Perceptually motivated customization of generic head-related transfer functions

Axel Ahrens

To perceive a virtual acoustic source over headphones in the same way as a real source, a variety of features need to be matched. Some of these  features are individually different for each person as they occur due to anatomical differences. The resulting change is also referred to as head-related transfer functions (HRTFs) The feature space on a HRTF is highly dimensional and is cumbersome to acoustically measure. Therefore, machine learning (ML) approaches have been used to determine these features for every individual. However, large sets of training data are needed to predict the large feature space. The aim of this project is to reduce this feature space using perceptual measures, thus, eliminating features that are not perceived or are not important. This reduction may allow for a reduced size of training data and to increase the individualization of HRTF.

This project is in collaboration with Facebook Reality Labs

From the 1st of March 2021, Axel Ahrens has started a new postdoc position at the University of Southern Denmark in collaboration with DTU, Boston University and WS Audiology. The project was awarded to Tobias Neher and funded by the Independent Research Fund Denmark. Hearing-impaired listeners are known to struggle to understand speech particularly in rooms with strong reverberation. The aim of this project is to understand the perceptual mechanisms of speech communication in listeners with hearing loss, particularly in the presence of reverberation. Furthermore, the influence of hearing aids and algorithms will be investigated.

This project is in collaboration with the University of Southern Denmark, DTU, Boston University, WS Audiology and funded by the Independent Research Fund Denmark

DTU Orbit

Contact

Axel Ahrens
Academic employee
DTU Health Tech