Interactive Sound Propagation for Massive Multi-user and Dynamic Virtual Environments Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 22, 2019
Creator
  • Taylor, Micah
    • Affiliation: College of Arts and Sciences, Department of Computer Science
Abstract
  • Hearing is an important sense and it is known that rendering sound effects can enhance the level of immersion in virtual environments. Modeling sound waves is a complex problem, requiring vast computing resources to solve accurately. Prior methods are restricted to static scenes or limited acoustic effects. In this thesis, we present methods to improve the quality and performance of interactive geometric sound propagation in dynamic scenes and precomputation algorithms for acoustic propagation in enormous multi-user virtual environments. We present a method for finding edge diffraction propagation paths on arbitrary 3D scenes for dynamic sources and receivers. Using this algorithm, we present a unified framework for interactive simulation of specular reflections, diffuse reflections, diffraction scattering, and reverberation effects. We also define a guidance algorithm for ray tracing that responds to dynamic environments and reorders queries to minimize simulation time. Our approach works well on modern GPUs and can achieve more than an order of magnitude performance improvement over prior methods. Modern multi-user virtual environments support many types of client devices, and current phones and mobile devices may lack the resources to run acoustic simulations. To provide such devices the benefits of sound simulation, we have developed a precomputation algorithm that efficiently computes and stores acoustic data on a server in the cloud. Using novel algorithms, the server can render enhanced spatial audio in scenes spanning several square kilometers for hundreds of clients in realtime. Our method provides the benefits of immersive audio to collaborative telephony, video games, and multi-user virtual environments.
Date of publication
Keyword
Resource type
Rights statement
  • In Copyright
Advisor
  • Manocha, Dinesh
Degree
  • Doctor of Philosophy
Graduation year
  • 2014
Language
Publisher
Parents:

This work has no parents.

Items