scared to death in horror games.

Games like Amnesia are scary enough on a flat screen in a lit room, there may be real issues with putting them in VR.

There's a thought. Gaming so realistic people stop playing them.

Did you try playing Team Fortress 2 with the Oculus Rift?

Yeah, it's pretty much the best possible demo of the Rift's abilities. The Tuscany demo in the SDK is IMO not great for showing the full promise of the technology - everything is static and the level geometry means you never really get close to anything for the depth perception to punch you in the face.

TF2 has all of that and more. My score has taken a punishment, but it's just so damn enjoyable.

I'm also that ass on your team that's walking around looking at stuff in amazement instead of shooting the enemy...

The space-time diagrams are a great visualization tool for various display artifacts.

Great! Was already considering buying GDC Vault annual fee of 500$ to just see this presentation... let's see if the read can spare me this money. :-)

roadtovr appears to be taking a beating. Mirrors?

The tracking problem seems easy to solve; but admitedly I never went to college so this may be just dumb, but why don't just use a external reference device? A little device that saves your initial position and triangulates the diff between your initial position and your current position.

Reply on news.ycombinator.com