pouët.net

I was actually doing my master thesis on volume rendering in OpenCL, was lazy and managed to do a simple raycaster but not any code/research on octree data structures as I was originally planning and because I realized that I have only less than three weeks to focus on a specific goal and write the thesis, I coded some realtime classic effects on the volume data instead and decided this will be my thesis. But me and my supervisor agree that the problem with this is what kind of literature to find about such effects?

So I am asking you, do you know any existing papers/thesis projects about the algorithms involved in demoscene specific effects? Especially those effects which you don't find in the regular computer graphics scientific literature. I remember there was one such paper where a scener presented bitmap manipulation/deformation effects or similar stuff I might not remember well. I can't remember where I found that last time.

I am not only in need but also curious about the papers written on this subject. Maybe if I can find that old paper or others I could also follow some past-references to scene effects (if any more) or more graphics related literature to this and find some more good stuff.

Be the first to write a scientific paper about it :)

Well, this guy wrote a thesis on raycasting using shaders. Maybe you can reference it. Though it's probably not exactly what you're looking for.

Something like this?

Burger, Boris; Paulovic, Ondrej; Hasan, Milos (2002): Realtime Visualization Methods in the Demoscene. Comenius University, Bratislava / Slovakia.

Online Version: http://www.cescg.org/CESCG-2002/BBurger/

sorry but why would they pass a thesis on "volume rendering of some, otherwise, uninteresting and useless 'classic' effects" in the first place...

I'd fail a thesis like that, also on account that you've only left yourself three weeks and still playing safe with your "hello, I'm 1994, give me back my plasmas".

I mean, c'mon, this is a computer graphics MSc, improvise god damn it :)

Come on, there is no time for something different. I actually also failed in the exams, so I don't care. I could go on next year just in case.

And somehow, I don't think I will not pass just for writting something different. I don't feel it's something so bad. There are thesis on "look I did this engine", I think at least. I think it would be ok.

I am also pissed with all these university thing and want to go back to my own stuff.

Boris Burger is seriously like the best name ever!

Even better than Curt Cool!

Or Dr. Dick and Mr. Sex!

Dont forget Dick Assman!
http://en.wikipedia.org/wiki/Dick_Assman

Hey this is optimus' plasma and looks really cool. What sort of framerate do you get on this? can you travel in/out of it? what about a sponge fractal and such... you can then make up a story about the shader being a good tool for testing performance against structures with set rules (i.e. intersections/ray or smthin)

Actually not much yet because it's just a slow test on CPU uploaded on 3D texture. I might port the code on OpenCL tonight.

structures with set rules? I don't get what this is

neither do I. But here is a nice pair of words for your academic work.

I may not include plasma3d. I could include perlin noise volume or something that is more in general literature.

Actually the thesis Sinar pointed is very recent and similar to the stuff I want, and I could take a read for some ideas, but I don't think I have much time anymore to decide for a different subject and more research.

This looks quite nice...like an update of the plasma cube from second reality. Would love to see that in motion.

perlin noise clouds is actually a fairly simple subject and you can get it done in a couple of days.

I would assume you'd do raymarching over the 3D perlin function and then add some palettes, animations, fogging etc. all easy, all manageable. I haven't done opencl but I assume it will do trilinear filtering for you.

Once you have all this you can talk until the cows come home about integration with the 3D graphics pipeline, blending, ordering, optimizations.

You can fill up at least 50 pages with all that shit in about a week tops. But better kill your internet connection first.

Man. People are fucked up here. In a good way. That's why I always come back.

Good luck with your thesis Optimus.

^someone had to do it ;-)

And also good luck with your thesis Optimouse.

robotski, that's one strangely organized article. first it gives a prosaic introduction to a range of effects - and then it goes into some what details on how to generate a train track ...

The title of your thesis will be: "On the metastability of multi-variate and self-describing Galois rings using Banach space operators as found in the demoscene."

oh, i would like to see a paper full of approximated infinite impulse response convergence in a bounded hilbert space using fixed precision arithmetic rings called.... fire effect. sure.

Wow, a thesis on scene effects?! Didn't know universities accepted that...! Quite open-minded, good luck Optimus with your thesis!