Anyone working in custom/in-house engines?

Absolutely :)

Well, not specifically for oculus, but in general yeah, 3D graphics code is my main interest, so this part of game development has always been very personal with me. I never touch another man's canned 3D engine :)

I've done a lot of graphics code and "3d engines" in the past, mostly for my own projects, but also occasionally in a professional capacity while working as a game programmer out there. Done software rendering, OpenGL, Direct3D, with or without shaders and stuff, mobile, etc.

So yeah, my plan to add oculus rift support in a couple of my hatching game projects involve manually working with the rift from within an OpenGL-based 3d engine of my own.

"nuclear" wrote:

So yeah, my plan to add oculus rift support in a couple of my hatching game projects involve manually working with the rift from within an OpenGL-based 3d engine of my own.

That's awesome :)

I'm curious with a question for you. I'm pretty well versed in C and somewhat in C++, and am currently working through OpenGL tutorials to get the 3D graphics bit down. How long have you been doing it? Curious how long you think it might take for someone like me to get a very simple 3D game/demo written (something along the lines of 3D pong, or walking around a cube room. Certainly nothing crazy like CoD or WoW :p)?

Yup, for sure! Created a little stub page like a year ago -

http://www.radioactivesoftware.com/content/oculus-rift-vr-development

Plan on updating it soon, for sure when I get DK2.

- Dan

I am working on a new pathtracer with Oculus support, the successor of Laguna.
(see http://www.geeks3d.com/20120502/laguna-real-time-opencl-path-tracer/ )
It will make fast pathtracing possible for dynamic scenes with minimal time needed to update the acceleration structure.
Can't wait to get my DK2 :)

Regards,

Arjan van Wijnen

I was thinking it would be cool if the community came together to produce a custom open source engine for VR, with an emphasis on high frame rate and resolution, 3d audio, VR UI, input, as well as ability to continually implement the latest oculus SDK.

The other engines are very cool, but things like fancy shaders, tricks to make things look better, etc, all have negative impacts on easily achieving high performance.

We have a custom engine we use for AirMech and will be messing around with some VR support for it. It's pretty lightweight, hopefully it's pretty straightforward to add Oculus support. :)

Not specifically for the Rift, but I'm writing a raytracing engine (in OpenCL) to explore some thoughts I have on computer vision and AI. The Rift would make a good platform for visualization.

A bit more about our engine, in case anyone was interested. We call it "Heat Engine" (so clever!) and it's multiplatform--everything from PC to Android to Xbox and PS4 all build out of the same sourcecode. The core bits are ours, but we use a ton of Open Source for things:

https://www.carbongames.com/opensource.html

For rendering we use bgfx: https://github.com/bkaradzic/bgfx

I think that a spectator mode with our game AirMech will be a great fit. We're super lowpoly and simple shaders, so the game runs crazy fast on decent hardware. It's nice having purpose built tech where you don't have any wasted performance overhead (that's not your fault at least).

Fan made video: https://www.youtube.com/watch?v=KE6DYx_a4lk

Just to add "one more yes" to the list .. "yes" ... I've been doing not one but "a number" of custom engines for the last 9 years or so at the point I could say "my job is to make graphic engines", Oculus been "another one" :)

What to say .. "the usual" .. "draw as much as you can with the fewer calls you can", "use as less as possible you can", "use vertex shader and pixel shader wisely and with care" and I could add " .. and you won't have other language other than C++" :)

Our motto was "60 fps or death" now with OR is starting to be "75 fps or death" ;)

I rarely used any sort of Open Source code .. just for the fact I normally find it "too tedious" to get the code and see how it works to use it rather than just writing what I need, by the time I spent 2 hours looking at things to make them work I think "I could have written the code in 1 hour" :mrgreen:

But yes sometime I looked at something "to get an idea" and then re-implemented it "my own way".

But yes .. "working on in-house engines" definitely.

You can add me to the list as well. Together with my friend, we are working on an open-source "in-house engine".

I don't want to share it yet, since we have a lot of reorganization to do (some leftovers from older times, when we thought we know how to code). Oculus support is one of the planned features, however we still have some work to do before even considering these planned features for real. Since we are students, our college projects are holding us up and causing huge delays...