When: Wednesday 22nd November 2023 @ 5:15 PM
Where: The Atrium (G.10), Alison House, 12 Nicholson Sq, University of Edinburgh
Title: Head-Related Transfer Function Interpolation: An Overview of Algorithmic Approaches
Speakers: Fabian Brinkmann (Technical University of Berlin)
Head-related transfer functions (HRTFs) describe the free field sound propagation from a source to the listener’s ears. Due to reflections and diffraction on the listener’s ears, head and torso, these filters contain information that is evaluated by the human auditory system to infer the position of the source. HRTFs are essential for almost all virtual and augmented reality systems that convey spatial auditory information. Since the auditory system is highly sensitive to changes in source position, such systems require HRTFs with high spatial resolution. In this talk I will give an overview of algorithmic, i.e. non-learning based, algorithms for HRTF interpolation that provide a way to obtain HRTFs in arbitrary spatial resolution. Finally, I will compare the performance of algorithmic and learning-based approaches. To make the talk accessible to audiences from other acoustic disciplines, I will start with a brief introduction to HRTFs and spatial hearing.
Fabian Brinkmann received his M.A. in Communication Sciences and Technical Acoustics in 2011 and his Dr. rer. nat. in 2019 from the Technische Universität Berlin, Germany. He focuses on signal processing and evaluation approaches for spatial audio, including simulation and psycho acoustics of spatial hearing, room acoustic simulation, and technical and perceptual evaluation approaches for spatial audio algorithms. Since 2019, he is a postdoctoral researcher at the Audio Communication Group at TU Berlin, where he co-leads the Binaural Processing and Perception team.