Today was not very exciting in comparison with the first two days. I had trouble figuring out which talks I wanted to go to and unlike the previous days, it wasn’t because there were too many good things to goto. It is probably because a lot of the talks today did not make too much sense to me. I went to the “Real-Time Rendering” session and the most interesting talk from that session was one on rendering smoke in real-time. The one on Global Illumination was also pretty interesting but a little over my head.
The next session was hard to decide. I ended up going to a session called “Particle Man” where they talked about how various particle effects were used in production movies. People from Dreamworks and Rhythm and Hues talked about the different stages of particle effects. It was pretty interesting but a little too fuzy for me. I left after the third talk and went to the last talk of “Faces and Reflectance.” I was hoping this paper on capturing material reflectance from a single camera view would be cool but I think I was too tired to really appreciate it.
Joel and I wandered around the Exhibition booth for a while and I eventually went to the session on “Global Illumination.” I was looking forward to this one a lot but I didn’t get much from the talks. I wrote down some of the neat ideas from the talks but I think I’ll need to read the papers to fully absorb what was going on. The first talk was on Radiance Caching for Participating Media which was pretty similar to my nebula project. I’m going to follow up and check this paper out.
I ended the day at the “Many Things” talk which I liked the most. I was originally planning on going to “The Game Show.” My brother gave me a recording of the one from GDC and it was pretty funny. When I got the room, it was pretty empty so I figured it wouldn’t be that good. I actually enjoyed the “Many Things” talks. Two of them were from Pixar talking about making the shaders for the many robots in Wall-E and also animating the crowds in Wall-E. Pixar has really got it down. They talk about the aspects of realistic movement of the future: hover, anticipation, and reaction. For both the shaders and animation, they ask an artist or animator to first create the ideal shader or animation. They then use procedural models based on these idealized shaders and animations and perturb them for the masses.
The third talk was about using crowd control software to animate the racecars in Speed Racer which was also really neat. They used a similar approach to Pixar’s in having an ideal artistic model that gets perturbed. They both use a system called Massive for crowd control. The system seemed to have an interesting interface. For instance, characters in Massive has specific ways they can interact with the world around them (like Karel but with more beepers). The artist could generate rules based on what it sensed around them. Of course, it was much more complicated but you get the idea. For Speed Racer, they would pain parts of the track blue to let cars know they should be breaking. They would make the track a gradient of red and black to indicate the distance to the edge of the track.
Sorry, this is pretty scatter-brain because I haven’t had too much sleep. Also, today was hectic for me as my indecisiveness got the best of me. I’ll leave with a quote from one of the Pixar guys, describing what a stop is to the physics people:
A stop is known as Big Negative Translational Acceleration.
For some reason, that makes me smile.
Leave a comment