How representative are Dummy head HRTFs? A subjective comparison of mannequin and human datasets

Katz BFG, Bellows SD. How representative are Dummy head HRTFs? A subjective comparison of mannequin and human datasets. In: AVARIG 2026: Audio for Virtual and Augmented Reality and Immersive Games. Audio Engineering Society; 2026. Paper 502. Available from: https://aes.org/publications/elibrary-page/?id=23349

@inproceedings{Katz2026_23349,
  author = {Katz, Brian F.G. and Bellows, Samuel D.},
  title = {{How representative are Dummy head HRTFs? A subjective comparison of mannequin and human datasets}},
  booktitle = {AVARIG 2026: Audio for Virtual and Augmented Reality and Immersive Games},
  note = {Paper 502},
  year = {2026},
  month = jun,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=23349}
}

TY  - CPAPER
TI  - How representative are Dummy head HRTFs? A subjective comparison of mannequin and human datasets
AU  - Katz, Brian F.G.
AU  - Bellows, Samuel D.
T2  - AVARIG 2026: Audio for Virtual and Augmented Reality and Immersive Games
M1  - Paper 502
PY  - 2026
DA  - 2026/06/30
UR  - https://aes.org/publications/elibrary-page/?id=23349
PB  - Audio Engineering Society
LA  - en
AB  - Head-related transfer functions (HRTFs) are central to binaural rendering in virtual and augmented reality. Whilst individual HRTFs offer the highest perceptual fidelity, the practical difficulty of personal HRTF acquisition drives widespread use of dummy-head (mannequin) measurements as a non-individualised substitute. Despite their ubiquity, systematic perceptual benchmarking of dummy-head HRTFs against human HRTFs remains limited. This study applies an established trajectory-based quality assessment paradigm to a mixed set of dummy-head and human HRTFs. Participants rated the perceptual quality of binaural renderings of predefined auditory trajectories, with HRTFs presented in randomised order across two sets of eight, with repeated items serving as inter-set normalisation anchors. The HRTF pool encompassed 7 human measurements alongside 7 dummy-head types: simplified head-only geometries, head-and-torso simulators (HATS), and models incorporating absorptive materials. The primary research question is whether, despite well-documented listener-dependent variability in HRTF suitability, population-level trends differentiate dummy-head HRTFs from human ones, and further, whether acoustic complexity of the mannequin correlates with perceptual performance. Results are discussed in terms of implications for HRTF database design and substitute HRTF selection strategies for immersive audio applications.
ER  -