Neural network for multi-exponential sound energy decay analysis

Georg Götz*, Ricardo Falcon Perez, Sebastian Schlecht, Ville Pulkki

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

2 Citations (Scopus)
5 Downloads (Pure)

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 20 000 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.
Original languageEnglish
Pages (from-to)942-953
Number of pages13
JournalJournal of the Acoustical Society of America
Volume152
Issue number2
DOIs
Publication statusPublished - 12 Aug 2022
MoE publication typeA1 Journal article-refereed

Keywords

  • sound energy decay analysis
  • multi-exponential sound decay
  • reverberation time
  • neural network

Fingerprint

Dive into the research topics of 'Neural network for multi-exponential sound energy decay analysis'. Together they form a unique fingerprint.

Cite this