David Romblom and Helene Bahu (Dysonics) Structural Head Models for Interactive 3D Audio
Date:
Fri, 11/09/2018 - 10:30am - 12:00pm
Location:
Seminar Room at the Knoll
Event Type:
Hearing Seminar At this week’s CCRMA Hearing Seminar, Helene Bahu and David Romblom from Dysonics in SF will be talking about their methods to estimate HRTFs from simple physical measurements. Can we generate good enough 3d audio using just the size of the head? And what is good enough?
Dick Duda, one of the original denizens of the hearing seminar started down this path many years ago. It’s nice to see these ideas coming back to CCRMA.
Who: Helene Bahu and David Romblom (Dynsonics)
What: Structural Head Models for Interactive 3D Audio
When: Friday November 9th at 10:30AM
Where: CCRMA Seminar Room (Top Floor of the Knoll at Stanford)
Why: Because we all want our sound in 3D
Come to CCRMA and we’ll discuss with real 3D audio how to make 3D audio.
- Malcolm
Structural Head Models for Interactive 3D Audio.
Helene Bahu and David Romblom
Abstract:
Structural models of Head Related Transfer Functions (HRTF) attempt to decompose complex acoustic phenomena into constituent signal processing models. In the absence of measurements, such models offer a simple way to predict HRTF from a small number of parameters that can ideally be related to morphological quantities. These models find wide application due to their low computational complexity for interactive applications, good sound quality, and potential for individualization.
This talk presents two recent papers about head modeling. The first pertains to the fitting of the spherical and ellipsoidal head models to the measured ITD of thirty seven subjects. Lower global errors are obtained when the interaural axis is no longer constrained to intersect the center of the sphere or ellipsoid. A well known paper by Brown and Duda modeled the head with two elements: a pure delay estimated from a ray-tracing formula by Woodworth and a 1-pole 1-zero shadowing filter. The ray-tracing formula is valid for frequencies above ≈ 2kHz while ITD is perceptually significant below ≈ 1.5kHz. Using a numerical approximation of Rayleigh’s solution as a reference, the second paper presents evaluation and revision of the 1-pole 1-zero shadowing filter.
Biographies:
Helene Bahu obtained her PhD from IRCAM (Paris) in 2016, where her work focused on auditory localization in the context of non-individual binaural synthesis. During this time she measured the BiLi HRTF database and conducted several auditory localization experiments. She is currently working as a spatial audio researcher at Dysonics in San Francisco. Her research focuses on the modeling of individualized auditory cues on compute-limited devices.
David Romblom received Music and Electrical Engineering degrees from the University of Michigan, a Master’s in Media Arts and Technology from the University of California at Santa Barbara, and his Ph.D. in Music Technology from McGill University. He has worked for E-mu Systems, Universal Audio, Sennheiser Research, and is current directing the research and development efforts at Dysonics in San Francisco. His research interests center on perceptually-transparent approximations of acoustic reality.
FREE
Open to the Public