Abstract
Variation in the radiated sound spectrum, p(rad), is an expressive technique that can be achieved on the saxophone by employing different vocal tract configurations that change the vocal tract impedance spectrum Z(mouth). However, the relation between p(rad) and the vocal tract impedance spectrum Z(mouth) has not previously been measured for orchestral instruments. In this study, p(rad) and Z(mouth) were measured simultaneously over the frequency range from 100 to 10000 Hz while saxophonists played. For notes sounded over the normal and altissimo playing range, experienced saxophonists are able to produce distinctive variations in the spectral envelope of the radiated sound, without changing the pitch or the amplitude of the fundamental, using different vocal tract configurations. When Z(mouth) was adjusted to have magnitudes comparable with the input impedance of the bore, Z(bore), harmonics of p(rad) were usually increased at nearby frequencies, both for the range over which the saxophone has strong resonances (100-2000 Hz) and for the higher range (2-10 kHz). Less experienced players who are unable to produce strong peaks in Z(mouth) produce much smaller variations in p(rad). p(rad) correlates more strongly with the series impedance Z(mouth) + Z(bore): for large values of \Z(mouth)\, larger series impedance at a particular frequency always produced larger radiated power. The change in sound inside the mouth is proportionally larger than that in p(rad), which explains why players judge the timbre to be more changed than do listeners.