Oculus Rift Support

[Adrianis]

Game programmer and Rift dev kit owner here. The Oculus SDK is pretty darn close to plug-and-play. The SDK gives you the head tracking input from the sensor fusion in the form of a transform matrix that you just apply to your camera's view matrix prior to rendering and they have documentation on the algorithm used to create the distortion effect needed to work with the lens distortion in the headset. It really is pretty straightforward - maybe 1-2 days to get it working properly with both tracking and rendering. The vast majority of the work in integrating Rift support for a game is typically related to HUD and controls, which is relatively simple in this case. This talk about the engine needing to be rewritten or even significantly reworked is greatly exaggerated. Posting because I would personally love to see this happen, although I would not blame them for holding off until the consumer version of the OR is closer to release - especially if adoption rates aren't looking great.

Check this page out
https://github.com/FrictionalGames/Penum...teract.cpp
Specifically 'cPlayerState_Grab::OnUpdate'

That's the Penumbra gameplay code in HPL1, specifically the object interaction. Assuming that the interaction in HPL2 is anything like in HPL1, the basis for moving objects that you are holding is the Camera position. It may be trivial to get the SDK implemented and tracking/rendering working, but that's the smallest and easiest part.

The hardest part is going to be redesigning the object interaction. Assuming you've played this game quite a lot, you know how important the feel of grabbing & moving objects around is. Separating the camera for head movement and the body for moving objects may not be a huge task (not trivial either), but think about the kind of consequences you would need to deal with for things as simple as knowing which way you're body is facing, since you're going to need to move things accurately, and it needs to feel natural.

I haven't picked through all the relevant code in HPL1, but I'm guessing they didn't build in the potential for a separate 'head' and 'body', so now it's not just picking up boxes that's the issue, never mind drawers & doors & wheels, but actually the way your 'body' is represented in the game world will need to be re-thought (if you're grabbing from the eyes as before it wouldn't feel right, but if you're grabbing from the body would the start position for grabbing an object be further down, what would that mean in the code above? Would that still work if it's a wheel you're turning?). I actually can't think of an interaction system that would be any harder to get working right with the Rift than the kind Amnesia uses.

Obviously the whole thing doesn't need re-writing, Statyk already did a good job pointing that out, but a good solution will be a lot more than a few days work. The design issues are fun to tackle, but when these guys need to make a living out of their development time, spending a lot of time thinking through UI implications for a niche hardware product doesn't make sense. Leave it to the hobbyists till there's a decent standard to work from. I'm not sure how you feel but there's no project on the Rift development showcase I've tried that I would point to and say 'that's how to do it'