Online Lombard-adaptation in Incremental Speech Synthesis
This repository contains the software described in the paper
Rottschäfer, S., Buschmeier, H., van Welbergen, H., & Kopp, S. (2015). Online Lombard-adaptation in incremental speech synthesis. In Proceedings of INTERSPEECH 2015, pp. 80–84, Dresden, Germany.
first developed as part of Sebastian Rottschäfer’s 2014 B.Sc.-Thesis “Implementierung des Lombard-Effekts für eine inkrementelle Sprachsynthese“.
The contained LombardModule is an extension for the speech synthesis software MaryTTS. It adjusts speech parameters such as pitch, speech rate or loudness, depending on environmental noise, to improve intelligibility. It simulates the behaviour of human speakers, called Lombard effect.
Copyright (C) 2014–2015 Sebastian Rottschäfer, Social Cognitive Systems Group, CITEC, Bielefeld University
This software makes use of the following third party components:
- MaryTTS for speech synthesis
- InproTK for an incremental version of MaryTTS
- Apache Commons Math for computing Fourier Transformations
- Lightweight Java Game Library for the NoisePlayer
The easiest way to get the LombardModule working is to import it into Eclipse. InproTk also needs to be a projects in your Eclipse workspace and set up to work with MaryTTS (see the InproTk documentation).
The source code is licensed under the GNU Lesser General Public License Version 3 (LGPL).
For questions, contact the author: firstname.lastname@example.org.