Modulation of Musical Sound Clips for Robot’s Dynamic Emotional Expression

Sound is an important medium for human-robot interaction. Single sound or music clip is not enough to express delicate emotions, especially it is almost impossible to represent emotional changings. This paper tries to express different emotional levels of sounds and their transitions. In this paper, happiness, sadness, anger, and surprise are considered as a basic set of robots’ emotion. By using previous proposed nominal sound clips of the four emotions, this paper proposes a method to reproduce the different emotional levels of sounds by modulating their musical parameters ‘tempo,’ ‘pitch,’ and ‘volume.’ Basic experiments whether human subject can discern three different emotional intensity levels of the four emotions are carried out. By comparing the recognition rate, the proposed modulation works fairly well and at least shows possibility of letting humans identify three intensity levels of emotions. Since the modulation can be done by dynamically changing the three musical parameters of sound clip, our method can be expanded to dynamical changing of emotional sounds.