Got a new job a few weeks ago

[PalmerTech]

And we have several motion capture stages, several Wide5 HMDs, a custom VR engine built in Unity, and all manner of fun things related to that.

Anything in particular you guys want pics/reviews of? Right now, I am planning on posting some pictures and impressions of the Wide5, and maybe a video if we end up running a simulation tomorrow.

[cybereality]

Please, whatever you are allowed to share. Please do.

[PalmerTech]

Will do! There are some things I cannot talk about Sounds like you guessed that), but what I can show is pretty amazing.

What I can say for now is that 6 months ago, I was losing hope in VR. After seeing what I have seen, working with the stuff we have, I am so, so incredibly excited about VR. The future we want IS here, but it is all hidden away in research labs.

[mAchiNE]

Where do i sign up?
Sounds awesome Palmer, will be good to see what the mysterious Wide5 HMD is really like

[Likay]

Really cool. It seems like they also got the right man to join their team. Gratz!

[cybereality]

What I would be interested in is what you think of Unity as a platform for VR. I just started evaluating it recently, and it looks promising. At least for prototyping, I think its the best thing out there (that is affordable for hobbyists). But I kind of wonder about some of the lower level stuff. For example, I'm not sure it will be able to directly support native 3D using Nvidia 3D Vision or AMD HD3D. I am also looking at Ogre (which I have been using off-and-on for a bit) or even rolling my own engine from scratch w/ DirectX. The DIY solution interests me more from an educational standpoint, but I am not sure how productive that will be. So currently I am going to work with Unity until I find any roadblocks. Should be more than enough to at least produce a proof of concept.

[nrp]

What I would be interested in is what you think of Unity as a platform for VR. I just started evaluating it recently, and it looks promising. At least for prototyping, I think its the best thing out there (that is affordable for hobbyists). But I kind of wonder about some of the lower level stuff. For example, I'm not sure it will be able to directly support native 3D using Nvidia 3D Vision or AMD HD3D. I am also looking at Ogre (which I have been using off-and-on for a bit) or even rolling my own engine from scratch w/ DirectX. The DIY solution interests me more from an educational standpoint, but I am not sure how productive that will be. So currently I am going to work with Unity until I find any roadblocks. Should be more than enough to at least produce a proof of concept.

The trouble with Unity is that it's closed source and the free version has limitations. If you buy a Unity Pro license to access the additional APIs, anyone else who is going to hack on your code needs a Pro license as well. If you're not planning on releasing your code to anyone or you don't need anything covered by the Pro license, Unity is probably fine, but I prefer starting with something that has no such limitations.

I've used Panda3D and Ogre in the past, but there is other stuff like Irrlicht, or even OpenSceneGraph if you really want to roll your own.

[cybereality]

@nrp: Yes, that is a good point. Its so hard to make a decision. That is why I almost feel like starting from scratch with DirectX. It might be a lot of blood and sweat, but at least I won't have to deal with anybody's BS.

[PalmerTech]

@nrp: The place where I work has no interest in ever releasing their engine to the public, so they do not really care about having to use a Pro license. On top of that, most of the environments we run use gigabytes and gigabytes of RAM, and eat more more processing power than average computers can handle.

@Cyber: If you were able to make a VR wrapper/driver along the lines of IZ3D or TriDef with more VR hardware support, there WOULD be money in it, enormous money. I mean, I would hope it would be for enthusiasts, too, but a commercial use license for a VR driver that plugs into DirectX would have a lot of buyers right off the bat.

In terms of using Unity as en engine, it is pretty much perfect, from what our software guys have explained to me. I will elaborate more tomorrow, got into a car accident on the way home today, not up to it now. I did get some pictures on my 3DS, though (I know, terrible quality. It was all I had at the moment, and 3D!)

[cybereality]

@PalmerTech: Thanks man! If I do actually make a VR driver it would be a paid product. Otherwise I don't think I could spend the time/money needed to support it. And really, anybody that could use the software would need all sorts of crazy hardware anyway, so surely they could afford $50 (or whatever) for the driver.

In terms of Unity, I was playing with it more last night and I really love the setup a lot. Very quick to prototype. However I am already running into some limitations. For example, the software physics don't seem as fast as I would like. I did a quick test with about 100 cubes, and it did run alright, but I could already see slowdown. The demo I am trying to do requires seriously intensive physics, maybe 1,000's of objects interacting, and I am not sure it can cut it. I mean, even with a dedicated hardware PhysX card I am not sure I could realize what I want. But I am going to at least prototype the basic stuff in Unity and see how far I get.

Also, when you feel better please tell us more about the setup they have at your job, and how the simulation works.

[PalmerTech]

When I meant big money, I meant more than $50. For commercial use, you could probably charge thousands of dollars and still get buyers, hundreds of dollars and even more buyers.

Like I noted above, Unity can be pretty intensive. We use hardware that would be considered "Bleeding-edge I have no life but computers hyper enthusiast" grade.

Review of the Wide5 here, I will cover the simulation chambers later.

[ido]

Palmer... Great to see you in a good job.

I'm pretty jealous that you get to use the Wide5. I'd love to see a video of it in use. Hopefully, one day, the consumer level HMDs can achieve that kind of FOV.

[brantlew]

Cyber - It's a bit older but have you considered the id Tech 3 engine (Quake III)? It's a mature and high performance engine and the source code and several variants are all out there for free development. You would certainly have to add in the NVidia and AMD 3D support, but at least you won't have to reinvent the wheel with all the geometry, texturing, and visual effects.

[cybereality]

id uses OpenGL and all the support for Nvidia 3D Vision and AMD HD3D is DirectX, I believe.

[brantlew]

Yeah I guess you are right about that. It seems that there is some nVidia 3D OpenGL support through the Quadro card line, but it's not entirely clear to me (or to anybody else it seems) how well it is supported on the consumer cards. For what it's worth here's a little bit of discussion...
http://forums.nvidia.com/index.php?showtopic=198902
http://forums.nvidia.com/index.php?showtopic=206673

[cybereality]

Quadro and 3D Vision Pro is something totally different, and I have neither the hardware nor the desire to support that. That is mainly for professional applications (scientific, etc.) not for gaming.