Acoustic Characteristics of Vocal Emotions Simulated by Actors

Abstract

This paper reports on a set of acoustic measures of vocal emotions simulated by a skilled actor and actress who read 12 short sentences. The emotions simulated were Neutral, Cold Anger, Hot Anger, Happiness, Sadness, Interest, and Elation. Acoustic measures were examined for digitised samples of the 168 sentences (2 speakers × 12 sentences × 7 emotions). These measures included the over-all mean energy, standard deviations of energy, mean fundamental frequency, mean values of the first six formant frequencies and their corresponding frequency bandwidths, mean durations of utterance, and mean articulation rate. Results displayed both similarities and differences in the acoustic measures of the two actors' simulations of the different emotions. These are presented and discussed.