A high-resolution head-related transfer function and three-dimensional ear model database

Bomhardt, Ramona; Klein, Mathias; Fels, Janina

doi:10.1121/2.0000467

Cited by 39 publications

(33 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…year # subjects # meas. # pairs ARI˚ [23] 2010 135 1150 138000 CIPIC [24] 2001 45 1250 56250 ITA˚˚ [25] 2016 46 2304 110592 Microsoft [9] 2015 252 400 100800 RIEC [26] 2014 105 865 90825…”

Section: Classification Performancementioning

confidence: 99%

Spatial Audio Feature Discovery with Convolutional Neural Networks

Thuillier

Gamper

Tashev

2018

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

The advent of mixed reality consumer products brings about a pressing need to develop and improve spatial sound rendering techniques for a broad user base. Despite a large body of prior work, the precise nature and importance of various sound localization cues and how they should be personalized for an individual user to improve localization performance is still an open research problem. Here we propose training a convolutional neural network (CNN) to classify the elevation angle of spatially rendered sounds and employing Layerwise Relevance Propagation (LRP) on the trained CNN model. LRP provides saliency maps that can be used to identify spectral features used by the network for classification. These maps, in addition to the convolution filters learned by the CNN, are discussed in the context of listening tests reported in the literature. The proposed approach could potentially provide an avenue for future studies on modeling and personalization of head-related transfer functions (HRTFs).

show abstract

Section: Classification Performancementioning

confidence: 99%

Spatial Audio Feature Discovery with Convolutional Neural Networks

Thuillier

Gamper

Tashev

2018

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…To estimate HRTF data sets on the basis of anthropometric dimensions, the weighting scores W for the PCs have to be expressed by a linear combination of the anthropometric dimensions (5). 10 The quality of this estimation is discussed in this section by a randomly-chosen subject first and by the subject-averaged ISSD between the original and estimated data sets of 47 subjects subsequently.…”

Section: E Anthropometric Estimation Of the Spectrummentioning

confidence: 99%

“…These dimensions are taken from the existing database 5 and are measured according to the specifications in Fig. 6.…”

Section: Regression Analysismentioning

confidence: 99%

“…A data base 5 with 47 individual HRTF data sets is used as input data for the PCA. These data sets contain 2 304 directions for each ear so that the PCA contains n = 216 576 observations with p = 129 frequency bins.…”

Section: A Principal Componentsmentioning

confidence: 99%

“…Often only the magnitude is taken into account where the phase has to be estimated using a minimum phase and a delay, 19 which can be estimated by an interaural time difference (ITD) model. 18 The present study is based on a ITA HRTF database 5 of 47 subjects which contains highresolution HRTF data sets as well as three-dimensional ear models and head dimensions. Using this database, the reconstruction of individual HRTF data sets is studied in the following section.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations