Noise tagging and perceptually-informed unsupervised clustering

Abstract

Airforce training was deemed both necessary and crucial for the country. However, the persistent issue of aircraft noise had long been a concern for residents living near the air base. This study aimed to explore the realm of sound analysis and data processing concerning the diverse and loud acoustic disturbances experienced by the residents. The objective was to investigate clustering techniques that could be applied for perceptually-informed unsupervised clustering of sound events, allowing for the automatic analysis of complex audio data. The study involved the development and execution of methods for extracting specific urban sound events from long-term recordings, which addressed the analysis measures used to detect or identify sound events.