Edinburgh Napier University

  Efforts to personalise Head-Related Transfer Functions (HRTFs) have traditionally focused on anatomical accuracy, modelling the head, ears and torso to predict how sound should reach an individual’s ears. While valuable, such approaches often require specialist equipment and yield variable results. They also struggle to accommodate changes in hearing or playback context and rarely provide users with opportunities to build or refine spatial perception.

This paper outlines an alternative, interaction-driven method: a perceptual approach to spatial calibration that learns from how listeners respond, rather than how they are built. Unlike anatomical or database-driven HRTF methods, this system adapts dynamically to how users perceive and respond to sound in space. It creates a calibration method that is accessible, reflexive and capable of evolving over time.