Comparative Spectral Analysis of Mongolian Sounds of Same Midlle Age Male and Female

This paper presents the results of spectral analysis of all vowels and consonants of Mongolian language using non-parametric and parametric methods of analysis. Early time for spectral analysis have been used the Period gram and it’s versions, Fast Fourier Transform, MUSIC and PMUSIC methods, but today except them are used the automated speech signal processing software as PRAAT and AUDACITY. Spectral analysis of all sounds of Mongolian language was performed by these methods. Detailed analysis was performed for separately pronounced all sounds and pronounced at the beginning, middle and end of the words. The spectral analysis parameters as the formant frequencies, bandwidth of the formants,？ formant power densities of the vowels and consonants were analyzed. Done comparative evaluation depending on position of sounds in words (at beginning, middle and end of word) and for male or female. Results of comparative spectral analysis show the followings. Firstly, should be noted differences of formant frequencies of same age male and female. Secondly, spectral parameters (formant frequencies, bandwidth of the formants,？ formant power densities) of separately pronounced sounds vary according to their positions in the words. It’s not observed deterministic or rational changes. First formant frequencies of female pronunciations？ predominantly？ exceed that of the male. But in next formants happened predominance of male formant frequencies depending on sound and position.？ Formant frequencies tend to increase at the beginning, middle, and end of the word