Datasets

1. Ressources for Text-to-Speech Synthesis:
   • Speech/text alignments for French end-to-end TTS: 5 speakers (60h15'') DOI
   • Ressources for End-to-End French Text-to-Speech Blizzard challenge: 33 speakers (33h22'') DOI

Softwares

1. Github for Tacotron2 with extensions: sTacotron2

Grenoble Images Parole Signal Automatique laboratoire

UMR 5216 CNRS - Grenoble INP - Université Joseph Fourier - Université Stendhal