HRTF Individualization using Deep Learning

Abstract

The research presented in this paper focuses on Head-Related Transfer Function (HRTF) individualization using deep learning techniques. HRTF individualization is paramount for accurate binaural rendering, which is used in XR technologies, tools for the visually impaired, and many other applications. The rising availability of public HRTF data currently allows experimentation with different input data formats and various computational models. Accordingly, three research directions are investigated here: (1) extraction of predictors from user data; (2) unsupervised learning of HRTFs based on autoencoder networks; and (3) synthesis of HRTFs from anthropometric data using deep multilayer perceptrons and principal component analysis. While none of the aforementioned investigations has shown outstanding results to date, the knowledge acquired throughout the development and troubleshooting phases highlights areas of improvement which are expected to pave the way to more accurate models for HRTF individualization.