Gesture-Based Articulatory Text-to-Speech Synthesis

Benjamin Weitz, Ingmar Steiner, Peter Birkholz

In: Jürgen Trouvain, Ingmar Steiner, Bernd Möbius (editor). 28th Conference on Electronic Speech Signal Processing (ESSV). Elektronische Sprachsignalverarbeitung (ESSV) March 15-17 Saarbrücken Germany Pages 324-331 TUD Press Dresden 3/2017.


We present work carried out to extend the text to speech (TTS) platform MaryTTS with a back-end that serves as an interface to the articulatory synthesizer VocalTractLab (VTL). New processing modules were developed to (a) convert the linguistic and acoustic parameters predicted from orthographic text into a gestural score, and (b) synthesize it to audio using the VTL software library. We also describe an evaluation of the resulting gesture-based articulatory TTS, using articulatory and acoustic speech data.

Weitere Links

Weitz.pdf (pdf, 2 MB )

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz