Abstract
This paper describes the relationships between the speaker's intended emotion and the prosodic features of the emotional Japanese speech. The purpose of this study is to clarify how much the speaker's intention is correctly conveyed to the listener, and what are the prosodic-feature parameters that can convey the speaker's emotion definitely. Listening tests are conducted using speech samples that consist of “neutral” speech as well speech with three types of emotions (“anger”, “joy”, and “sadness”) of three degrees (“light”, “medium”, and “strong”). We use 120-word speech samples uttered by 4 announcers, and 144-word speech samples uttered by 4 radio actors/actresses. Results are summarized as follows. The agreement rate of the listener's receptivity with the speaker's intention for speech uttered by radio actors/actresses is higher than that for speech uttered by announcers. There are significant differences between prosodic-feature parameters of emotional speech uttered by announcers and those uttered by radio actors/actresses.
