AES Fall Convention 2023

Thanks to the development and wide adoption of the SOFA file format, researchers and developers can access multiple published repositories of HRTF data. However, due to a lack of standardization for the capture, post-processing and publication of datasets, publicly available databases cannot be readily aggregated into extensive corpora. Areas of discrepancy include spatial distribution, measurement distances, temporal alignment, spectral equalization, and extrapolation at low frequencies (where acoustic measurements are often not reliable). In this session, panelists will share their experience and perspectives on the collection and processing of measured HRTF data, their application to the realization of binaural rendering systems, and their use in machine learning and analytics for HRTF personalization and individualization. An open initiative is considered to facilitate interoperability and data aggregation across existing and future HRTF databases.