Head-related Transfer Function Synthesis Research Articles

Regularization Approaches for Synthesizing HRTF Directivity Patterns

As an alternative to traditional artificial heads, it is possible to synthesize individual head-related transfer functions (HRTFs) using a so-called virtual artificial head (VAH), consisting of a microphone array with an appropriate topology and filter coefficients optimized using a narrowband least squares cost function. The resulting spatial directivity pattern of such a VAH is known to be sensitive to small deviations of the assumed microphone characteristics, e.g., gain, phase and/or the positions of the microphones. In many beamformer design procedures, this sensitivity is reduced by imposing a white noise gain (WNG) constraint on the filter coefficients for a single desired look direction. In this paper, this constraint is shown to be inappropriate for regularizing the HRTF synthesis with multiple desired directions and three alternative different regularization approaches are proposed and evaluated. In the first approach, the measured deviations of the microphone characteristics are taken into account in the filter design. In the second approach, the filter coefficients are regularized using the mean WNG for all directions. The third approach additionally takes into account several frequency bins into both the optimization and the regularization. The different proposed regularization approaches are compared using analytic and measured transfer functions, including random deviations. Experimental results show that the approach using multiple frequency bands mimicking the spectral resolution of the human auditory system yields the best robustness among the considered regularization approaches.

System for automatic personalization of head-related transfer functions based on computer vision, photo-anthropometry, and inference from a database

A software system and method for automatic personalization of head related transfer functions (HRTF) is presented. The system pursues the objective of personalizing HRTF using the anthropometry of the subject, as proposed originally by [Zotkin et al. WASPAA 2003, 157–160], which in this work, is measured automatically from a set of photographs of the subject. The system operates in three stages. First, a computer vision algorithm, known as active shape models, is used over the portraits to recognize and adjust geometric form profiles of the ears, head, and torso of the subject. Then, anthropometry is performed by measurement of pixel distances between specific points in the model, and then converted into metric units. Finally, HRTF are estimated from the anthropometry using a choice of three methods, whose performance is compared: (1) HRTF selection of the best anthropometric match in the CIPIC database, (2) HRTF synthesis by multiple linear regression and principal component analysis, and (3) HRTF synthesis using an artificial neural network. Results are analyzed, concluding that automatic personalization of HRTF is attainable automatically from the subject portraits, using computer vision and inference from a database. Further analysis, however, reveals the need for more complete HRTF and anthropometric databases.

On the Relation Between Pinna Reflection Patterns and Head-Related Transfer Function Features

This paper studies the relationship between head-related transfer functions (HRTFs) and pinna reflection patterns in the frontal hemispace. A pre-processed database of HRTFs allows extraction of up to three spectral notches from each response taken in the median sagittal plane. Ray-tracing analysis performed on the obtained notches' central frequencies is compared with a set of possible reflection surfaces directly recognizeable from the corresponding pinna picture. Results of such analysis are discussed in terms of the reflection coefficient sign, which is found to be most likely negative. Based on this finding, a model for real-time HRTF synthesis that allows to control separately the evolution of different acoustic phenomena such as head diffraction, ear resonances, and reflections is proposed through the design of distinct filter blocks. Parameters to be fed to the model are derived either from analysis or from specific anthropometric features of the subject. Finally, objective evaluations of reconstructed HRTFs in the chosen spatial range are performed through spectral distortion measurements.

Head-Related Transfer Function Synthesis Method Based on SOCP

Head-Related Transfer Function Synthesis Method Based on SOCP

Insights into head-related transfer function: Spatial dimensionality and continuous representation

This paper studies head-related transfer function (HRTF) sampling and synthesis in a three-dimensional auditory scene based on a general modal decomposition of the HRTF in all frequency-range-angle domains. The main finding is that the HRTF decomposition with the derived spatial basis function modes can be well approximated by a finite number, which is defined as the spatial dimensionality of the HRTF. The dimensionality determines the minimum number of parameters to represent the HRTF corresponding to all directions and also the required spatial resolution in HRTF measurement. The general model is further developed to a continuous HRTF representation, in which the normalized spatial modes can achieve HRTF near-field and far-field representations in one formulation. The remaining HRTF spectral components are compactly represented using a Fourier spherical Bessel series, where the aim is to generate the HRTF with much higher spectral resolution in fewer parameters from typical measurements, which usually have limited spectral resolution constrained by sampling conditions. A low-computation algorithm is developed to obtain the model coefficients from the existing measurements. The HRTF synthesis using the proposed model is validated by three sets of data: (i) synthetic HRTFs from the spherical head model, (ii) the MIT KEMAR (Knowles Electronics Mannequin for Acoustics Research) data, and (iii) 45-subject CIPIC HRTF measurements.

Investigation of phase distortion in the synthesis of head-related transfer functions

Synthesizing over headphones the free-field to ear-canal transfer functions, or head-related transfer functions (HRTFs), of the human auditory system poses several difficulties from the standpoint of digital signal processing. Since the order of the HRTF is unknown, it is necessary to assume a particular value when estimating the HRTF from responses obtained in an anechoic chamber. Additionally, while standard filter designs minimize the squared error between the desired and designed response, applications such as HRTF synthesis are likely to require error metrics that are sensitive to disparities between the log magnitude response of each system. A new design algorithm is proposed that is based on a logarithmic distortion metric and it is shown how this algorithm can be used to obtain minimum phase, low-order pole-zero approximations to HRTFs. Psychophysical results characterizing audibility sensitivity to such HRTF approximation errors are presented in order to establish an audible limit for the distortion in such designs. These results are compared with similar experimental results for the widely used least-squares metric. [Research supported by a grant from the NIDCD of the National Institute of Health.]