Emotions in Hindi speech- analysis, perception and recognition

Human Speech conveys speaker´s emotional state along with linguistic intelligence. Meaning of a speech sample changes when it is uttered with different emotions. The present paper gives a description of different types of studies conducted to analyze, perceive and recognize commonly occurring emotions in Hindi speech. These have been classified as anger, happiness, fear, sadness, surprise in addition to neutral. Intonation, intensity and duration patterns changes due to changes in sentence types as well as due to changes in emotions. A relationship among the measured acoustic parameters and the patterns has been used to classify them. Experiments have been conducted to study and recognise emotions based on phonetic as well as prosodic parameters in the speech samples due to changes in emotions. These parameters include MFCC & their derivatives and prosodic parameters as the F0, A0 and Duration. In one of the experiment vowel segments taken from continuously spoken sentences and in another experiment Hindi digits were used as speech samples for machine recognition of emotions using the Neural Net classifiers. Human perception experiments have been conducted at all levels of experiments and compared the results with machine recognition performance. In most cases it has been found that machine recognition was found to be better compared to human performance. Both Phonetic as well as prosodic parameters play role in identification of emotions.