Investigation of the Relation between Emotional State and Acoustic Parameters in the Context of Language

Abstract

Acoustic analysis is the most basic method used for speech emotion recognition. Speech records are digitized by signal processingmethods, and various acoustic features of speech are obtained by acoustic analysis methods. The relationship between acoustic featuresand emotion has been investigated in many studies. However, studies have mostly focused on emotion recognition success or the effectsof emotions on acoustic features. The effect of spoken language on speech emotion recognition has been investigated in a limitednumber. The purpose of this study is to investigate the variability of the relationship between acoustic features and emotions accordingto the spoken language. For this purpose, three emotions (anger, fear and neutral) of three different spoken languages (English, Germanand Italian) were used. In these data sets, the change in acoustic features according to spoken language was investigated statistically.According to the results obtained, the effect of anger on the acoustic features does not change according to the spoken language. Forfear, change in spoken language shows a high similarity in Italian and German, but low similarity in English.