Audio Data Toolbox - WiGE

... audio data management

Whilst working with speech data, be it creating new synthetic voices, training speech recognizers, seeking new and informative characteristics of the speech signal, or any other task you can imagine, you have to manage a great number of files with audio data, labels and features that are highly intertwined. You must also be aware of dependencies and inheritances as well as undertaking many routine tasks. In order to maintain an overview of the above-mentioned tasks and to ease the entire process, we developed WiGE for you!

Features

upwards

• Constructed upon a flexible framework with the aid of the widely-used scripting language TCL
• The possibility to split the working process into classes relevant to the management of speech data (context, element, collection, export, etc.)
• Maintain the connections between different data types (which context does this element came from?)
• Automation of all routine tasks
• Automatically create lists of different elements (diphone, triphone) out of the contexts
• Search for the collected elements or contexts with sophisticated queries
• Use the wavesurfer from KTH to examine audio data in depth
• Support for many audio and other related formats e.g. wav, bin, au, phondat.
• Labelling of ESPS, phondat, HTK and many other types of audio data

These current features can easily be broadened and tailored according to your needs.

WiGE helps you to focus on the fundamental tasks at hand!