Final Thesis

Investigating the spatial variation of characteristic spectral features in HRTFs

Key Info

Basic Information

Research Area:


Virtual reality (VR) plays an increasingly important role in the development of audiological health applications, as well as for entertainment purposes. The DFG project “Individual binaural synthesis of virtual acoustic scenes” (in cooperation with Prof. Dr.-Ing Sascha Spors, Institute of Communications Engineering, University of Rostock) aims at evaluating the quality of virtual acoustic reproduction, which makes use of so-called “head-related transfer functions” (HRTFs). HRTFs describe the direction-specific filtering of sound, as it arrives at the ear canal entrance from an arbitrary direction. The interaction of soundwaves with the listener’s geometry leads to characteristic peaks and notches in the frequency spectrum. These spectral maxima and minima vary with the direction of sound incidence, as well as between different listeners, due to individual head and pinna shapes. In this thesis, dominant peaks and notches of HRTFs are identified, examining characteristic changes with spatial direction. Methods like pole-zero analysis [1] and/or spherical harmonics decomposition (SHD) can be applied for this purpose. A distance metric is developed for comparison of different HRTF sets, which is later to be linked to perceptual attributes in spatial audio reproduction. Requirements  Basic knowledge in acoustics and digital signal processing  Interest in the topic Recommended  Basic MATLAB skills