An Acoustic Study of Emotional Speech Produced by Italian Learners of Japanese

Motoko Ueyama, Xinyue Li


Pioneering research on L2 emotional speech provides evidence for crosslinguistic similarities and differences. However, limited research has been conducted at the production level. This study examines three types of emotional speech—both L1 and L2 of Italian learners of Japanese, with L1 Japanese as the baseline. Our goal is to find the basic acoustic patterns of emotional speech in this L2-type, and further observe whether there is a transfer effect. We conducted an acoustic analysis of single-word utterances with five emotion types (neutral, happy, angry, sad and surprised) for six acoustic parameters (F0mean, F0max, F0min, F0range, intensity and utterance duration). The results showed crosslinguistic similarities (e.g., higher and greater pitch range for happy and surprised). We also realized a grouping of the patterns of the learners’ L1 and L2. The two speech types showed a very similar mean distribution across the emotion types for F0mean, F0max and F0min, which differed distinctively from the distribution of L1 Japanese. Additionally, they were both lower in overall pitch level for all emotion types; higher in F0mean and F0max for anger and sad with respect to neutral, which were contrary to L1 Japanese patterns. These findings indicate towards the effects of L1 transfer.


 DOI: 10.21437/SpeechProsody.2020-8

Cite as: Ueyama, M., Li, X. (2020) An Acoustic Study of Emotional Speech Produced by Italian Learners of Japanese. Proc. 10th International Conference on Speech Prosody 2020, 36-40, DOI: 10.21437/SpeechProsody.2020-8.


@inproceedings{Ueyama2020,
  author={Motoko Ueyama and Xinyue Li},
  title={{An Acoustic Study of Emotional Speech Produced  by Italian Learners of Japanese}},
  year=2020,
  booktitle={Proc. 10th International Conference on Speech Prosody 2020},
  pages={36--40},
  doi={10.21437/SpeechProsody.2020-8},
  url={http://dx.doi.org/10.21437/SpeechProsody.2020-8}
}