---

* Research

Ph.D. Theses

Algorithms for Aural Representation and Presentation of Quantitative Data to Complement and Enhance Data Visualization

By Christopher Richard Volpe
Advisor: Ephraim Glinert
April 2, 2002

Large quantities of complex data are part of almost every industry and science. For some time now, experts in these domains have relied on computers to present their data to them in a form which is easier to understand. One way this is done is through a process called visualization, which refers to generating graphical images that capture essential characteristics of the data and highlight interesting relationships. Another approach, which has received far less attention, is to use sound as a means of presenting complex information. This approach, called auralization, is the auditory analog of visualization. This thesis involves novel techniques for auralization of scientific data and focuses on the general concept of an auditory illusion as a means of addressing some of the issues associated with more conventional techniques. In this thesis we propose, and implement, a general framework for incorporating auralization techniques with scientific visualization. We define a platform-independent software architecture that allows for a pipeline of sound processing components, including sources, filters, and device-mappers. We implement the necessary infrastructure that permits new auralization techniques to be developed and tested, including a hardware-specific sound device mapper that facilitates interactive sound presentation synchronous with an animated visualization. We examine an auditory illusion which produces a sound that seems to ascend or descend endlessly in pitch, and show how to use visualization data to control this sound. We demonstrate the applicability of this illusion for presenting Computational Fluid Dynamics data. We also show that an analysis of this illusion leads to general principles that can be adapted to construct another auditory illusion that may be used for auralization, one that produces pulses that seem to increase or decrease endlessly in rate. And we demonstrate through formal user testing and statistical analysis that an aural data presentation using an auditory illusion can improve performance in locating key data characteristics, a task that demonstrates a certain level of understanding of the data. Our data show that this holds true even when the user expresses a subjective preference and greater confidence in a visual presentation. Finally, we discuss some open questions and opportunities for future work, and provide information on how to obtain our software.

* Return to main PhD Theses page


---

---