Related Experiment Videos
Infinite-impulse-response models of the head-related transfer function
Abhijit Kulkarni1, H Steven Colburn
1Hearing Research Center and Department of Biomedical Engineering, Boston University, 44 Cummington Street, Boston, Massachusetts 02215, USA.
The Journal of the Acoustical Society of America
|April 23, 2004
Summary
This study approximates head-related transfer functions (HRTFs) using efficient infinite-impulse-response (IIR) filter models. These models accurately represent spatial audio, enabling indistinguishable virtual sound perception.
Area of Science:
- Acoustics
- Signal Processing
- Psychoacoustics
Background:
- Head-related transfer functions (HRTFs) are crucial for realistic 3D audio.
- Accurate HRTF modeling is computationally intensive.
- Efficient low-order models are needed for real-time applications.
Purpose of the Study:
- To develop and validate computationally efficient, low-order infinite-impulse-response (IIR) filter models for HRTFs.
- To assess the perceptual indistinguishability of sounds synthesized using model HRTFs compared to measured HRTFs.
Main Methods:
- Approximated measured HRTFs using rational transfer functions (poles and zeros).
- Employed linear prediction for all-pole models and weighted least-squares for pole-zero models.
- Conducted psychophysical experiments using a four-interval, two-alternative, forced-choice paradigm with noise token stimuli.
Main Results:
- Low-order all-pole and pole-zero IIR models were derived efficiently.
- Sounds synthesized with model HRTFs were perceptually indistinguishable from those using measured HRTFs for most tested locations.
- The models offer significant computational advantages.
Conclusions:
- Efficient low-order IIR models can accurately represent HRTFs.
- These models maintain perceptual realism in virtual auditory environments.
- The developed techniques have implications for both basic research and applied spatial audio systems.