Nick Thieberger, University of Melbourne

Well It Works! Reflections on the Audiamus Model for Corpus Building and Where It Could Go from Here

Audiamus is software that I began working on in 1998 for accessing all of my time-aligned transcripts of field recordings via a unified interface. It allows me to verify my analysis with instant access via text to a corpus (some 20 hours) of media in a previously undescribed language and to write a grammatical description in which examples and sample texts are all playable and cited to the utterance level in archival data files. These examples could then be presented on a DVD in a set of playlists, also built in Audiamus as it functions both as a means of presentation and of editing the data. Like many such tools, it was built for a particular purpose and suits me very well, but is not going to fit other ways of working with field data. In this paper I discuss the need for piecing together appropriate existing tools and of knowing what each will do within a larger workflow. I will then critique Audiamus and suggest ways in which the functions I have found so useful could be built using open source tools.


References
Thieberger, Nicholas. 2004. Documentation in practice: Developing a linked media corpus of South Efate.
Austin, Peter (ed). Language documentation and description. Volume 2. London: Hans Rausing            Endangered Languages Project, SOAS.
Audiamus: http://www.linguistics.unimelb.edu.au/thieberger/audiamus.htm