Several studies from this laboratory have shown that when listeners are given a set of six emotions to choose from, they rate semantically neutral sentences as sounding angry more often when spoken in a clear speaking style (clear speech) than when the talker was instructed to speak in a conversational manner (conversational speech). The present study expanded this line of research to a new corpus of conversational and clear speech in which talkers heard their own voice through headphones in four different simulated speaking environments: quiet, two levels of white noise (i.e., “Lombard” speech), and reverberation. In an exploratory perceptual experiment, listeners with normal hearing heard 10 instances of a single semantically neutral sentence (“I looked up the word in the dictionary” with a variety of /bVd/ keywords) spoken by two male and two female talkers in four conditions: conversational speech in quiet, conversational speech in 63 dB SPL of white noise, clear speech in quiet, and clear speech in 63 dB SPL of white noise. Data analyses will determine whether, similar to clear speech, Lombard's speech is perceived as sounding angry more often than conversational speech, and whether the effects on perceived emotion of speaking style and speaking environment interact.
Robison et al. (Tue,) studied this question.