Abstract
This work attempts to build an automatic 1D-CNN based tone recognizer of Mizo, an under-studied Tibeto-Burman language of North-East India. Preliminary research findings have confirmed that along with four canonical tones of Mizo (High, Low, Rising and Falling), a phenomenon of Rising tone sandhi (RTS) with distinct phonetic characteristics are also observed. As per the authors’ knowledge, no work has been reported to identify the RTS along with four distinct tones. Moreover, previous tone recognition works have explored hand-crafted features derived from F0 contour which may not provide the explicit representation of a specific tone category. To address these issues, current work attempts to incorporate the RTS along with four lexical tones and learn tone specific features directly from F0 contours using a 1D-CNN model. Experimental results conducted for speaker independent case show that the proposed 1D-CNN model achieves an accuracy of 68.18%.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Boersma, P.: PRAAT, a system for doing phonetics by computer. Glot Int. 5(9/10), 341–345 (2001)
Chang, P.C., Sun, S.W., Chen, S.H.: Mandarin tone recognition by multi-layer perceptron. In: International Conference on Acoustics, Speech, and Signal Processing, pp. 517–520. IEEE (1990)
Chen, C., Bunescu, R., Xu, L., Liu, C.: Tone classification in mandarin Chinese using convolutional neural networks. In: Interspeech 2016, pp. 2150–2154 (2016). https://doi.org/10.21437/Interspeech.2016-528. http://dx.doi.org/10.21437/Interspeech.2016-528
Chen, X.X., Cai, C.N., Guo, P., Sun, Y.: A Hidden Markov Model applied to Chinese four-tone recognition. In: ICASSP 1987. IEEE International Conference on Acoustics, Speech, and Signal Processing, vol. 12, pp. 797–800. IEEE (1987)
Chhangte, L.: Mizo Syntax. Ph.D. thesis, University of Oregon (1993)
Chollet, F., et al.: Keras link url. https://keras.io/
Dey, A., et al.: Robust mizo continuous speech recognition. In: INTERSPEECH, pp. 1036–1040 (2018)
Fanai, L.: Some aspects of the lexical phonology of mizo and english: an autosegmental approach. Ph.D. thesis, CIEFL, Hyderabad, India (1992)
Gogoi, P., Dey, A., Lalhminghlui, W., Sarmah, P., Prasanna, S.R.M.: Lexical tone recognition in mizo using acoustic-prosodic features. In: Proceedings of The 12th Language Resources and Evaluation Conference, pp. 6460–6463. European Language Resources Association, Marseille, France, May 2020
Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press, Cambridge (2016)
Govind, D., Sarmah, P., Prasanna, S.R.M.: Role of pitch slope and duration in synthesized Mizo tones. In: Speech Prosody (2012)
Kaur, J., Singh, A., Kadyan, V.: Automatic speech recognition system for tonal languages: state-of-the-art survey. Arch. Comput. Methods Eng. 28(3), 1039–1068 (2020). https://doi.org/10.1007/s11831-020-09414-4
Kothapalli, V., et al.: Robust recognition of tone specified mizo digits using CNN-LSTM and nonlinear spectral resolution. In: 2018 15th IEEE India Council International Conference (INDICON), pp. 1–5. IEEE (2018)
Lalhminghlui, W., Mazumdar, P., Sarmah, P.: Behaviour of tone sandhi in different morphological structures in Mizo. In: Paper presented at The 23rd Himalayan Languages Symposium, Tezpur University, Tezpur (2017)
Lalhminghlui, W., Sarmah, P.: Production and perception of rising tone sandhi in Mizo. In: Proceedings TAL, pp. 114–118 (2018)
Lee, T., Ching, P.C., Chan, Y.H., Mak, B.: Tone recognition of isolated Cantonese syllables. IEEE Trans. Speech Audio Process. 3(3), 204–209 (1995)
Lee, T., Lau, W., Wong, Y.W., Ching, P.: Using tone information in Cantonese continuous speech recognition. ACM Trans. Asian Lang. Inf. Process. (TALIP) 1(1), 83–102 (2002)
Li, X., et al.: Mandarin Chinese tone recognition with an artificial neural network. J. Otol. 1(1), 30–34 (2006)
Liu, L.C., Yang, W.J., Wang, H.C., Chang, Y.C.: Tone recognition of polysyllabic words in Mandarin speech. Comput. Speech Lang. 3(3), 253–264 (1989)
Peng, G., Wang, W.S.Y.: Tone recognition of continuous Cantonese speech based on support vector machines. Speech Commun. 45(1), 49–62 (2005)
Gao, Q., Sun, S., Yang, Y.: ToneNet: a CNN model of tone classification of mandarin Chinese. In: Interspeech 2019 (2019)
Rose, P.J.: Considerations on the normalization of the fundamental frequency of linguistic tone. Speech Commun. 10(3), 229–247 (1991)
Ryant, N., Slaney, M., Liberman, M., Shriberg, E., Yuan, J.: Highly accurate mandarin tone classification in the absence of pitch information. In: Proceedings of Speech Prosody (2014)
Ryant, N., Yuan, J., Liberman, M.: Mandarin tone classification without pitch tracking. In: 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (2014)
Sarma, B.D., Dey, A., Lalhminghlui, W., Gogoi, P., Sarmah, P., Prasanna, S.: Robust Mizo digit recognition using data augmentation and tonal information. In: Proceedings 9th International Conference on Speech Prosody, vol. 2018, pp. 621–625 (2018)
Sarma, B.D., Sarmah, P., Lalhminghlui, W., Prasanna, S.R.M.: Detection of Mizo tones. In: Interspeech, pp. 934–937 (2015)
Sarmah, P., Dihingia, L., Lalhminghlui, W.: Contextual variation of tones in Mizo. In: Interspeech, pp. 983–986 (2015)
Sarmah, P., Wiltshire, C.R.: A preliminary acoustic study of Mizo vowels and tones. J. Acoust. Soc. India 37(3), 121–129 (2010)
Wang, W.S., Li, K.P.: Tone 3 in Pekinese. J. Speech Lang. Hear. Res. 10(3), 629–636 (1967)
Wang, X.D., Hirose, K., Zhang, J.S., Minematsu, N.: Tone recognition of continuous mandarin speech based on tone nucleus model and neural network. IEICE Trans. Inf. Syst. E91.D(6), 1748–1755 (2008)
Weidert, A.: Componential Analysis of Lushai Phonology, vol. 2. John Benjamins Publishing (1975)
Yang, W.J., Lee, J.C., Chang, Y.C., Wang, H.C.: Hidden Markov model for Mandarin lexical tone recognition. IEEE Trans. Acoust. Speech Signal Process. 36(7), 988–992 (1988)
Acknowledgements
The speech corpus used in this work was developed for the project titled “Acoustic and Tonal Features based Analysis of Mizo”, funded by the Ministry of Electronics & Information Technology (MeitY), Ministry of Communication & Information Technology (MC&IT), Government of India.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Gogoi, P., Kalita, S., Lalhminghlui, W., Sarmah, P., Prasanna, S.R.M. (2021). Learning Mizo Tones from F0 Contours Using 1D-CNN. In: Karpov, A., Potapova, R. (eds) Speech and Computer. SPECOM 2021. Lecture Notes in Computer Science(), vol 12997. Springer, Cham. https://doi.org/10.1007/978-3-030-87802-3_20
Download citation
DOI: https://doi.org/10.1007/978-3-030-87802-3_20
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-87801-6
Online ISBN: 978-3-030-87802-3
eBook Packages: Computer ScienceComputer Science (R0)