Audio Data Toolbox - WiGE

Manage Audio Data Easily and Clearly.

Screenshots of software WiGe

Working with speech data in order to create synthetic voices, to train speech recognizers or to search for new features of the speech signal requires the management of a huge number of audio and data files as well as labels and features. Most of the time these files, labels and features are highly intertwined and there are dependencies, which need to be considered.

Product Features

  • Flexible framework implemented with TCL
  • Platform independent (Windows, Linux, OS X)
  • Possibility of splitting up the workflow into different categories which are important for managing speech data (context, elements, collections, export etc.)
  • Maintains cross relations between data types ( What context did this element come from?)
  • Automation of routine tasks
  • Automated list creation of different elements (like diphones, triphones) out of contexts
  • Several search opportunities
  • Use of the wavesurfer by KTH to examine audio data in depth
  • Support of many formats: wav, bin, au, phondat
  • Labeling: ESPS, phondat, HTK and other audio data
  • Support for recording and management of mass data