Geometrical Acoustics (GA) is a popular family of methods for simulating acoustics in virtual environments. Unfortunately, most GA techniques fail to take into account the effects of diffraction — an important physical effect that involves some of the wave-like properties of sound. Because geometric acoustics is founded on the approximation that sound travels like a ray rather than a wave diffraction can be difficult to accurately model. We propose using a signed distance function as a heuristic to help determine how to simulate diffraction. In this project, we endeavor to explore various models of diffraction to see how we can minimize aliasing while still running the simulation at interactive rates.

In this SuperUROP project, I want to gain more research experience in the field of computer graphics. I have been interested in computer graphics and systems engineering for a couple of years, and this year, I am taking Computer Graphics (6.837). I am excited to continue work with the Computer Graphics Group, especially because of the focus on performance engineering that this project will entail.