The aim of this paper was to create an automatic sound source classification framework for recordings captured with a microphone array and evaluate the sound source separation algorithm impact on the classification results. The preprocessing related to the said evaluation concerned convolving the dataset samples with impulse responses captured with a microphone array, as well as mixing the samples together to simulate their co-presence in a virtual recording scene. This way, the evaluation of the separation algorithm impact on classification results was possible. Furthermore, such approach saved multiple hours of labour that would need to be spent on the recording process itself. Finally, the classification results delivered by different models were evaluated and compared.

Authors: Chrul, Michal; Ruminski, Andrzej; Zernicki, Tomasz; Lukasik, Ewa
Affiliations: Zylia sp. z o. o., Poznan, Poland; Gdansk University of Technology, Gdansk, Poland(See document for exact affiliation information.)
AES Convention: 150 (May 2021) Paper Number: 10481
Publication Date: May 24, 2021 Import into BibTeX
Subject: Music Analysis
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=21074

Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!

This paper costs $33 for non-members and is free for AES members and E-Library subscribers.

Learn more about the AES E-Library

E-Library Location:   /conv/150/10481.pdf

Start a discussion about this paper!