Automatic Tone Recognition of Ao Language

Parismita Gogoi, Moakala Tzudir, Priyankoo Sarmah, S. R. M. Prasanna

Ao is an under-resourced Tibeto-Burman tonal language spoken in the North-Eastern state of India, Nagaland. Preliminary research findings have confirmed that Ao has three lexical tones, namely, High, Mid and Low. There are three distinct dialects of Ao, namely, Chungli, Mongsen and Changki, differing in tone assignment in lexical words. In this work, tone distributions in trisyllabic words are considered with 4320 iterations consisting of 4176 high, 5473 mid and 3311 low tones, collected from 36 speakers, for the three dialects of Ao. An attempt is made to automatically recognise the phonological tones in Ao using SVM with zero-frequency filtering (ZFF) derived F0 profile as the preliminary feature.

 DOI: 10.21437/SpeechProsody.2020-205

Cite as: Gogoi, P., Tzudir, M., Sarmah, P., Prasanna, S.R.M. (2020) Automatic Tone Recognition of Ao Language. Proc. 10th International Conference on Speech Prosody 2020, 1005-1008, DOI: 10.21437/SpeechProsody.2020-205.

  author={Parismita Gogoi and Moakala Tzudir and Priyankoo Sarmah and S. R. M. Prasanna},
  title={{Automatic Tone Recognition of Ao Language}},
  booktitle={Proc. 10th International Conference on Speech Prosody 2020},