The Effect of Sex on the Turkish Voice Spectrum

<strong>Objective:</strong> The Long-Term Average Spectrum (LTAS) of voice is a functional method used to discriminate various voice qualities. In this study, the voice quality characteristics of Turkish female and male speakers were examined and compared through voice spectra created with LTAS. <strong>Material and Methods: </strong>30 male and 30 female speakers with typical voices, aged 18 to 50, were required to read a text at a habitual pitch and loudness. The recorded speech samples were used to create voice spectra containing 75 bands with a bandwidth of 160 Hz and covering the range from 0 to 12 kHz. The spectral tilt (ST) of the voice spectra was also compared between the sexes. <strong>Results: </strong>Multiple comparisons revealed that there were significant intensity differences between sexes at the band regions of 0–160 Hz, 160–320 Hz, 640–960 Hz, 2880–3040, 4000–5120 Hz where females had higher intensities than males except for the band of 0–160 Hz. STs of females were found to be significantly lower than those of males. <strong>Conclusion:</strong> All significant band differences between the sexes were below 6 kHz, and both sexes had a rather flat-spectrum above 6 kHz. Females had significantly higher intensities around the region of 2880–3040 Hz, because of aspiration noise caused by the possible posterior glottal gap during phonation. Females had shallower STs than males, indicating that females had greater hyperfunctional voice quality. The voice spectrum characteristics of Turkish speakers were compared with those of other studies and found more similarities with Spanish than Persian.