Efficient Light and Sound Propagation in Refractive Media with Analytic Ray Curve Tracer Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 19, 2019
  • Mo, Qi
    • Affiliation: College of Arts and Sciences, Department of Computer Science
  • Refractive media is ubiquitous in the natural world, and light and sound propagation in refractive media leads to characteristic visual and acoustic phenomena. Those phenomena are critical for engineering applications to simulate with high accuracy requirements, and they can add to the perceived realism and sense of immersion for training and entertainment applications. Existing methods can be roughly divided into two categories with regard to their handling of propagation in refractive media; first category of methods makes simplifying assumption about the media or entirely excludes the consideration of refraction in order to achieve efficient propagation, while the second category of methods accommodates refraction but remains computationally expensive. In this dissertation, we present algorithms that achieve efficient and scalable propagation simulation of light and sound in refractive media, handling fully general media and scene configurations. Our approaches are based on ray tracing, which traditionally assumes homogeneous media and rectilinear rays. We replace the rectilinear rays with analytic ray curves as tracing primitives, which represent closed-form trajectory solutions based on assumptions of a locally constant media gradient. For general media profiles, the media can be spatially decomposed into explicit or implicit cells, within which the media gradient can be assumed constant, leading to an analytic ray path within that cell. Ray traversal of the media can therefore proceed in segments of ray curves. The first source of speedup comes from the fact that for smooth media, a locally constant media gradient assumption tends to stay valid for a larger area than the assumption of a locally constant media property. The second source of speedup is the constant-cost intersection computation of the analytic ray curves with planar surfaces. The third source of speedup comes from making the size of each cell and therefore each ray curve segment adaptive to the magnitude of media gradient. Interactions with boundary surfaces in the scene can be efficiently handled within this framework in two alternative approaches. For static scenes, boundary surfaces can be embedded into the explicit mesh of tetrahedral cells, and the mesh can be traversed and the embedded surfaces intersected with by the analytic ray curve in a unified manner. For dynamic scenes, implicit cells are used for media traversal, and boundary surface intersections can be handled separately by constructing hierarchical acceleration structures adapted from rectilinear ray tracer. The efficient handling of boundary surfaces is the fourth source of speedup of our propagation path computation. We demonstrate over two orders-of-magnitude performance improvement of our analytic ray tracing algorithms over prior methods for refractive light and sound propagation. We additionally present a complete sound-propagation simulation solution that matches the path computation efficiency achieved by the ray curve tracer. We develop efficient pressure computation algorithm based on analytic evaluations and combine our algorithm with the Gaussian beam for fast acoustic field computation. We validate the accuracy of the simulation results on published benchmarks, and we show the application of our algorithms on complex and general three-dimensional outdoor scenes. Our algorithms enable simulation scenarios that are simply not feasible with existing methods, and they have the potential of being extended and complementing other propagation methods for capability beyond handling refractive media.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Bishop, Gary
  • Lastra, Anselmo
  • Lin, Ming
  • Niethammer, Marc
  • Manocha, Dinesh
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2015
Place of publication
  • Chapel Hill, NC
  • There are no restrictions to this item.

This work has no parents.