|
|
Home |
|
speak.plug |
|
|
v.2.0.5 |
Mbrola speech for the Wavesurfer |
|
|
|
|
|
|
What does it do? The speak.plug adds Mbrola speech functionality to Wavesurfer. Since Mbrola can only accept phoneme labels and not straight text, it is not a Text-to-Speech application but rather a tool for linguists and Mbrola experimenters. The speak.plug v.2.0.5 adds to Wavesurfer the following functionality:
|
|
Contact: dzach_at_hol.gr subject=speak.plug
|
|
|
Wavesurfer is a sound visualization and manipulation tool, writen in tcl. It is Open Source, fast and extendable and uses the sound processing tcl library, Snack. Wavesurfer and Snack are actively maintained by Kåre Sjölander and Jonas Beskow at the Center for Speech Technologie (CTT) at KTH in Stockholm. One of the typical uses of Wavesurfer is the transcription of spoken language. For the transcription process, a user inserts proper graphemes or labels at specific positions in time, along a recorded speech track, so that the start and end of a label , is aligned with its corresponding speech segment. The label and timing data can be saved and used for speech analysis by other applications.
|
|||
|
|
Mbrola is a free speech engine producing high quality speech. It uses diphone voice databases, and comes with 62 different voices for about 27 languages. A large number of free applications, including Text-to-Speech readers, has been developed for Mbrola. Mbrola is extensively used by researchers and Universities and is a project maintained by the team of Thierry Dutoit of the Faculté Polytechnique de Mons,in Belgium.
|
|||
|
The speak.plug |
|
The speak.plug plugin adds speech functionality with natural voice intonation to Wavesurfer. It provides a connection to the Mbrola engine, to which it passes phoneme and voice intonation data entered or generated in Wavesurfer. The speak plugin is Open Source, under the GPL license, and works along with Wavesurfer and Mbrola on many platforms, including Windows and Linux. The speak.plug works together with transcription panes in wavesurfer. Each transcription pane can have its own voice. Right clicking on a transcription pane brings up the popup menu where the item "Speak" plays the transcription to the speakers. Clicking on a pwline pane with prosody targets plays the transcription with the local prosody. See more in the "Installation" and "Soundshots" pages. Mentioning of use of this tool in scientific or other work will be much appreciated.
|
||
|
|
|
Last modified on: 29-09-2004 |