WHAT ARE THE SPECTRAL DIFFERENCES FOR SINGING SOUNDS PRODUCED BY DIFFERENT LARYNGEAL MECHANISMS IN THE SAME PITCH RANGE?

This paper deals with the vocal quality differences observed between two sounds produced at the same pitch and SPL in the two different laryngeal mechanisms M1 and M2. Glottal parameters such as spectral tilt would be relevant to describe these differences but inverse filtering is problematic for high pitch sounds. Instead, the energy distribution over the whole spectra, relatively to the SPL, was studied. Three frequency bands were defined: FB1 = 0-2 kHz, FB2 = 2-4.5 kHz (which corresponds to the singer's formant frequency region) and FB3 = 4.5-8 kHz. One professional countertenor and two amateur tenors recorded their phonetogram in each mechanism. The results show that the energy ratio of FB1 at same SPL and pitch is greater in M2 than in M1, which can be related to the difference in open quotient. The energy ratio of FB2 is smaller in M2 than in M1, where a correlation has been found with the SPL. This difference is smaller for the professional singer who is used to sing in both mechanisms while performing. But the strongest difference between the two mechanisms is observed on the energy ratio of FB3. Differences in FB1 and FB3 might be of laryngeal origin