Research / Game audio

Full paper · 2026

Spherical Acoustic Spatial Entropy: Predicting Acoustic Scene Complexity in Virtual Environments

Resti, L., Gully, A., McLoughlin, M., Kearney, G., & Denisova, A.

JournalIEEE Open Journal of Signal Processing, 2026

Teaser figure: Spherical Acoustic Spatial Entropy (SASE). The method splits the space around the listener into equal-area regions, weights each source by perceived loudness and measures how evenly sound is spread. In a listening experiment (N = 21), SASE best predicted perceived complexity (R² = 0.714), ahead of energy and spectral entropy; together with spectral variability it explained 84.4% of the variance.
Open full-size image

In short

How complex a soundscape feels depends not only on which sounds are present but on where they are. We introduce Spherical Acoustic Spatial Entropy (SASE), a measure that predicts how complex a virtual acoustic scene sounds from how loud each sound source is in each direction around the listener.

Key findings

  • In a listening experiment with 21 participants, SASE predicted perceived complexity better than spectral or energy entropy (R² = 0.71).
  • The relationship also held for individual participants (pseudo-R² = 0.74).
  • Combined with a measure of spectral variability, SASE explained 84% of the variance in perceived complexity, so spatial and spectral measures capture complementary aspects of a scene.

Read and cite

PDF (accepted version) Publisher's version (DOI)

APA citation

Resti, L., Gully, A., McLoughlin, M., Kearney, G., & Denisova, A. (2026). Spherical Acoustic Spatial Entropy: Predicting Acoustic Scene Complexity in Virtual Environments. IEEE Open Journal of Signal Processing.

More on this topic