Abstract:Several individualization methods have recently been proposed to estimate a subject's Head-Related Transfer Function (HRTF) using convenient input modalities such as anthropometric measurements or pinnae photographs. There exists a need for adaptively correcting the estimation error committed by such methods using a few data point samples from the subject's HRTF, acquired using acoustic measurements or perceptual feedback. To facilitate this, we introduce a Convolutional Conditional Neural Process meta-learner specialized in HRTF error interpolation. In particular, the model includes a Spherical Convolutional Neural Network component to accommodate the spherical geometry of HRTF data. It also exploits potential symmetries between the HRTF's left and right channels about the median plane. In this work, we evaluate the proposed model's performance purely on time-aligned spectrum interpolation grounds under a simplified setup where a generic population-mean HRTF forms the initial estimates prior to corrections instead of individualized ones. The trained model achieves up to 3 dB relative error reduction compared to state-of-the-art interpolation methods despite being trained using only 85 subjects. This improvement translates up to nearly a halving of the data point count required to achieve comparable accuracy, in particular from 50 to 28 points to reach an average of 20 dB relative error per interpolated feature. Moreover, we show that the trained model provides well-calibrated uncertainty estimates. Accordingly, such estimates could inform the sequential decision problem of acquiring as few correcting HRTF data points as needed to meet a desired level of HRTF individualization accuracy.

Analysis on minimum-phase characteristics of measured head-related transfer functions affected by sound source responses

Measurement and Analysis of Near-Field Head-Related Transfer Function

An asynchronous HRTF measurement method based on phase alignment

Head-Related Transfer Function Modeling Based on Finite-Impulse Response

Effect of Sound Source Scattering on Measurement of Near-Field Head-Related Transfer Functions

Experimental Verification of the Minimum Bhattacharyya Distance-Based Source Bearing Estimator.

Numerical Modeling of Head-Related Transfer Functions Using the Virtual Source Representation

Identification of Anthropometric Measurements for Individualization of Head-Related Transfer Functions

Magnitude-Corrected and Time-Aligned Interpolation of Head-Related Transfer Functions

Distance-Dependent Head-Related Transfer Functions Measured with High Spatial Resolution Using a Spark Gap

A Study of Morphological Influence on Head-Related Transfer Functions

Minimum-Phase HRTF Modeling of Pinna Spectral Notches using Group Delay Decomposition

Time-of-arrival Estimation and Phase Unwrapping of Head-related Transfer Functions With Integer Linear Programming

A Study on Calibration Methods of Noise Annoyance Data from Listening Tests

Dependency of the Finite-Impulse-Response-Based Head-Related Impulse Response Model on Filter Order

Individualized Head-Related Transfer Functions Based on Population Grouping

Interpolation Method of Head-Related Transfer Functions Based on Common-Pole/zero Modeling

HRTF Interpolation using a Spherical Neural Process Meta-Learner

Comparison of headphones and equalization for virtual auditory source localization

A Sparse Spherical Harmonic-Based Model in Subbands for Head-Related Transfer Functions.