Method of Extracting Speech Characteristics of Bugis Regional Language

Sriwijanaka Yudi Hartono; Saidah Suyuti

doi:10.31963/elekterika.v22i1.5477

Authors

Sriwijanaka Yudi Hartono Universitas Muslim Indonesia
Saidah Suyuti Universitas Muslim Indonesia

DOI:

https://doi.org/10.31963/elekterika.v22i1.5477

Keywords:

characteristic extraction, short time energy, pitch, formant

Abstract

his research focuses on developing an application to extract speech signal characteristics of Bugis regional speakers using Matlab and Wavesurfer software. The study centers on feature extraction, specifically Short Time Energy (STE), Autocorrelation Cepstral for pitch extraction, and Cepstrum Linear Predictive Coding (LPC) for formant analysis. Results indicate that male voices have greater signal energy than female voices. The pitch values for male speakers range from 52.29 Hz to 135.76 Hz, while female speakers' pitch ranges from 73.39 Hz to 242.42 Hz. The formant analysis showed that male and female speakers have distinct formant frequencies, with male speakers having lower formant frequencies. For instance, the first formant (f1) in male speakers ranges from 433.0 Hz to 590.3 Hz, while in female speakers it ranges from 273.9 Hz to 452.8 Hz. The pitch of male voices is concentrated between 50 Hz and 250 Hz, whereas female voices fall between 120 Hz and 500 Hz. Additionally, differences in formants (f1 to f5) were observed in the word "na'baca," with female speakers generally having higher formant frequencies. This analysis provides valuable insights into the acoustic characteristics of the Bugis regional language. The developed application aims to enhance speech signal processing technologies, supporting applications like speech recognition, linguistic analysis, and voice synthesis, contributing to the preservation of the Bugis language through signal processing techniques.

References

[1]. Hartono , Y.S, Basalamah A. Ekstraksi Pitch Pada Suara Laki-laki Dan Perempuan Dengan Metode Algoritma Autokorelasi Dan AMDF Untuk Penentuan Tipe Jenis Jangkauan Suara Berdasarkan Tipe Suara. Makassar Sulawesi Selatan; 2016.

[2]. Umar R, Riadi I, Hanif A. Analisis Bentuk Pola Suara Menggunakan Ekstraksi Ciri Mel-Frequency Cepstral Coefficients ( MFCC ) Voice Pattern Form Analysis Using Feature Extraction Mel-Frequency Cepstral Coefficients ( MFCC ). 2018;4(2):294–304.

[3]. Wang J, Li Y, Xue Z, Li C. Research on Speech Feature Extraction System in Oral English. 2019;347(Icedem):9–11.

[4]. Mazumder MA, Salam RA. Feature extraction techniques for speech processing: A review. Int J Adv Trends Comput Sci Eng. 2019;8(1.3 S1):285–92.

[5]. Hartono, S.Y.,, Suyuti S. Ekstraksi Pitch Sinyal Wicara ( Speech ) Bahasa Daerah Bugis Dengan Algoritma Autokorelasi. 2024;9(April):125–34.

[6]. Gill AS. A Review on Feature Extraction Techniques for Speech Processing. 2016;5(10):18552–6.

[7]. R A, K L, Shivakumar DM. Feature Extraction of Speech Signal using LPC. Ijarcce. 2016;5(12):189–90.

[8]. Hartayu R, Anam C, Aziz DA, Studi P, Elektro T, Teknik F, et al. Simulasi Ekstraksi Fitur Suara menggunakan Mel-Frequency Cepstrum Coefficient. 2022;8:80–7.

[9]. Athira G, Anjusha VK. Analysis of Feature Extraction Methods for Speech Recognition. 2017;4(4):328–33.

[10]. GIANNAKOPOULOS T. Introduction to AUDIO ANALYSIS: A MATLAB Approac. First Edit. Amsterdam, New York: Elsevier; 2014. 33–38 hal.

[11]. Thakuria L, Pran P, Talukdar H. Analysis and Detection of Emotional states of a person by extracting Pitch and Formants with respect to Assamese Language. 2014;5(2):890–3.

[12]. Pitch Estimation using Autocorrelation Method Pitch Estimation using Autocorrelation Method and. 2016;(March).

[13]. Caesarendra, W., Ariyanto M. Panduan Belajar Mandiri MATLAB. Jakarta: PT.Elex media Kompetindo; 2016. page 37-64.