Audio-Based Tools for Decontextualisation Detection

This chapter addresses the underexplored challenge of detecting audio decontextualisation, the misleading reuse of audio recordings outside their original context. While visual media benefit from established tools for provenance tracking and verification, comparable methods for audio remain limited, despite its high persuasive impact. A central focus of this chapter is audio provenance analysis, a key method for tracing the origin, reuse, and transformation of audio segments. We present provenance analysis techniques for audio reuse detection, clustering, and transformation tracking, and demonstrate their application in real-world use cases. Complementing this, we explore methods for context analysis, such as inferring recording locations from acoustic evidence. The chapter concludes with an outlook on future extensions and multi-modal approaches, including the integration of textual information derived from speech transcriptions.