Compact statistical encoding of early reflections for auditory virtual reality

Raghuvanshi N, Allen A, Snyder J, Chemistruck M, Cross N, Walker C, Willette A, McWilliams N, Myrbeck S. Compact statistical encoding of early reflections for auditory virtual reality. In: AES Conference: AES 2022 International Audio for Virtual and Augmented Reality Conference. Audio Engineering Society; 2022. Paper 36. Available from: https://aes.org/publications/elibrary-page/?id=21866

@inproceedings{Raghuvanshi2022_21866,
  author = {Raghuvanshi, Nikunj and Allen, Andrew and Snyder, John and Chemistruck, Michael and Cross, Noel and Walker, Christopher and Willette, Aaron and McWilliams, Nathan and Myrbeck, Shane},
  title = {{Compact statistical encoding of early reflections for auditory virtual reality}},
  booktitle = {AES Conference: AES 2022 International Audio for Virtual and Augmented Reality Conference},
  note = {Paper 36},
  year = {2022},
  month = aug,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=21866}
}

TY  - CPAPER
TI  - Compact statistical encoding of early reflections for auditory virtual reality
AU  - Raghuvanshi, Nikunj
AU  - Allen, Andrew
AU  - Snyder, John
AU  - Chemistruck, Michael
AU  - Cross, Noel
AU  - Walker, Christopher
AU  - Willette, Aaron
AU  - McWilliams, Nathan
AU  - Myrbeck, Shane
T2  - AES Conference: AES 2022 International Audio for Virtual and Augmented Reality Conference
M1  - Paper 36
PY  - 2022
DA  - 2022/08/06
UR  - https://aes.org/publications/elibrary-page/?id=21866
PB  - Audio Engineering Society
LA  - en
AB  - Early reflections pose a major challenge for parametric virtual acoustics systems. Reflections within a spatial impulse response (IR) must be compactly encoded to allow interactive rendering within limited resources. Encoding individual echoes is far too expensive, while the exponentially-decaying-noise model only befits late reverberation. We propose a novel statistical formulation for compactly encoding salient early reflection properties with six interpretable, orthogonal parameters capturing the loudness, echo density, temporal mean & spread, and directional mean & spread of energy arrival within the IR. Further, we propose a robust approach for additive separation of the impulse response into a coherent (specular) and incoherent (diffuse) component, rather than partitioning at a mixing time. We discuss initial results on a variety of measured IRs in indoor and outdoor spaces suggesting that salient acoustical variations across scenes are captured by the proposed parameters.
ER  -