Towards an audio attribute framework for understanding the perception of reverberant spaces through elicitation and clustering methods based on participant expectation

Child, Luke; Ford, Natanya

Abstract

When used in perceptual audio evaluation, elicitation methods produce a wide variety of raw and unorganized text data. Although at first ambiguous, elicited data can be organized into themes and attributes that are intrinsic to the listener experience. This paper seeks to compare the trends found in descriptions of reverberant locations from memory, isolating key attributes and phrases present in descriptions. These attributes are then cleaned, validated, and clustered to form a series of key parent attributes that encompass the descriptions of the original attributes. Methods for the optimization of each stage are discussed, alongside applications for understanding and utilizing the attributes in future implementations of digital reverberation.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: Audio Engineering Society Convention 150
• Start Date: May 25, 2021
• End Date: May 28, 2021
• Acceptance Date: Apr 19, 2021
• Online Publication Date: May 25, 2021
• Publication Date: 2021
• Deposit Date: May 28, 2021
• Publisher: Audio Engineering Society
• Book Title: Audio Engineering Society Convention 150
• Public URL: https://uwe-repository.worktribe.com/output/7428941
• Publisher URL: http://www.aes.org/e-lib/browse.cfm?elib=21060
• Related Public URLs: https://www.aes.org/tmpFiles/elib/20210528/21060.pdf