[font=arial,helvetica,sans-serif]I'm sorry if this isn't appropriate for this forum but I figured with all the great minds here I might as well ask.[/font]

[font=arial,helvetica,sans-serif]I'm currently [color=#333333]to think of a focus for my independent study next semester at school. My major is Software Engineering so I have pretty much free-reign over what I can focus on but in the past I've done more computer graphics-related stuff such as volume marching, orbital simulations, ray tracing etc. (check my profile for a few examples). More recently I've been exploring the mobile platform with OpenGL as I got a tablet with a decent video chip (Tegra 3) but I'm drawing a serious blank on ideas at the moment in any of these areas.[/font]

[font=arial,helvetica,sans-serif][color=#333333]Any thoughts? I'm open to literally just about any realistic suggestion. My school's on the quarter schedule so I'd have approximately 10 weeks to work on it, but that shouldn't limit my options too much. It doesn't by any means have to be something that's never been done before either, it's more about you as a student exploring the domain.[/font]

Thanks for the response!

A lot of your previous projects seem to be in cutting-edge graphics. When it comes to mobile, that's really not where the game is at - a tablet is all about the touchscreen.

Consider anything that exploits multi-touch as a control surface.

That's what makes this so tough. From my point of view I don't see what I've done as too "cutting-edge". The ray-tracer was just a reproduction of Whitted that I toyed around with, and the volume marcher was me applying existing theory (that had been implemented much better before) in OpenGL for an independent project. Yeah I wrote both from scratch, but I had no previous education in the subjects really. I just saw something that looked interesting and read up on it enough to try it out. I'm more than willing to do the same again, but as I said with the new tablet I'd like to try to do something with it that HASN'T really been done before since everyone's playing catch-up. This is my first touch-screen device and I'm just completely oblivious to what I should be approaching with it. I know people have done fluid simulations with them before, so I don't know what I could really explore there. All I know for sure is that if I stay PC-oriented I'd do something somewhat advanced like my past projects, and if I stick with the tablet I'd like to try something that balances the relatively decent video chip in mine with the multi-touch surface. I apologize if this sounds like rambling but it's just that I'm really unclear on what I could / should do.

Realtime global illumination is the biggest area for improvement in realtime stuff, or animation but that's an area you'd need background in.

Textures? We've got fine grained texture streaming and can make tradeoffs between unique texturing, texel density, and disc space fairly effectively. Geometry? We've got tessellation, realtime impostors, a thousand parallax mapping techniques, etc. Materials? The biggest thing needed is more power and memory bandwidth, then we can get cook-torrance and subsurface scattering hacks and etc.

But light bounce, that's the biggest thing Hollywood has over realtime right now. Lightbounce is hard, big time complexity problem and crucial to looking not just real, but good.

GI is something I'm very interested in but I feel might be a little over my head. All I've done that is even remotely close to radiosity is raytracing which seems considerably simpler. Do you guys think that I'd be able to transition from virtually no experience in GI to competent enough in just under 10 weeks? Maybe I don't a clear idea of the domain.