shva facilitates careful listening and precise alignment of the annotations against the acoustic signals. It is designed for accessing speech databases over the web. It opens from the browser when the user clicks on a link for an utterance. It then runs as a separate application.
The underlying data format is ALingA (Aligned Linguistic Annotations), a multi-stream format designed for interleaving linguistic annotations with the acoustic or other derived signals for web transport. The annotation part of ALingA is based on the annotation graph model of Bird and Lieberman (2001). shva provides the GUI to manipulate the nodes and arcs.