Date: Tue, 21 Aug 2007 11:21:34 -0600
From: Mike Farnsworth
Subject: Re: Status update that might interest you
To: "Dr. Robert F. Erbacher" <Robert.Erbacher@usu.edu>
I haven't talked to you for a good while (since I presented that paper
at the GMVAG '05 conference with you about efficient renderer
architecture; if you recall, it was called "Global Illumination:
Efficient Renderer Design and Architecture"). I wanted to share a
success story that might inspire your students.
I've accepted a position on the R&D team at Tippett Studio, a movie
visual effects firm in Berkeley. You should go have a look at their
resume of movies at http://www.tippett.com/ to see what they've done.
Tippett is of Phil Tippett fame, the guy who was the lead animator for
Star Wars, Robocop, Dragonslayer, Jurassic Park, and others. He is
off directing movies now, but his wife runs the day-to-day operations.
The whole software development team supporting the artists/producers
at the studio is all of 20 guys, and the R&D team has 5 (including me
once I start), so it should be an impactful place to be for me to be. I
have been working at Headgate Studios, now EA Salt Lake, doing game
development. That is an awesome game studio, and working there
has been a blast, but this offer was too good to pass up.
I will get to work on graphics/asset pipeline tools for the artists, do
some UI work to make life easier on them, but of course I get to do
rendering and graphics research. One of the current things they are
working on is the LightSpeed Architecture paper, which was presented
at Siggraph in San Diego this year. There is a lot of RenderMan work,
and they're primarily a linux development house and they love open
source; they are giving a little bit of help to the author of Pixie, an
open-source RenderMan implementation, for example.
Much of the reason I got a job at Headgate/EA and now Tippett was due
to the paper you urged me to work on, so I wanted to say thanks for
pushing me and helping me get published. The renderer testbed that I
created for that has been improved, and I rewrote it in C# to make it
nice and clean so I could implement even better rendering algorithms
in it. I've got a blog that any of your graphics students (for when you're
teaching the graphics class) might like to peek at about that, at
http://renderspud.blogspot.com/ in case you are interested. It talks
about issues in implementing various rendering techniques, and the
pitfalls of all aspects of a ray-tracing rendering application;
RenderSpud continues to be a learning tool for me, and it is a lot of
fun to work on.