|
|
 |
|
|
|
| Measurement of Personalized HRTFs |
|
|
Approach and Objectives
Previous work has resulted in the development of a very high-speed digital signal processor, the Convolvotron, for presentation of spatial audio cues in real time over headphones. This device uses filtering by time-domain finite impulse response (FIR) filters based on head-related transfer functions (HRTFs). While both successful and useful for presenting spatial information in a variety of aerospace applications, the Convolvotron is still limited in the complexity of environmental (room) modeling that can be achieved and which requires enormous computational resources. This is important because research indicates that synthesis of purely anechoic sounds can result in a variety of perceptual errors and these errors tend to be exacerbated when virtual sources are generated from non- personalized HRTFs, a common circumstance for most spatial auditory displays. Other research suggests that such errors may be mitigated by providing more complex acoustic cues from reflective environments. Thus, there is a need to resolve the issue of using non-personalized HRTFs and to develop a much more computationally-efficient method of synthesizing complex acoustic environments in real time than is currently realizable. The primary objective of our advanced technology development effort is to develop display systems which maximize perceptual accuracy by presenting a carefully designed set of veridical acoustic cues to the listener that have been validated in psychoacoustical studies. The approach taken to achieve this goal is two-fold: (1) improve the ability to make reliable measurements of HRTFs in any environment for any listener and (2) develop techniques for rendering complex environmental cues in real-time (see section 3.3).
Accomplishments
Development of a relatively low-cost portable system for measuring HRTFs, called Snapshot (Figure 3.2), was completed at Crystal River Engineering and installed at the NASA Ames laboratory. The system employs a single speaker mounted on an adjustable arm. The subject, seated on a swivel chair, is outfitted with blocked-ear canal microphones and HRTFs are measured using Golay code sequences. By constraining the geometry to ensure that no early reflections arrive within the first few milliseconds after the direct path arrival, the need for an anechoic chamber is eliminated -- simple windowing techniques are used to extract the desired HRTF from the measured impulse response. Initial acoustical studies have verified that, for the majority of subjects, measurements using a blocked-ear canal microphone are comparable to those made with probe (inside the ear canal) microphones (Wightman, Kistler, Foster, & Abel, 1995).
Future plans
Initial measurements of individual subjects' HRTFs have been made at NASA Ames, but we are still developing experimental techniques for ensuring the accurate and replicable placement of the subject and the sound source during a set of measurements, e.g., by integrating a head-tracker into the measurement chain. We are also developing software extensions to allow measurement of materials properties and room reverberation. Once these features are in place we will have the ability to personalize virtual acoustic displays for both our basic and applied studies and to more readily extend these studies to more complex acoustic environments.
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.
Wenzel, E. M. (1992). "Localization in virtual acoustic displays," Presence, 1, 80-107.
Wightman, F. L., & Kistler, D. J. (1989a). Headphone simulation of free-field listening. I: Stimulus synthesis. Journal of the Acoustical Society of America, 85, 858-867.
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.
|
|
|
|
|
