Item

Abstract

While recurrence plots (RPs) are a well-known tool in various fields such as physics, astronomy, and health sciences, their application in audio signal processing remains limited. RPs are a data-analysis tool that visualizes recurrences of states, which are typically measured by the Euclidean norm. When analyzing audio data, however, β-divergences are more common than the Euclidean norm because of their adaptability and suitability for audio-specific characteristics. Therefore, we propose the use of β-divergence-based RPs to gain additional insight into audio data. In this paper, we explore the properties of such RPs, providing a fundamental understanding of their characteristics and an indication of possible future applications. Our findings show that β-divergence-based RPs can provide additional information over traditional RPs, making them well-suited for audio analysis.