27.2. The influence of subglottal pressure changes on the modelled EGG waveform.

The effects of subglottal pressure changes were tested within the simulation of the three vowels /a:/, /e:/ and /u:/ as described by the vocal tract configuration in Table 17. During the simulation all other parameters are kept constant. The subglottal pressure varies between 2 and 10 cm H2O in 2 cm H2O steps, which corresponds to soft to moderate voice levels.

The effect of subglottal pressure on fundamental frequency (see Fig.48) is about 5 Hz/cm H₂O, which is twice as much as reported by Ishizaka and Flanagan (1972:1257). When comparing both results it should be noted that most of the measurement points of this experiment lie before the saturation area (ibid.) and, thus, the slope of the curve can be steeper than in the Flanagan and Ishizaka results.

Figure 48. The dependency of the fundamental frequency on the subglottal pressure for [a:], [e:] and [u:] . The subglottal pressure changes from 2 to 10 cm H2O in 2 cm H2O steps. The other modelling parameters remain unchanged

The Open Quotient computed for the glottal airflow is presented in Fig. 49a. The computation was done in the same way as for the EGG waveform by using the first method described in section 12, thus the values are lower than those of Ishizaka and Flanagan (1972:1259). The general tendency, however, remains unchanged - the OQ decreases as the subglottal pressure increases, but the change is relatively small.

Figure 49. The dependency of the Open Quotient on the subglottal pressure for [a:], [e:] and [u:] for a) the simulated glottal airflow b) the simulated EGG waveform. The subglottal pressure changes from 2 to 10 cm H2O in 2 cm H2O steps. The other modelling parameters remain unchanged.

The results of the EGG simulation exhibit the opposite dependency - the duty ratio (ie.e the Open Quotient) grows with the increased subglottal pressure (Fig.49b). This tendency is most obvious for /a/ vowels, for the other vowels the growth is almost saturated at the higher pressures. There is a simple explanation for this effect - at lower pressures, there is partial contact between folds and the growing contact area and the shrinking opening area overlap, causing a shortening of the no-contact phase. As the subglottal pressure increases, the motion of the folds becomes faster and the phase difference between the margins decreases, resulting in sharper changes in the waveform at the instants of closure and opening. Thus, the duty ratio grows unexpectedly. This surprising result is, however, in line with the results of the word stress experiment, where a small but significant increase in the EGG duty ratio was found for stressed vowels.

Additionally, it was expected, that the skewness of the glottal airflow pulse (reflected in Fig. 50a in the inverse of the Speed Quotient) changes with the increase in subglottal pressure. The airflow pulses are more skewed at higher pressures, which validates the results of previous experiments (Sluijter, 1996; Hanson, 1995; Claßen et al., 1996). The results of the EGG simulation are not as obvious. As follows from Fig. 50b the Speed Quotient grows, which means that the EGG pulses are more symmetrical at higher subglottal pressures. As anticipated, this results in an effect which is contrary to the airflow pulse dependency. However, for very low pressures (Fig.50b, pressures lower than 5 cm H₂O) the SQ of the EGG waveform is much higher than for higher pressures, which suggests a limited usability of the Speed Quotient in the EGG domain.

Figure 50. The dependency of the Speed Quotient on the subglottal pressure for [a:], [e:] and [u:] a) inverted Speed Quotient of the simulated glottal airflow b) Speed Quotient of the simulated EGG waveform. The subglottal pressure changes from 2 to 10 cm H2O in 2 cm H2O steps. The other modelling parameters remain unchanged.

The main goal of the experiment was to relate the steepness of the EGG slopes to the subglottal pressure. The results presented in Fig. 51 (a) for the rising and b) for the falling slope of the EGG waveform) are fully in line with the outcome of the word stress experiment. The steepness of the modelled slopes increases with the increased subglottal pressure for all modelled vowels, especially in a lower pressure range. The average increase amount to about 25 amplitude units per 1 cm H₂O for the rising contact slope and -23/cm H₂O for the falling contact slope (measured with a linear regression model for all vowels, the ANOVA measure of the regression model is F(1,13)=16.597, r2=0.74 and F(1,13)=247, r2=0.95 for contact rise and fall, respectively).

Figure 51. The dependency of EGG slope steepness on the subglottal pressure for [a:], [e:] and [u:] a) the steepness of the rising slope b) the steepness of the falling slope. The subglottal pressure changes from 2 to 10 cm H2O in 2 cm H2O steps. The other modelling parameters remain unchanged.

As was expected, the increase in subglottal pressure causes an increase in the glottal flow velocity Ug. However, the peak-to-peak amplitude of the EGG waveform remains unchanged. There are two classes of EGG timing parameters: those which do not depend on the Ps alterations (e.g. the duration of both closing and both opening phases) and those which change in accordance with the already mentioned parameters (e.g. the full-contact phase shortens with the increase in subglottal pressure).