This paper presents an indexing and retrieval method for podcasts where transcripts exist. The indexing units are speech segments limited by speech pauses. Podcasts and their corresponding texts are synchronized automatically in terms of these indexing units by the software SpeechIndexer. A web-based interactive player function makes each podcast fully accessible. This is demonstrated with PodClub – a podcast service for language learning offered by the largest language school in Switzerland. The interactive player received enthusiastic feedback in a pilot test phase and is online since January 2014. The search across podcast archives is performed with SpeechConcordancer, which belongs to the SpeechIndexer software suite.
Ulrike Glavitsch, Dennis Küpper, Tobias Stamm, Jozsef Szakos, "Podcast Archives: Access Through SpeechIndexer Technology" in Proc. IS&T Archiving 2014, 2014, pp 197 - 200, https://doi.org/10.2352/issn.2168-3204.2014.11.1.art00044