Differentiable Acoustic Radiance Transfer
Abstract
Geometric acoustics is an efficient framework for room acoustics modeling, governed by the canonical time-dependent rendering equation. Acoustic radiance transfer (ART) solves the equation by discretization, modeling time- and direction-dependent energy exchange between surface patches with flexible material properties. We introduce DART, an efficient, differentiable implementation of ART that enables gradient-based optimization of material properties. We evaluate DART on a simpler variant of acoustic field learning that aims to predict energy responses for novel source-receiver configurations. Experimental results demonstrate that DART generalizes better under sparse measurement scenarios than existing signal processing and neural network baselines, while maintaining simplicity and full interpretability. We open-source our implementation.
Cite
@article{arxiv.2509.15946,
title = {Differentiable Acoustic Radiance Transfer},
author = {Sungho Lee and Matteo Scerbo and Seungu Han and Min Jun Choi and Kyogu Lee and Enzo De Sena},
journal= {arXiv preprint arXiv:2509.15946},
year = {2026}
}
Comments
Accepted to TASLPRO