Oculus + unlimited detail (Euclideon)

raytracing in stereoscopic mean to "throw" ray for each eye's camera's pixel, right ?

More or less, all you are doing basically is calculating example paths how light can bounce around based on how it could bounce around based on the scenery. (Geometry, reflective properties, etc...) You have to do that over and over again that's why it takes so long.

In the most basic form you just do it backwards from every pixel of the camera(s) (it doesn't matter how many or where they are the principle stays the same) until you hit a light source. Then you take the color of the light source and multiply it by every color of every surface that the light bounced off from and you got part of the color of your pixel.
Then you repeat it until the image shows up. That is how naive path tracing works.

The main reason that is so inefficient is you have to do that several times per pixel because what I described above is just one of countless possible paths light can take from the light source to our pixel.

Everything after that are optimization where you try to avoid calculating light paths which have very little effect on the image in order to concentrate on the ones which do make a difference.

Anyway, from the raytracer i coded when i was at school, I remember that we had a possibility to increase performance
by limitating the "bounce/recursivity" of transparency/reflection.

I discovered something interesting, real time global illumination is possible right now! As I understand it it works by using the rasterizer to sample the rays. It's not as fast as rendering the old fashioned way but it looks a whole lot prettier and lightning is entirely plausible and physically based.




There are papers on the subject too: http://www.jp.square-enix.com/info/library/


another demo:

That does look pretty good, and at interactive framerates.

"geekmaster" wrote:

Please give me credit if you build upon my ideas.

Write and submit a technical whitepaper at the next Siggraph. Plenty of people would then see your work, as well as build upon it for everyone's benefit and give you the proper citations.

"geekmaster" wrote:

BTW. I would love to work with Mark Zuckerberg and crew, getting loaned to Oculus VR now and then. That would instantly pre-empt my current project.
:D

Following to my above thought, Johnny Chung Lee was a PhD student who had some cool/novel ideas about what you could easily do with the Wii remote. He released his research, concept videos, and code for free on his blog http://johnnylee.net/. This generated lots of web buzz for him, as well as giving others a jumping off point for their own ideas. He wasn't worried about "making bank", or what other might steal. It also showed people in high places that he could walk the walk, not just talk the talk. He went on to be hired by Microsoft to work on the first Kinetic team, then worked for Google X on Glass, and now works for the crazy Advanced Technology and Projects group at Google doing stuff like this: https://www.google.com/atap/projecttango/

My point being:

You are clearly a REALLY smart guy, who has thought about and researched lots of stuff. However, if even half the things you claim on this and others forums are true, you could never really get that deep into everything you want to do in one lifetime. And if you got hit by a bus tomorrow, no one would ever know you had done any of it, and lots of people would waste lots of time re-doing it all at some point.

If you took the time to fully explain and release code of some of your findings for others to use as open source, without worrying so much about "who might steal it", you might be surprised to find that there are really great companies out there who would be much more interested in the person who had all of those ideas than just stealing individual ones.

Not so worried about stealing ideas. Ideas are cheap. When an idea's time has come, it will emerge in the minds of many. I have published code in other forums under the MIT license so even companies can use it. I have lots of "proof-of-concept" code in VR, but not ready to publish. You are about fouth in line telling me to publish for SIGGRAPH, so perhaps I should, if time. Juggling a lot of balls here. This stuff has been my life, my job, and my hobby since about 50 years ago. I rarely sleep more than a few hours per night, and I often work up to three days nonstop in critical project phases, so I have had more than typical amounts of time to actually do this stuff. Testing has shown me to be smarter than average too, but my best code was all done in coding trances -- hard to do with WIMP interference (Window/Icon/Mouse/Pointer eye-hand coordination) consciously distracting me. Text and WordStar keyboard layout all the way! TurboC rocked! I have been told to start a video blog too, so I guess it is time to open up even more. ... and stop trying to reform the trolls...

I required a manager to come with me on interviews even when young, because my experience and abilities were unbelievable even back then... ...but most work sought me out based on my (pre-internet) "industry-wide" reputation. Of course, these days, stuff not found in the Google database never existed, eh?

I kew even as a young child that I had enough plans to last SIX lifetimes, and my plans have only grown since then. I began training an apprentice to help me THREE times in past decades, but they all got girlfriends or wives who were jealous of time they spent with me. Now some tell me I need disciples, not apprentices, but word are important a that sounds like dangerous terminology to me. Whatever we call it, I need a staff of bright minds to help me accomplish my dreams.

I need to figure out how to write a SIGGRAPH paper, and what to write about. Any suggestions?

Just wanted to clear up a few errors. First off. prior to sept 2014 Euclideon was not asking for money but was actively turning down investors since they received the Australia grant. And they had released a commercial product, just not for the gaming industry.

They now have a game engine and two games on the way to be released this year. Also this year, they're opening up to investors in the game industry. Shame it didn't happen years ago as graphics have improved exponentially while we've been waiting for a decent product. But here's hoping they don't suck anyway. Also nice to meet you all.

"Regless" wrote:
Just wanted to clear up a few errors. First off. prior to sept 2014 Euclideon was not asking for money but was actively turning down investors since they received the Australia grant. And they had released a commercial product, just not for the gaming industry.

They now have a game engine and two games on the way to be released this year. Also this year, they're opening up to investors in the game industry. Shame it didn't happen years ago as graphics have improved exponentially while we've been waiting for a decent product. But here's hoping they don't suck anyway. Also nice to meet you all.

Since I'm too lazy to look at what the company is doing now, has Unlimited Detail (Geospatial?) released anything for virtual reality or in particular the Oculus Rift?