Actors vocally portrayed happiness, sadness, anger, fear, and disgust with weak and strong emotion intensity while reading brief verbal phrases aloud. The portrayals were recorded and analyzed according to 20 acoustic cues. Listeners decoded each portrayal by using forced-choice or quantitative ratings. The results showed that (a) portrayals with strong emotion intensity yielded higher decoding accuracy than portrayals with weak intensity, (b) listeners were able to decode the intensity of portrayals, (c) portrayals of the same emotion with different intensity yielded different patterns of acoustic cues, and (d) certain acoustic cues (e.g., fundamental frequency, high-frequency energy) were highly predictive of listeners' ratings of emotion intensity. It is argued that lack of control for emotion intensity may account for some of the inconsistencies in cue utilization reported in the literature.
Juslin et al. (Mon,) studied this question.