In this paper, the Acoustic Vowel Parameters which is based upon the Dialect Classification for Punjabi Speech is provided. The information from the formant’s dynamics F1, F2, and F3 was analyzed and further used in the process. The sound was evaluated with open source software PRAAT for the acoustic assessment. Multilingual speakers having age between twenty to thirty years has been selected for recording from Malwai and Doaba dialects of Punjab. The data of the total 20 people from Doaba region and 20 from Malwai region in which 7 females and 13 males has been taken for each dialect. In the proposed work MATLAB platform is used. First all the training dataset for the Doabi and the Malwai sound files were collected. The total training set has 140 sound files and the testing file has 19 sound files. Various parameters were analyzed in the training process. These parameters are Duration of the Sound file, Pitch, and Formants (F1, F2, and F3). The Formants (F1, F2, and F3) values were analyzed through PRAAT also. The formants are evaluated using the LPC method in MATLAB. The classifier used in the work was LDA and has classified the input sound file as per Doabi and Malwai sound file. The overall accuracy achieved in the system is 94.44%.

Acoustic, Punjabi, Formants, classification, training, testing, LDA (Linear Discriminant Analysis), and MATLAB

[1] C. Themistocleous, “Dialect classification using vowel acoustic parameters”, Elsevier Speech Commun., Vol. 92, Issue.5, pp. 13–22, 2017.
[2] S. Mittal and R. Kaur, “Implementation of Phonetic Level Speech Recognition System for Punjabi Language”, IEEE, 2016.
[3] D. Kaur and G.B. Kumar, “Vowel Analysis for Indian English”, In the Proceedings of the 2016 6th International Conference On Advances In Computing & Communications, ICACC, 6-8 September 2016, Cochin, India, Vol. 93, pp.533–538, 2016.
[4] Y. Liuxin, “The five main vowels acoustic analysis of Tibetan Lhasa dialect”, In the Proceedings of the 2015 Fifth Int. Conf. Instrum. Meas. Comput. Commun. Control, pp.0–4, 2015.
[5] Z. Yue and L. Wei, “Study of the formant and duration in Chinese whispered vowel speech”, Elsevier Applied Acoustics, Vol. 114, pp.240–243, 2016.
[6] H. Liu and J. Liang, “Vowels as acoustic cues for sub-dialect identification in Chinese”, IEEE, Vol. 1, pp.3–7, 2016.
[7] P. Singh “Formant Analysis of Punjabi non-nasalized Vowel Phonemes”, In the Proceedings of the International Conference on Computational Intelligence and Communication Systems, India, pp. 382–387, 2011.
[8] O. Kalasnhnik and J. Fletcher, “An acoustic study of vowel contrasts in North Indian English”,ICphS XVI, No.8, pp.953–956, 2007.
[9] P. Martiand and S. P. V. S. W. Beet, “Analysis of ten vowel sounds across gender and regional/cultural accent”, pp. 2231–2234, 2016.
[10] H. Kaur and R. Bhatia, “Speech Recognition System for Punjabi Language”, International Journal of Advanced Research in Computer Science and Software Engineering, Vol. 5, Issue 8, August 2015.
[11] V. Gupta and G.S. Lehal, “Automatic Keywords Extraction for Punjabi Language”, IJCSI International Journal of Computer Science Issues, Vol. 8, Issue 5, No 3, 2017.
[12] W. Ghai and N. Singh, “Continuous Speech Recognition for Punjabi Language”, International Journal of Computer Applications, Vol. 72, No.14, pp. 0975 – 8887, 2013.