Ending Course Notes in cs348b

The following are the ending notes Professor Hanrahan left us in his class on image synthesis.

1. It is all about light. People under-appreciate the effects of light bouncing around and hitting things.
2. Radiometry is weird but key to understanding. Always keep track of power.
3. Monte Carlo. What a great hammer that is! In graphics, you cannot have an algorithm because it just won’t scale. This is why finite elements approaches to global illumination just aren’t tractable. Monte Carlo is primitive, but in the hands of a master, it can do great things. Key reasons to its success:
   • no curse of dimensionality
   • only want solution for a few points (rather than the whole thing)
   • simple and robust (i.e. hammer)
4. Tricks are things we don’t understand. Some people (like Pat) used to hate tricks, but now realized their utility. As an example, Rembrandt used many layers of paint to get skin to look just right. Rembrandt most likely did not understand why this trick worked, but when Pat saw this, it inspired him to think of skin coloring as subsurface scattering techniques.
5. The goal of computer graphics is not to simulate reality. It is to produce images that are fun and/or useful.

Working on my final project, I have certainly hit all five of these points. As Mehran would say, “it warms the cogs of my heart.”