Overview

On the week of July 26th - August 1st 2003, nine people from Florida State University's School of Computational Science & Information Technology (eight of them in the Vis Lab run by Dr. David Banks) went to the ACM's SIGGRAPH 2003 in San Diego, California. Those people are: Christopher Baker, Kevin Beason (me), Wilfredo Blanco, Brian Bouta, Neil Druckmann, Micha Faw, Brad Futch, Matt Scragg, and Hui Song.

Day 1 - Saturday July 26th

Five minutes into the road to Jacksonville, Chris' van broke down. The radiator cap was missing. We all stopped for an hour until we sorted things out -- they switched to Chris' wife's car, and we all continued non-stop to Jacksonville.

At the Atlanta connection, Brian and Chris nearly missed the connecting flight because they lost track of time. Similarly, there was some confusion over Brad's bar tab and I covered him after he left, while I arguably should have left his bar tab unpaid, since it was a ripoff and the service was the worst.

When we re-boarded the plane Wilfredo found out the cleaning crew disposed of his glasses.

After a four hour flight, we arrived in San Diego and ate at a Mexican-run diner near our hotel. It was expensive (~$14/person), but this turns out to be normal for San Diego. There also was some confusion about how far the diner was from our hotel. The hostess said 7 blocks, a drunk next to the diner said 10 more blocks. All in all it was 20 minutes from the hotel by foot.

Day 2 - Sunday July 27th

There was a course called Interactive Geometric and Scientific Computations Using Graphics Hardware all day, but this was pretty boring so I left and returned later for the section "Ray Tracing and Global Illumination Using Graphics Hardware" (3:30-4:15p), by Tim Purcell from Stanford University.

After that, Matt and I met up with Kate Kligman and went downtown to Alambres, a Mexican restaurant for lunch. On the way there, we bumped into Josh Grant, a graduate from FSU's Vis Lab and now an employee at Pixar. On the way back, we bumped into Elizabeth Bellenot, a friend of mine from FSU who was there with her father, Steven Bellenot, who is a professor in FSU's Mathematics department.

At 6:00pm Matt and I attended a Special Event, the Fast-Forward Papers Preview. This was mildly useful, but there were several no-shows, and this actually didn’t cause me to go to any extra papers.

For dinner, Matt and I ate at Ricki's next to our Hotel.

Day 3 - Monday July 28th

Josh, Kate, Lai (Hui's friend from Missouri), and the entire Vis Lab group ate lunch together at China Too, a downtown restaurant.

From 1:45-5:30 I attended the course Global Illumination for Interactive Applications and High-Quality Animations. Afterwards, I talked to the following people:

• 
  Cyrille Damez, a course speaker from Max-Planck-Institut für Informatik. I asked Dr. Damez what the most efficient representation of a vector in g++ is. He (I think) told me that most people use their own class for vectors, and that the main thing I should do is not use inherited classes, use inlining, and let the compiler handle optimization.
• Ingo Wald, a course speaker from Universität des Saarlandes. Dr. Wald showed me his real-time ray tracer, in particular the RQMC checkbox (little effect in his demo), and explained Virtual Point Lights (used for instant radiosity).
• 
  Philipp Slusallek, a course speaker also from Universität des Saarlandes. Dr. Slusallek said intrinsics were important for speed (by this I think he means inlining).
• 
  Per H. Christensen, a course speaker and senior developer at Pixar. I told Dr. Christensen that I read his paper on Faster Photon Map Global Illumination and asked if I understood it correctly, and he confirmed.

Henrik Wann Jensen said hey to me. I said hello and followed him into a RenderMan course but stopped short of sitting with him, because I didn't want to be annoying.

From 5:45 - 7:00pm I went to the Papers session Precomputed Radiance Transfer, which was good but still difficult to understand, despite having read the papers beforehand.

From 7:30 - 9:30pm, the CSIT group attended a Special Session: Creatures, Critters & Clones: Styles and Techniques Unique to Industrial Light + Magic. It was okay.

Day 4 - Tuesday July 29th

• 
  Henrik Wann Jensen, from University of California, San Diego. I asked Dr. Jensen how he achieved his quick rendering times. He shrugged and noted he wrote his first irradiance cache 10 years ago, which I think means two things: Irradiance caching and ray tracing coding experience.
• 
  Alexander Keller, from Universität Kaiserslautern. I told him I used his SamplePack for RQMC sampling, but was not having any luck. He says it takes a lot of time, and that the two days I spent was not enough. I need to try it for a week, he says.
• Art Owen, a statistician from Stanford University. I explained my RQMC sampling woes, and that I read a paper of his. With only a rough description of my problem, he suggested I make sure I use a power of two for the number of samples I use (when using a Sobol sequence).
• 
  James Arvo, from California Institute of Technology. Dr. Arvo, at my prompting, explained that a point set that is evenly stratified over a triangle, when projected onto a sphere, does not evenly stratify the projection.
• 
  Philip Dutré, from Katholieke Universiteit Leuven gave me a coupon for a discount on his book (which I bought and he signed, see below).
• 
  Matt Pharr, from NVIDIA. I asked Mr. Pharr if the second path mutation strategy he mentioned in his Metropolis Light Transport talk would handle lighting through a grate, and he said it depends on if the grate's thickness is smaller than the jump parameter, etc.

I also met Benedict Brown, a student from Princeton studying applying photon mapping to sound waves. I also met his instructor,
Thomas Funkhouser.

At 7:00pm the CSIT group went to the Electronic Theatre presentation. I sat in the 4th row with Matt and Chris, we had good seats. The presentation was impressive and entertaining, except for the last clip, which was sleep-inducing.

We almost went back to hotel but I got everyone (Hui, Micha, Chris, Matt, and I) off the bus and we went to Rockbottom Brewery for drinks. Later we were joined by Brad and Wilfredo.

Day 5 - Wednesday July 30th

Everyone went to the San Diego Zoo until 3:00pm. I saw mother and child hippopotami, a polar bear, a panda, a lion, some hyenas, two Alaskan brown bears, some sun bears (think Winnie the Pooh), some chimpanzees, snakes, turtles, some elephants, a rhinoceros, some sleeping koalas, and more.

For lunch, Brad, Wilfredo, Matt, Neil, and I ate at The Field, a great Irish restaurant downtown.

After lunch Matt and I checked out the Exhibition.

At 8:00pm we all went to the Papers Reception in the Sails Pavilion, where free food was had. Josh, Lai, and Kate were also there.

After that several of us went for drinks at Rockbottom Brewery (again): Wilfredo, Neil, Rick, Matt, Chris, Brad, Brian, Micha.

Day 6 - Thursday July 31st

Some interesting products:

• Point Grey Researchs bumblebee, a depth sensing stereo camera
• Rose Brand's Chromatte, a glass curtain for video matting
• Vrex's passive (polarized) stereoscopic displays
• Keyhole's EarthViewer 3D, a 3D model of the earth streamed over the net
• NVIDIA's Dawn demo (hey, it's pretty) on a niiice NEC LCD3000 30″ LCD display.
• Air²'s Vapir Digital Air, a digital smokeless vaporizer (pipe). By the way, I don't smoke, I just thought this was weird.

From 1:00-3:00pm, we saw a Special Session called: Finding Nemo: Story, Art, Technology, and Triage. For some reason, people cheered when it was announced this session was not a technical How-To. Personally, I was disappointed with both special sessions because they were not How-To's or creation documentaries.

At 5:00pm, I bought Realistic Ray Tracing (2nd Edition) by Peter Shirley and Keith Morley (down), and also Advanced Global Illumination by Philip Dutré (whom I talked to, as mentioned earlier), Philippe Bekaert (down), and Kavita Bala, the three of which signed my copy.

Next I went to the 3:45pm Papers session Computation on GPUs, but it was boring so I left early.

That night we took group photos at the conference center and then had dinner at Hooters. Hui claimed it was her birthday and danced with the Hooters girls, and we had a $335 tab.

After dinner, Matt and I met with my highschool friend, Bryan Blank, now Chief of Security at Sony Online Entertainment. He took us to a hilltop overlooking San Diego and La Jolla, down to the beach, to his nice apartment (12′ TV), and to his aforementioned workplace, including the data center. Thank you, Bryan.

Day 7 - Friday August 1st

We departed San Diego. The airport was awfully slow. On the first flight, a piece on the wing was vibrating and scaring us. After the flight, we told the pilot and he said he would tell the mechanic.

On the second flight, there was a lot of turbulence, and Neil and I were both grateful when we had landed.

Driving back, we made a wrong turn (North on I-95) and had to back track before leaving Jacksonville.

On the way back I finished one of Jensen’s papers: Structured Importance Sampling of Environment Maps. Whew, I started this on the first day!

After the drive back I said goodbye to Neil, and Brian, Matt, Matt's brother, Mike, and I went to Schlotzsky’s for dinner.

Summary

Best of Siggraph 2003

Best Course

Monte Carlo Ray Tracing - Henrik Jensen, James Arvo, Philip Dutré, Alexander Keller, Art Owen, Matt Pharr, ? (Maybe Tim Purcell), filling in for Peter Shirley

Best Paper

All-Frequency Shadows Using Non-linear Wavelet Lighting Approximation - Ren Ng, Ravi Ramaoorthi, Pat Hanrahan

Best Animation

Gone Nutty - Blue Sky Studios

Best Product

EarthViewer 3D - Keyhole

Best Restaurant

The Field

Best Animal

Polar Bear - San Diego Zoo

Best Location

Mount Soledad

Things that could have been better

• Reception location, number. This year there was only one reception and it was at the conference. Boring!
• The best animations were left off the Electronic Theater DVD. Lame copyrights.
• No Career Center or Job Fair. No place to print my resume. Boo! Shame on me for forgetting, too.
• No panels.
• I didn't get to see Glider Port.
• I didn't get to visit UCSD's Computer Graphics Laboratory.

Acknoweledgements

Thanks to Dr. Banks for funding to go to SIGGRAPH 2003.

Kevin Beason /