Year of the project: 2014
A text-to-speech (TTS) system converts normal language text into speech. This is part of my diploma thesis, which I successfully defended. I chose English as a language for speech synthesis.

Neural network based approach was used to solve grapheme-to-phoneme conversion problem.

Phonemes duration prediction problem was solved by means of stochastical model.

Speech signal generation problem was solved using the digital signal processing module MBROLA.