ISCA Archive SIGUL 2023
ISCA Archive SIGUL 2023

Towards Automatic Marking of Pepeha: a Formulaic Māori Language Speech

Catherine I. Watson, Piata Allen, Peter J. Keegan, Keoni Mahelona, Peter-Lucas Jones

This study looks at the assessment of the pronunciation of te reo Māori (Māori language) within a short formulaic speech, known as a pepeha. It is a comparison between marks awarded to the pehepa by trained markers, and scores awarded to the pepeha by Arero, a speech recognition platform which is purpose-build to assess te reo Māori. Pepeha recordings of 304 people were analysed. The study found that there were many similarities between the two assessment methods and they were correlated, albeit weakly. It is argued that the results suggest automatic marking of pepeha is feasible. The next step is to understand acceptable phonetic variation in the pepeha pronunciation via phonetic analysis of a large number of the pepeha recordings.


doi: 10.21437/SIGUL.2023-27

Cite as: Watson, C.I., Allen, P., Keegan, P.J., Mahelona, K., Jones, P.-L. (2023) Towards Automatic Marking of Pepeha: a Formulaic Māori Language Speech. Proc. 2nd Annual Meeting of the ELRA/ISCA SIG on Under-resourced Languages (SIGUL 2023), 124-128, doi: 10.21437/SIGUL.2023-27

@inproceedings{watson23_sigul,
  author={Catherine I. Watson and Piata Allen and Peter J. Keegan and Keoni Mahelona and Peter-Lucas Jones},
  title={{Towards Automatic Marking of Pepeha: a Formulaic Māori Language Speech}},
  year=2023,
  booktitle={Proc. 2nd Annual Meeting of the ELRA/ISCA SIG on Under-resourced Languages (SIGUL 2023)},
  pages={124--128},
  doi={10.21437/SIGUL.2023-27}
}