Head-Related Transfer Function Estimation
A research project exploring how spatial acoustics and binaural cues can be estimated and interpreted through data-driven methods.
Project Overview
This work focuses on the estimation of head-related transfer functions, a core component of spatial hearing that links physical sound fields to the cues used for localization. The project investigates how these effects can be captured and modelled in a way that is both interpretable and useful for practical audio systems.
The motivation comes from the need to understand how acoustic filtering differs across listeners, environments, and device geometries, and how that information can be used to improve immersive audio experiences.
Methodology
The work combines measurement, signal processing, and estimation techniques to characterise the relationship between source position, acoustic propagation, and the resulting binaural response. The analysis pipeline is designed to make visible the structure of the cues while remaining robust to noise and limited measurement data.
This approach enables a direct link between measured acoustic responses and the computational models used to reproduce them in simulation or audio rendering systems.
Results
The project demonstrates that a careful estimation framework can recover meaningful spatial cues while preserving the underlying structure of the transfer function. The resulting models provide a practical basis for further research into spatial audio, auditory perception, and immersive rendering.
The work also highlights the value of combining physical insight with signal-processing methods when the measurement setting is constrained or noisy.