OwnSurround HRTF Service for Professionals
3 likes•1,835 views
The document outlines a high-resolution Head-Related Transfer Function (HRTF) service aimed at professionals, utilizing cloud computing technology for accurate simulations. It emphasizes the importance of geometry and surface impedances in achieving high fidelity in audio rendering and discusses the challenges associated with HRTF measurements, including variability and inaccuracies in near-field recordings. Customizable input data and output formats, including 3D scanning for personalized simulations, are also featured.
OwnSurround HRTF Service for Professionals
- 1. HRTF Service for Professionals 2016
- 2. In a nutshell • Simulation based service for high-resolution HRTF acquisition • Full control of Head-and-Torso (HAT) model and environment: • Ideal point sources/receivers at exact locations • Perfectly anechoic environment • Easiness of geometric changes (CAD model) • Freely adjustable surface impedances • Superior spatial resolution • Easily utilization of results in the SOFA-format
- 3. Scan, simulate and listen   OwnSurround uses full-wave cloud computing technology for the fast simulation of the HRTF
- 4. Challenges in HRTF measurements • Inter-measurement variability “Results revealed spectral magnitude variations up to 12.5 dB for frequency bands below 6 kHz and up to 23 dB above that.” “Nevertheless, the observed ITD variations of up to 235 sec are alarming as they often exceeded reported JND values. Such findings highlight the potential impact of physical spaces, measurement routines, and equipment types on the collected HRTF data.”* • Inaccurate in near-field “Note, that for the distance of 0.5m the used Genelec loudspeaker presents not really a point source. In addition, the HRTFs for 0.5m include reflections of the sound going from the KEMAR head back to the loudspeaker and back to the dummy head.”** *Andreopoulou et al: Inter-Laboratory Round Robin HRTF Measurement Comparison, IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING, VOL. 9, NO. 5, 2015 **http://twoears.aipa.tu-berlin.de/doc/1.2/database/impulse- responses/#anechoic-measurements-hrtfs
- 5. Simulated HRTFs • “… virtualisation by using the simulated HRTFs is preferred over virtualisation by the measured HRTFs …”* • Fast boundary element solution of the acoustic wave equation** • The HRTF is solely defined by the geometry of model and surface impedances full control of accuracy *Turku et al: Perceptual evaluation of numerically simulated Head-Related Transfer Function, AES 124th Convention, 2008. **Huttunen et al: Rapid generation of personalized HRTFs, AES 55th International Conference: Spatial Audio, 2014.
- 6. High accuracy • Increased spatial and spectral resolution • Simulations can be easily computed for thousands of sound directions (e.g. one degree spacing in both azimuth and elevation) • High spectral resolution (e.g. 32 - 24 000 Hz in 32Hz steps) • Near-field HRTFs (eg. close to ear sources) • Ideal point sources guarantee the high accuracy in near- field
- 7. 3d surface mesh • The simulation needs a triangulated surface mesh for the head and torso • Obtained using 3d scanning: • Structured light scanning • Photogrammetry • Laser scanning • Well-known requirements for resolution*: • <2mm on pinna • <5mm elsewhere*Ziegelwanger et al: Numerical calculation of listener-specific head-related transfer functions and sound localization: Microphone model and mesh discretization, The Journal of the Acoustical Society of America, vol 138, no 1, pp. 208-222, 2015
- 8. Pinna • A key role in personalized HRTFs • Visibility of 3d scanning to the entrance of the ear-canal Simulation corresponds to blocked ear measurements
- 9. 3d scanning • Use a local 3d scanning service • OwnSurround provides a scanning guide • Scanning service by OwnSurround • At your own premises
- 10. Input data • The mesh (e.g. in STL-format) • Customizable parameters (default values in parenthesis): • Frequencies (64 - 24000Hz in 64 Hz steps) • Sound directions and distances (838 points on 2m sphere) • Location of the “measurement”point at the entrance of the ear-canal (at the center of the ear-canal entrance) • Surface impedances (for skin, hair and torso)
- 11. Output data • Head-Related Impulse Responses (HRIRs) used in binaural renderers (e.g. in SOFA-format) • Custom output for frequency domain data: • e.g. Tabulated complex valued pressure for each direction, distance and frequency
- 12. Advanced topics • Ear-canals can be included to the simulation model if the geometry is available (e.g. using ear-canal scanning, see www. unitedsciences.com/efit-scanner/ ) • The model can include hearing-aids, headphones, mobile devices, etc. • Surface impedances: • Complex valued as a function of location and frequency • Other custom simulations
- 13. Contact information CEO Antti Vanne tel +358 50 3566 744 [email protected] Microkatu 1, FI-70211 Kuopio Finland