Subspace-Based HRTF Synthesis from Sparse Data: A Joint PCA and ML-Based Approach

Bharitkar SG, Mauer T, Wells T, Berfanger D. Subspace-Based HRTF Synthesis from Sparse Data: A Joint PCA and ML-Based Approach. In: AES Convention 145. Audio Engineering Society; 2018. Paper 476. Available from: https://aes.org/publications/elibrary-page/?id=19740

@inproceedings{Bharitkar2018_19740,
  author = {Bharitkar, Sunil G. and Mauer, Timothy and Wells, Teresa and Berfanger, David},
  title = {{Subspace-Based HRTF Synthesis from Sparse Data: A Joint PCA and ML-Based Approach}},
  booktitle = {AES Convention 145},
  note = {Paper 476},
  year = {2018},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=19740}
}

TY  - CPAPER
TI  - Subspace-Based HRTF Synthesis from Sparse Data: A Joint PCA and ML-Based Approach
AU  - Bharitkar, Sunil G.
AU  - Mauer, Timothy
AU  - Wells, Teresa
AU  - Berfanger, David
T2  - AES Convention 145
M1  - Paper 476
PY  - 2018
DA  - 2018/10/06
UR  - https://aes.org/publications/elibrary-page/?id=19740
PB  - Audio Engineering Society
LA  - en
AB  - Head-related transfer functions (HRTF) are used for creating the perception of a virtual sound source at an arbitrary azimuth-elevation. Publicly available databases use a subset of these directions due to physical constraints (viz., loudspeakers for generating the stimuli not being point-sources) and the time required to acquire and deconvolve responses for a large number of spatial directions. In this paper we present a subspace-based technique for reconstructing HRTFs at arbitrary directions for the IRCAM-Listen HRTF database, which comprises a set of HRTFs sampled every 15 deg along the azimuth direction. The presented technique includes first augmenting the sparse IRCAM dataset using the concept of auditory localization blur, then deriving a set of P=6 principal components, using PCA for the original and augmented HRTFs, and then training a neural network (ANN) with these directional principal components. The reconstruction of HRTF corresponding to an arbitrary direction is achieved by post-multiplying the ANN output, comprising the estimated six principal components, with a frequency weighting matrix. The advantage of using a subspace approach, involving only 6 principal components, is to obtain a low complexity HRTF synthesis ANN-based model as compared to training an ANN model to output an HRTF over all frequencies. Objective results demonstrate a reasonable interpolation with the presented approach.
ER  -