Audio Meta Data Transcription from Meeting Transcripts for the Continuous Media Web
Steve Cassidy, Silvia Pfeiffer · 2004
The Continuous Media Web (CMWeb) integrates time{continuous media into the searching, linking, and browsing functionality of the World Wide Web. The le format underlying the CMWeb technology, Annodex, streams the media content multiplexed with XML{markup in the Continuous Media Markup Language (CMML). CMML contains information relevant to the whole media le (e.g., title, author, language) as well as time{sensitive information (e.g., topics, speakers, time{sensitive hyperlinks). This paper discusses the challenges of automatically generating Annodex streams from complex annotated recordings collected for use in linguistic research. We are particularly interested in annotated recordings of meetings and teleconferences and regard Annodex and its media browsing paradigm as a novel and rich way of interacting with such recordings. The paper presents our experiments with generating CMML and their corresponding Annodex les from hand annotated meeting recordings.