EstherKlabbers
Research areas: Text-to-speech synthesis (TTS), prosody, computational linguistics (NLP), machine learning
I have been doing research in the field of TTS since I started my PhD at the Technical University of Eindhoven in 1996. I am passionate about making TTS sound more natural and intelligible. In this day and age, machine learning takes a very prominent role in TTS. I created various machine learning models to be used in the TTS linguistic preprocessing module to predict pronunciations for OOV words, accentuation, and phrase boundaries. I believe that it takes more than heaps of training data to produce high-quality output out of neural networks. I am passionate about using my linguistic and prosodic knowledge to inform neural networks and generate accurate and natural sounding speech for different voices and languages in different speaking styles and emotions.