Tuesday, 2 February 2010
Just a little experiment I had after I watched a similar thing (but with much less points) in a NDS demo. I always avoided trying this because I thought it would be hard to visualize a 3d plasma since the inside would be occluded (although I thought then about alpha too) and it would be hard for the viewer to recognize and enjoy the 3d effect. But seeing that the NDS demo gave me somehow the impression that if a box of 16*16*16 non-blended thick pixels with 3d plasma coloring would work, then this experiment with more and blended would work too. Their version also reminded me of all these videos.
My next curiosity is to try out and see how the 2d water ripples version would work out in 3d. I am very curious about this. Probably like expanding spherical ripples that fade out by time. And maybe I could use the final result to also do displacement with the background? (because that's that's the standard thing I did in the old 2d effect too). How would that work? Also, I have first to add the effect of alpha per dot, because now all dots are additively blended by the same tiny percentage on the screen buffer (I also don't sort the points from back to front) and apply it on the plasma so that you can see some plasma bands appearing and some not being there, instead of a full blown box with more color in the middle. I am not sure about the result too but I will try..
p.s. Btw, my final project in my MSc will have to do with volume rendering in OpenCL and possibly fluid dynamics. I love those stuff and I hope I will have something here soon with nice teaser screenshots (and maybe youtube videos).