no code implementations • LREC 2018 • Ingmar Steiner, Sébastien Le Maguer
We present a new workflow to create components for the MaryTTS text-to-speech synthesis platform, which is popular with researchers and developers, extending it to support new languages and custom synthetic voices.