|
|
 |
|
|
|
| Localization in Reflective Environments |
|
|
Approach and Objectives
One problem in synthesizing veridical acoustic images over headphones is the fact that such stimuli sometimes fail to externalize. Environmental cues such as the ratio of direct to reflected energy and reverberation time appear to enhance the externalization of images. Localization of virtual sounds may also improve if the listener is allowed to become familiar with sources as they interact in a particular artificial acoustic world. Such simulations may be achieved using techniques based on ray-tracing or an algorithm known as the image model (Figure 3.3). For example, simulation of an asymmetric reverberant room may aid the listener in distinguishing front from rear locations by strengthening spectral or timbral differences between the two positions and externalization may also be enhanced by the presence of reflective cues from the environment. The specific parameters used in such a model must be investigated carefully if localization accuracy is to remain intact. It may be possible to discover an optimal trade-off between environmental parameters which improve externalization and reduce confusions while minimizing the impact of the resulting expansion of the spatial image which can interfere with the ability to judge the direction of the source.
Accomplishments
Begault has investigated the effects of synthetic reverberation (based on a ray-tracing model of an asymmetric room) on inexperienced listeners' localization and externalization of static, virtual sound sources using the nonperosnalized HRTFs. He found that, compared to anechoic stimuli, adding reverberation to speech stimuli nearly eliminated non- externalized judgments but tended to decrease localization accuracy. Reversals, on the other hand, were relatively unaffected. Average rates were 33% for both anechoic and reverberant conditions, although some individual differences were observed in the relative bias toward front-to-back versus back-to-front reversals in the two stimulus conditions.
A pilot study by Begault also recently focused on determining echo thresholds using non- realtime stimuli in which the intensity of a single early reflection at 180 degrees, 45 degrees or 90 degrees azimuth was manipulated in the context of a direct path at 0 degrees azimuth plus 2 early reflections at +/- 90 degrees azimuth (all at 0 degrees elevation). The simulated acoustic environment was based on the image model. Preliminary data suggest that, because of the degree of individual differences among subjects, the threshold for inclusion of an early reflection in a virtual acoustic environment should be quite conservative, e.g., 20 dB below the direct path intensity, especially for reflections closer to the plane of the direct path.
Future plans
Future work will extend the above investigations to localization of virtual sources in interactive, reflective environments and test the hypothesis that appropriately-chosen environmental cues can improve localization accuracy. For example, preliminary observations suggest that (1) the upward bias in elevation seen for anechoic sounds is reduced when a floor reflection is added, and (2) lateral reflections may be particularly important for a realistic, externalized auditory perception. These studies will examine reflection cues in dynamic contexts using advanced technology developed in collaboration with Crystal River Engineering.
Recent technology development efforts have been concerned with developing synthesis techniques based on reduced-complexity representations of the HRTF, such as minimum- phase estimates, principal components analysis, or ARMA architectures. Because of their reduced complexity, such representations facilitate the real-time implementation of more realistic room models. Currently, we plan to adapt a Crystal River Acoustetron II system to enable real-time, head-tracked simulation of one or more direct paths plus up to about 12 early reflections. We are also developing software to allow experimental control of spatial sound parameters for studies investigating the number, placement, and fidelity of reflections.
Key references
Abel, J. S. and Foster, S. H. (1994) Snapshot HRTF Measurement System User's Guide. Crystal River Engineering, 490 California Ave., Suite 200, Palo Alto, CA 94306, USA.
Abel, J. S. and Foster, S. H. (1995) Measuring HRTFs in a Reflective Environment. In G. Kramer & S. Smith (Eds.), Proceedings of the 1994 International Conference on Auditory Displays, (p. 265). Santa Fe, NM.
Begault, D. R. (1992). Perceptual effects of synthetic reverberation on three-dimensional audio systems. Journal of the Audio Engineering Society, 40, 895- 904.
Begault, D. R. (in press) Audible and inaudible early reflections: thresholds for auralization system design. Audio Engineering Society 100th Convention , New York: Audio Engineering Society, preprint.
Wenzel, E. M. (in press) Research in virtual acoustic displays at NASA. [Invited Paper] Proceedings of SimTecT 96, The Simulation Technology and Training Conference, March 25- 27, 1996, Melbourne, Victoria, AUSTRALIA.
Wightman, F. L., Kistler, D. J., Foster, S. H., Abel, J. (1995). A comparison of head- related transfer functions measured deep in the ear canal and at the ear canal entrance. Abstracts of the 17th Midwinter Meeting, Association for Research in Otolaryngology, 71.
|
|
|
|
|
