Voice Analysis Using PRAAT Software and Classification of User Emotional State

Abstract

During the last decades the field of IT has seen an incredible and very rapid development. This development has shown that it is important not only to shift performance and functional boundaries but also to adapt the way human-computer interaction to modern needs. One of the interaction possibilities is a voice control which nowadays can‘t be restricted only to direct commands. The goal of adaptive interaction between man and computer is the human needs understanding. The paper deals with the user's emotional state classification based on the voice track analysis, it describes its own solution - the measurement and the selection process of appropriate voice characteristics using ANOVA analysis and the use of PRAAT software for many voice aspects analysis and for the implementation of own application to classify the user's emotional state from his/her voice. In the paper are presented the results of the created application testing and the possibilities of further expansion and improvement of this solution.

Keywords

• classification
• emorec
• emotion recognition
• praat
• voice analysis