Speech Emotion Analysis in the Romanian Language

In the communication in the international negotiation, the people must adopt certain ways to deliver the message in order to support and enforce in a most effective strategies and tactics adopted by the negotiating team. The analysis of the emotion expressiveness in the voice, setting new theoretical methodologies and algorithms to identify and classify the emotions are based on several practical and scientific applications in the inter-disciplinary field which touches: computers sciences, psychology and cognitive sciences. Several scientists are focused on the improvement of the performances of classification and emotional recognition methods, in order to find new optimum parameters to solve the constraints (classification errors, complexity of the algorithms etc.) met in several applications. The aim of this paper consists of developing and improving the analysis for the emotional expressiveness of the vocal signals in Romanian language. We use the only emotional corpus annotated for the Romanian language, which is part of the Romanian Language Sounds project (SRoL) which contains more than 300 folders with emotional stales (i.e., 900 speeches) as well as some folders with standard tonalities. The studied emotions are: happiness, sadness, fury and neutral tone. This inter-disciplinary theme has been chosen in order to contribute to the development of several principles, methodologies and possible concepts concerning the emotional classification and recognition.