Indian Vowels based Evaluation of First Fundamental Frequency and its Variant Bandwidths

Authors

  • Praveen. K. B
  • B. Siva Kumar

Abstract

The study presents analysis of Indian English vowels based on fundamental frequency along with the bandwidths of the four fundamental frequencies for an untrained database for the speakers age ranging between 16 to 21 years. An autoregressive modelling is incorporated in addition to Linear Prediction Coding (LPC) for the analysis and estimation of fundamental frequency, bandwidth are calculated for various vowel recording speech samples. These parameters are considered as the tool for the phonetic distinction, speaker unique among the range of individual speakers under consideration. The research concern is to employ and trace the speakers with the ingrained speech parameters in it. The researchers propose is to use this autoregressive model for utterances made by speakers from south India taken on different vowels and SWIPE algorithm for the fundamental frequency estimation. The speech samples are recorded for the neutral condition of the speaker. The frequency components obtained are comparatively plots against the various individual bandwidths along the fundamental frequency of the every speaker which are uttered over an span of time for thirty individual times by the speaker, the mean values are taken into account for the comparative analysis for the investigation of the vowel uniqueness and its variability as a parameter for speech recognition criteria, the entire demonstration is done using MATLAB.   

Downloads

Published

2020-01-12

Issue

Section

Articles