Font Size: a A A

Measurement Of Anthropometric Parameters And Individualization Of Head-Related Transfer Function

Posted on:2019-07-27Degree:MasterType:Thesis
Country:ChinaCandidate:R X WuFull Text:PDF
GTID:2428330566986449Subject:Acoustics
Abstract/Summary:PDF Full Text Request
Head-related transfer functions(HRTFs)are essential for virtual auditory display system.One of the effective way to enhance the performance of virtual auditory display system is employing the individualized HRTFs.There is difficulty in practice to obtain accurate individualized HRTFs by experiment or numerical simulation since it is both time consuming and costly.Relatively feasible approach is that individualization of HRTFs based on anthropometric parameters.However,there are still deficiencies in individualization of HRTF based on anthropometric parameters,such as lack of standard on anthropometric parameters measurement,insufficient subject with complete anthropometric parameters and incomplete spatial sampling of HRTF.In the present work,individualization of HRTF based on anthropometric parameters is studied systematically.First,three-dimensional(3D)scanned head models are obtained using laser 3D scanner.A set of anthropometric parameters is defined on the 3D scanned head model and the process of measurement of the whole set of parameters is designed.An anthropometric parameter database including 62 subjects is thereby setup.Further,corresponding individualized HRTFs with fullspace sampling are simulated using boundary element method(BEM).Principle component analysis(PCA)is applied to extract the effective features and reduce the dimension of HRTFs.Further,a method to select the key anthropometric parameters based on correlation analysis is proposed.8 anthropometric parameters are selected as the key parameters that correlates with individualized HRTFs.Further,individualized method of HRTFs based on anthropometric parameters is proposed.Individualized model based on GRNN is setup.Spectral distortion analysis and the psychoacoustics experiments show that the method proposed individualize the HRTF for new subjects effectively.The individualized method of HRTFs proposed in the present work is able to individualize HRTFs for new subjects by 8 key anthropometric parameters and help to enhance the performance of virtual auditory display system.Thus,it is considered to have good application prospect.
Keywords/Search Tags:Head-related transfer function, anthropometric parameters, individualized, general regression neural network
PDF Full Text Request
Related items