Visualisation in speech corpora: maps and waves in the Glossa system
Chapter by Michał Kosek, Anders Nøklestad, Joel Priestley, Kristin Hagen and Janne Bondi Johannessen in Proceedings of the Workshop on Innovative Corpus Query and Visualization Tools at NODALIDA 2015, 2015.
We present the Glossa web-based system for corpus search and results handling, focussing on two modes of visualisation implemented in the system. First, we describe the use of maps to show the geographical distribution of search results and its utility for exploring dialectal variation and discovering new isoglosses. Secondly, we present a functionality for speech visualisation, yielding dynamically generated representations of spectrograms, pitch and formants. The analyses are accompanied by the ability to replay selected parts of the waveform, as well as export and compare maximum, minimum and average values of the parameters for different selections. Among other things, this can be used to explore in more detail the set of spoken variants revealed by the geographical map view.