Neural Network Program Package for Prosody Modeling

Abstract

This contribution describes the programme for one part of the automatic Text-to-Speech (TTS) synthesis. Some experiments (for example [14]) documented the considerable improvement of the naturalness of synthetic speech, but this approach requires completing the input feature values by hand. This completing takes a lot of time for big files. We need to improve the prosody by other approaches which use only automatically classified features (input parameters). The artificial neural network (ANN) approach is used for the modeling of prosody parameters. The program package contains all modules necessary for the text and speech signal pre-processing, neural network training, sensitivity analysis, result processing and a module for the creation of the input data protocol for Czech speech synthesizer ARTIC [1].