Blind Acoustic Parameter Estimation Through Task-Agnostic Embeddings Using Latent Approximations

Götz P, Tuna C, Brendel A, Walther A, Habets E (2024)

Publication Type: Conference contribution

Publication year: 2024

Publisher: Institute of Electrical and Electronics Engineers Inc.

Pages Range: 289-293

Conference Proceedings Title: 2024 18th International Workshop on Acoustic Signal Enhancement, IWAENC 2024 - Proceedings

Event location: Aalborg, DNK

ISBN: 9798350361858

DOI: 10.1109/IWAENC61483.2024.10694126

Abstract

We present a method for blind acoustic parameter estimation from single-channel reverberant speech. The method is structured into three stages. In the first stage, a variational auto-encoder is trained to extract latent representations of acoustic impulse responses represented as mel-spectrograms. In the second stage, a separate speech encoder is trained to estimate low-dimensional representations from short segments of reverberant speech. Finally, the pre-trained speech encoder is combined with a small regression model and evaluated on two parameter regression tasks. Experimentally, the proposed method is shown to outperform a fully end-to-end trained baseline model.

Authors with CRIS profile

Philipp Götz International Audio Laboratories Erlangen (AudioLabs) Emanuël Habets Lehrstuhl für wahrnehmungsbasierte räumliche Audiosignalverarbeitung (AudioLabs)

Involved external institutions

Fraunhofer-Institut für Integrierte Schaltungen (IIS) DE Germany (DE)

How to cite

APA:

Götz, P., Tuna, C., Brendel, A., Walther, A., & Habets, E. (2024). Blind Acoustic Parameter Estimation Through Task-Agnostic Embeddings Using Latent Approximations. In 2024 18th International Workshop on Acoustic Signal Enhancement, IWAENC 2024 - Proceedings (pp. 289-293). Aalborg, DNK: Institute of Electrical and Electronics Engineers Inc..

MLA:

Götz, Philipp, et al. "Blind Acoustic Parameter Estimation Through Task-Agnostic Embeddings Using Latent Approximations." Proceedings of the 18th International Workshop on Acoustic Signal Enhancement, IWAENC 2024, Aalborg, DNK Institute of Electrical and Electronics Engineers Inc., 2024. 289-293.

BibTeX: Download