Abstract: Emotion detection through voice is a method that provides information regarding the actual emotion of a person irrespective of the words being used in conversation. Here the emotion is detected through the voice as an input instead of text in conversation. In this paper we focus on various techniques proposed by authors to detect the emotion through userís voice. Estimating the degree of natural emotion through voice corpora by different authorís experimental measures is the main aim of this paper.

Keywords: Acoustic Features, Speech Recognition, Support Vector Machines, Anger.