WWWSigTranscribe is a tool for annotating single and multichannel speech signals via the WWW. It was originally developed for the SpeechDat(II) project for an orthographic annotation of telephone speech, and has been extended to handle standard (WAV, AIFF, au) as well as generic binary audio file formats.
The major motivation for providing annotation via the WWW is
platform independence and world wide access to speech data, and central storage of annotations and descriptive data.
This means that e.g. students can access their labs data from home. Annotations, lexicons, and other descriptive data are held on the speech server; updates e.g. of the lexicon are thus immediately available to the annotators.
WWWSigTranscribed is implemented in Java in a clean object-oriented design. It features a signal display window and an editing pane with editing buttons. These editing buttons speed up the annotation process by performing common conversions. Syntactical correctness of the annotation is checked, and a dictionary lookup is performed prior to saving the annotation to a DBMS.
Due to its object-oriented design, WWWSigTranscribe can easily be adapted to new file formats or languages. Because it is implemented in pure Java it supports Unicode - i.e. IPA symbols or ideographic languages, and can make use of the standard interface to relational DBMSs.