Oculus Rift support

[DRayX]

Any update on this?

[biolizard89]

(05-30-2013, 05:51 AM)DRayX Wrote: Any update on this?

I've been working on it a bit, but nothing to report yet.

[braindx]

Sorry everyone. I've been pretty busy lately and this kind of dropped off the radar.

As mentioned before, the compatibility is pretty low. I'm going to see if I can just translate the image in screen space for the IPD rather than apply a translation to the projection matrix because I think that will cause far fewer issues.

Also some games (e.g. Goldeneye) use a really low FOV perspective matrix when rendering UI rather than an orthographic matrix. This means that if you force it to a large FOV you just get really tiny (unusable) menus. So, I'm going to hack something in so that it detects a really low perspective FOV and does something else (not sure what yet, maybe just leaves it untouched).

[garrlker]

Letting you both know, I greatly appreciate what ya'll are doing. I plan on buying an oculus rift just for this.

[biolizard89]

(05-30-2013, 08:29 AM)braindx Wrote: Sorry everyone. I've been pretty busy lately and this kind of dropped off the radar.

As mentioned before, the compatibility is pretty low. I'm going to see if I can just translate the image in screen space for the IPD rather than apply a translation to the projection matrix because I think that will cause far fewer issues.

Also some games (e.g. Goldeneye) use a really low FOV perspective matrix when rendering UI rather than an orthographic matrix. This means that if you force it to a large FOV you just get really tiny (unusable) menus. So, I'm going to hack something in so that it detects a really low perspective FOV and does something else (not sure what yet, maybe just leaves it untouched).

My linear algebra skills are lacking, am I correct in assuming that translating the projection matrix in the X direction is equivalent to translating the image in screen space, except that there will be a small black space on the side where the image wasn't rendered? If that's correct, I'm probably going to go that route as well, because it seems that the GameCube/Wii GPU doesn't support X translations of the projection matrix.

Progress update: I'm in the process of modifying the GeckoTunnel script format to handle head tracking data separately from player stats. Unfortunately at the moment when I run the code on F-Zero GX the Wii just crashes. I'm pretty confident that I just made a minor stupid mistake when adjusting the code, so most likely I'll have to disable the code and then start re-enabling it line-by-line until I see which line of my code causes the crash. Normally this wouldn't take much time/effort, but due to the fact that I'm attending a robotics conference in early July and I need to have stuff ready for that, plus some recent medical issues in the family, I haven't had much time as of late. But, I really want this project to succeed, so I'll try to find some time. Also, my brother's Rift order just transitioned from "Ready" to "Processing", meaning that most likely I'll have access to a physical Rift within a couple weeks.

[AnyOldName3]

Wait, you've been doing this without one so far?

[braindx]

(06-03-2013, 04:41 AM)biolizard89 Wrote: My linear algebra skills are lacking, am I correct in assuming that translating the projection matrix in the X direction is equivalent to translating the image in screen space, except that there will be a small black space on the side where the image wasn't rendered? If that's correct, I'm probably going to go that route as well, because it seems that the GameCube/Wii GPU doesn't support X translations of the projection matrix.

Progress update: I'm in the process of modifying the GeckoTunnel script format to handle head tracking data separately from player stats. Unfortunately at the moment when I run the code on F-Zero GX the Wii just crashes. I'm pretty confident that I just made a minor stupid mistake when adjusting the code, so most likely I'll have to disable the code and then start re-enabling it line-by-line until I see which line of my code causes the crash. Normally this wouldn't take much time/effort, but due to the fact that I'm attending a robotics conference in early July and I need to have stuff ready for that, plus some recent medical issues in the family, I haven't had much time as of late. But, I really want this project to succeed, so I'll try to find some time. Also, my brother's Rift order just transitioned from "Ready" to "Processing", meaning that most likely I'll have access to a physical Rift within a couple weeks.

Yes, that is correct. A translation of the projection matrix is equivalent to a screen-space translation of the scene. So, if you are rendering side-by-side images, with each side having a translation of 1 unit on the X-axis away from the center of the screen, the two renders would seamlessly connect in the center and make one large image.

[biolizard89]

(06-03-2013, 08:30 AM)AnyOldName3 Wrote: Wait, you've been doing this without one so far?

That's correct; the stuff I've been doing so far is testable using Stereoscopic Player and a Nvidia 3D Vision display, which I have. I'm quickly nearing the end of the phase where I can do useful stuff without a Rift, so I'm happy that the Rift is arriving soon.

(06-03-2013, 08:42 AM)braindx Wrote: Yes, that is correct. A translation of the projection matrix is equivalent to a screen-space translation of the scene. So, if you are rendering side-by-side images, with each side having a translation of 1 unit on the X-axis away from the center of the screen, the two renders would seamlessly connect in the center and make one large image.

Excellent, that's what I figured. Thanks for the confirmation.

[ExtremeDude2]

Lead Oculus VR Engineer Andrew Reisse Dies in a Senseless Accident:

http://www.maximumpc.com/article/news/lead_oculus_vr_engineer_andrew_reisse_dies_senseless_accident123

[Zee530]

WTH?!