Immersive and Personalised Podcasting Using AI-driven Audio Production Tools

From the first podcasts to the current diverse content landscape, there has been a drastic expansion of what we consider a podcast. Genres have emerged and died, survived, and thrived, and the podcast landscape in 2024 is nothing like the one from 2004. So what will podcasts in 2044 sound like, and what does Next-Generation Podcasting entail? After highlighting the gaps within podcasting innovation literature, this thesis proposes definitions and frameworks to advance podcast research, informed not only by context, but by involving a group of podcasters whose opinions and expectations of podcasting are gathered through interviews and workshops. This participatory methodology is not only valuable to map the current practices and production habits of professionals, but also to substantiate the development of new tools for immersive and personalised podcasting. The peculiarities and requirements linked to such tools are explored, and the iterative development process that occurs results in the creation and evaluation of a web-app for modular podcasting, Podulr, and automatic chapterisation algorithm, pod-CLIPR (Podcast Chapter Localisation through Intelligent Pattern Recognition).

The contributions of this thesis are: 1/ A definition of podcasting alongside a framework for podcasting innovation; 2/ A contemporary archetypal workflow for podcasting; 3/ A summary of expectations of producers for Next-Generation Podcasting, views on new technologies, and a reflection on the systems already in place and how they’ll need to adapt to enable it; 4/ A system for automatic podcast audio chapterisation, pod-CLIPR, comprising of a sound recognition model combined with a rule-based algorithm, and its evaluation; and 5/ A reflection on participatory design for developing media tools and a practical application in the form of the modular podcasting web-app Podulr. This work has interdisciplinary impact, in podcasting, audio production, interactive media, and participatory design for new media tools.