Automatic Detection of Expressiveness in Oral Reading

Kamini Sabu, Kanhaiya Kumar, Preeti Rao


We present a Computer-Aided Language Learning (CALL) system that assesses a child's oral reading skill including the prosodic aspects. With children who have otherwise achieved word decoding automaticity, prosodic fluency is a reliable predictor of comprehension. Prosody includes attributes such as pace, phrasing and expression. Based on the acoustic correlates of prosodic events, we propose and test features that discriminate expressive speech from monotonous speech and further detect whether the expression is meaningful or simply a rhythmic cadence with no relation to the underlying syntax or semantics of the text. Finally the system based on processing short samples of recorded oral reading and providing feedback on the goodness of both lexical and prosodic aspects is described.


Cite as: Sabu, K., Kumar, K., Rao, P. (2018) Automatic Detection of Expressiveness in Oral Reading. Proc. Interspeech 2018, 1489-1490.


@inproceedings{Sabu2018,
  author={Kamini Sabu and Kanhaiya Kumar and Preeti Rao},
  title={Automatic Detection of Expressiveness in Oral Reading},
  year=2018,
  booktitle={Proc. Interspeech 2018},
  pages={1489--1490}
}