Neural network for multi-exponential sound energy decay analysis

TL;DR

This paper introduces a neural network method for accurately estimating parameters of multi-exponential sound energy decay functions from measurements, demonstrating robustness and efficiency across diverse acoustic environments.

Contribution

It presents a novel neural network approach trained on synthetic data to estimate sound energy decay parameters from large real-world datasets, improving robustness and computational efficiency.

Findings

Robust estimation of decay parameters across various environments

Lightweight neural network architecture for real-time analysis

Publicly available implementation for broader use

Abstract

An established model for sound energy decay functions (EDFs) is the superposition of multiple exponentials and a noise term. This work proposes a neural-network-based approach for estimating the model parameters from EDFs. The network is trained on synthetic EDFs and evaluated on two large datasets of over 20000 EDF measurements conducted in various acoustic environments. The evaluation shows that the proposed neural network architecture robustly estimates the model parameters from large datasets of measured EDFs, while being lightweight and computationally efficient. An implementation of the proposed neural network is publicly available.