Comparative Analysis of Pitch and Formant for Recognizing Emotions of Isolated Marathi Speech

Recognizing emotions from speech is a tuff task as we are not aware of the features which will accurately classify the emotions. This paper is an approach to show which speech feature classifies the emotions more accurately. The features compared here are Pitch and Formant while the classifier used is Linear Discriminant Analysis (LDA). The database used in this experiment was developed using 50 male and 50 female Marathi speaking native speakers. The emotions used here are Neutral, Happy, Sad, Surprise and Boredom. At the end of the experiment it was observed that formant recognized the emotions very efficiently and accurately with respect to that of energy