Head-related transfer function interpolation through multivariate polynomial fitting of principal component weights

Head-Related Transfer Function (HRTF) representation and interpolation is an important problem in spatial audio. We present a kernel regression method based on Gaussian process (GP) modeling of the joint spatial-frequency relationship between HRTF measurements and obtain a smooth non-linear representation based on data measured over both arbitrary and structured spherical measurement grids. This representation is further extended to the problem of extracting spectral extrema (notches and peaks). We perform HRTF interpolation and spectral extrema extraction using freely available CIPIC HRTF data. Experimental results are shown.

Head related transfer function HRTF is the main technique of binaural synthesis, which is used to reconstruct spatial sound image, and the HRTF data only can be obtained by measurement. A high resolution HRTF database contains too many HRTFs, the workload of measurement is too huge to be finished. As a solution, in order to calculate new HRTF by measured HRTFs, many researchers concentrate on the interpolation of HRTF. But, before interpolating, HRTFs should be aligned because there is time delay between different HRTFs. Some researchers try to implement aligning operation based on phase, but the method is not appropriate since the periodicity of phase. Another idea to align HRTFs is by detecting method, however, the time difference is too tiny to detect exactly. None of the methods can provide a good and stable performance. In this paper, we propose a new method to align HRTFs based on correlation. And the experiments show that the proposed aligning method improves the accuracy index SDR 18.5i¾?dB for the most, furthermore, the proposed method could improve the accuracy for all positions.

Head-related transfer function (HRTF) plays an important role in the construction of 3D auditory display. This paper presents an individual HRTF modeling method using deep neural networks based on spatial principal component analysis. The HRTFs are represented by a small set of spatial principal components combined with frequency and individual-dependent weights. By estimating the spatial principal components using deep neural networks and mapping the corresponding weights to a quantity of anthropometric parameters, we predict individual HRTFs in arbitrary spatial directions. The objective and subjective experiments evaluate the HRTFs generated by the proposed method, the principal component analysis (PCA) method, and the generic method. The results show that the HRTFs generated by the proposed method and PCA method perform better than the generic method. For most frequencies the spectral distortion of the proposed method is significantly smaller than the PCA method in the high frequencies but significantly larger in the low frequencies. The evaluation of the localization model shows the PCA method is better than the proposed method. The subjective localization experiments show that the PCA and the proposed methods have similar performances in most conditions. Both the objective and subjective experiments show that the proposed method can predict HRTFs in arbitrary spatial directions.

이 연구는 중앙면 상에서 주성분 분석법을 이용하여 시간 및 주파수 영역에서 머리전달함수의 모형화 기법들을 다룬다. 시간영역의 머리전달함수, 복소수 값의 머리전달함수, 확장된 머리전달함수, 로그 크기의 머리전달함수에 기반하여 각각 주성분 분석법을 수행하여 얻은 네 가지 머리전달함수 모형들에 대해서 최소자승오차 관점에서 모형화 성능을 비교하고, 모형들간의 이론적인 관계를 살펴보는 것이 이 연구의 목적이다. 모형화에 사용되는 기저함수의 수가 동일하다면, 시간영역에서의 머리전달함수 혹은 확장된 머리전달함수에 기반한 모형이 복소수 값의 머리전달함수에 기반한 모형보다 최소자승오차 관점에서 더 효율적인 모형화 성능을 지닌다. 시간영역에서의 머리전달함수에 기반한 모형과 확장된 머리전달함수에 기반한 모형은 이론적으로 동일한 모형이며 서로 푸리에 변환 관계가 있다. 로그 크기의 머리전달함수에 기반한 모형은 다른 모형들과 모형화 성능 및 이론적인 관계를 비교할 수가 없는데, 이는 로그 크기의 머리전달함수에 기반한 모형은 머리전달함수의 크기 정보만을 로그 크기로 다루는 반면에 다른 모형들은 선형 크기로 머리전달함수의 크기와 위상정보를 모두 다루기 때문이다. This study deals with modeling of head-related transfer functions(HRTFs) using principal components analysis(PCA) in the time and frequency domains. Four PCA models based on head-related impulse responses(HRIRs), complex-valued HRTFs, augmented HRTFs, and log-magnitudes of HRTFs are investigated. The objective of this study is to compare modeling performances of the PCA models in the least-squares sense and to show the theoretical relationship between the PCA models. In terms of the number of principal components needed for modeling, the PCA model based on HRIR or augmented HRTFs showed more efficient modeling performance than the PCA model based on complex-valued HRTFs. The PCA model based on HRIRs in the time domain and that based on augmented HRTFs in the frequency domain are shown to be theoretically equivalent. Modeling performance of the PCA model based on log-magnitudes of HRTFs cannot be compared with that of other PCA models because the PCA model deals with log-scaled magnitude components only, whereas the other PCA models consider both magnitude and phase components in linear scale.

A high voltage damped oscillating wave generator with variable frequency has been developed using power devices (IGBT), photo devices, a designed noncore transformer and a function-generator in order to diagnose insulating materials as power apparatus (transformer, cable), enameled wires, printed circuit board (PCB), wire-harness in automobiles. The generated maximum voltage (2.5 kV) and the frequency dependency (0.01-2 kHz) were confirmed in output tests.

Efficient prediction of individual head-related transfer functions based on 3D meshes

This paper presents a deep-learning technique for modeling Head-related Transfer Function (HRTF) using a large data-set from IRCAM [7] of 49 subjects with 115 directions/subject. We use a log-spectral distortion metric that we have used to optimize the performance of the deep-learning model. The presented model first uses a sparse autoencoder (AE) for capturing the salient representation (peaks, and notches for elevation) from the user-captured HRTFs, in its latent representation, and then uses a generalized regression neural network (GRNN) to approximate the latent representation of the corresponding HRTFs at arbitrary angles. The sparse AE then decodes the GRNN output to reconstruct the HRTF. The proposed approach is shown to outperform, objectively using the log-spectral distortion, the state-of-the-art linear-least-squares optimum solution (viz., derived from Principal Component Analysis) to perform better generalization over the HRTFs in the IRCAM dataset. Towards this, Bayesian optimization has been performed to tune the hyper-parameters of the sparse AE and the GRNN.

The work quantifies the anthropometric dependency of the head-related transfer functions(HRTFs). The human body modifies the soundwaves while traveling to the ear canal from different spatial locations, and the HRTF defines the transformation. The peaks and notches in the HRTF magnitude spectrum, generated by the complex asymmetrical shape of the outer ear (pinna), are essential for sound localization in the elevation plane. However, the anthropometric dependency of HRTF is not well quantified in the literature. The huge experimental cost of conventional HRTF measurement methods is a hurdle in these studies. To overcome this, distance-dependent HRTFs have been evaluated using the finite element-based computational tool on surface meshes of the human upper body with different pinnae. The size and interaural position of the pinna were altered, and various HRTF simulations were carried out in the study. The frequency shift of the spectral peaks and notches in the magnitude spectrum was examined to understand its anthropometric dependency. The dependency of HRTF on human anthropometric features was appraised in the proposed work. It can be beneficial in predicting personalized HRTF according to the size and interaural position of the subject’s pinna.

Spatial interpolation of head-related transfer functions using enhanced generative adversarial networks

Head-related transfer functions (HRTFs) vary with both frequency and source position. The near-field HRTFs with source distance less than 1.0 m are particularly complicated due to their distance dependence. Principal component analysis (PCA), which is conventionally carried out in the frequency or time domain, has been widely applied to reduce the dimensionality of far-field HRTF data with source distance greater than 1.0 m. The present work firstly extends the conventional PCA, and then proposes a spatial PCA method in the spatial domain rather than in the frequency or time domain to reduce the dimensionality of near-field HRTF data. An illustrative case indicates that near-field HRTF magnitudes at 9 distances with 493 directions for each distance can be approximately represented by the weighted sum of 15 spectral shape basis functions using the conventional PCA or the weighted sum of 15 spatial basis functions using the spatial PCA. Both representations account for more than 98 % energy variation of the original data, and reduce the dimensionality of the original data to about a quarter. The proposed method is also applicable to the head-related impulse responses in the time domain. Furthermore, the spatial PCA scheme is potentially applicable to simplify near-field HRTF measurement.

3-dimensional sound field with high accuracy is an important topic for MR (Mixed Reality) to create the high immersion sense. HRTF (Head-Related Transfer Function) can achieve 3-dimensional sound field, whereas the measurement of the HRTF for each person is difficult because the measurement of the HRTF requires the expensive equipment and environment. In this paper, we propose a method to design HRTF by piecing several HRTFs selected from a large database of HRTFs. As a result of evaluation experiments, we confirmed that the proposed method can design the useful HRTF.

An apparatus for implementing a 3-dimensional virtual sound and method thereof are disclosed, in which computational and storage complexity are reduced, in which system stability is secured, and by which the 3-dimensional virtual sound can be implemented in such a mobile platform failing to be equipped with expensive instruments for the implementation of the 3-dimensional sound as a mobile communication terminal and the like. The present invention includes a first step of giving an inter-aural time delay (ITD) to at least one input sound signal, a second step of multiplying output signals of the first step by principal component weight, and a third step of filtering result values of the second step by a plurality of low-order approximated IIR filter models of basis vectors extracted from a head related transfer function (HRTF). The basis vectors, extracted from the head related transfer function database are approximated using balanced model approximation technique.

While various methods for sound source localization have been developed, most of them utilize on the time difference of arrival (TDOA) between microphones or the measured head related transfer functions (HRTF). In case of a real robot implementation, the former has a merit of light computation load to estimate the sound direction but can not consider the effect of platform on TDOAs, while the latter can, because characteristics of robot platform are included in HRTF. However, the latter needs large resources for the HRTF database of a specific robot platform. We propose the compensation method which has the light computation load while the effect of platform on TDOA can be taken into account. The proposed method is used with spherical head related transfer function (SHRTF) on the assumption that robot platform, for example a robot head, installed microphones can be modeled to a sphere. We verify that the proposed method decreases the estimation error caused by the robot platform through the simulation.

The Head Related Transfer Function (HRTF) means a process related to acoustic transmission from 3d space to the listener`s ear. In other words, it contains the information that human can perceive locations of sound sources. So, we make virtual 3d sound using HRTF, despite it doesn`t actually exist. But, it can deteriorate some three-dimensional effect by the confusion between front and back directions due to the non-individual HRTF depending on each listener. In this paper, we proposed the new algorithm to reduce the confusion of sound image localization using human`s acoustic characteristics. The frequency spectrum and global masking threshold of 3d sounds using HRTF are used to calculate the psychoacoustical differences among each directions. And perceptible cues in each critical band are boosted to create effective 3d sound. As a result, we can make the improved 3d sound, and the performances are much better than conventional methods.

The significant amount of variance in head-related transfer functions (HRTFs) resulting from source location and subject dependencies have led researchers to use principal components analysis (PCA) to approximate HRTFs with a small set of basis functions. PCA minimizes a mean-square error, and consequently may spend modeling effort on perceptually irrelevant properties. To investigate the extent of this effect, PCA performance was studied before and after removal of perceptually irrelevant variance. The results indicate that from the sixth PCA component onward, a substantial amount of perceptually irrelevant variance is being accounted for.