Acoustic Impact of Ventricular Folds on Phonation Studied in Ex Vivo Human Larynx Models

A reliable and objective evaluation of voice quality is of essential importance in the therapy of pathological voices. The cepstral peak prominence (CPP) is calculated in the cepstral domain and is insensitive to variations of the fundamental frequency or pitch-tracking errors. Owing to the inconsistent information about its calculation procedure in the literature, the quantitative comparison of CPP values is difficult. This study provides insight into different definitions for the cepstrum calculation and also parameter variations. Their influence on the resulting CPP is presented on the basis of an acoustic dataset of sustained phonation obtained from an ex vivo human larynx model. Applying the different definitions for cepstrum calculation, the absolute CPP values varied within six orders of magnitude. Variation of the window length for partitioning the acoustic time signal and use of the averaging method over the windows yielded both quantitative and qualitative deviations of the CPP distribution. The results showed that it is essential to provide detailed information about the CPP calculation procedure to permit quantitative comparisons with the literature. After its evaluation, the CPP was used to investigate the acoustic influence of the ventricular folds as a function of the sub-glottal pressure (P-sub). In the present study, the CPP definition of Hillenbrand et al. [1], which has become established in voice research, was applied. In vitro experiments with two excised human larynx models, each in the configuration with and without ventricular folds, were executed. The results were compared with those of available studies with canine larynges. The analysis turns out that the ventricular folds stabilise and enhance the acoustic output at low P-sub ranges at which the ventricular folds do not oscillate. After oscillation onset of the vocal folds, the CPP increases for increasing Psub and reaches its maximum at relatively low Path values. By further increasing Psub the CPP decreases much faster for the configuration with ventricular folds than without. This could be attributed to the movement of the ventricular folds. We suggest that the ventricular folds support and stabilise the vocal fold phonation process as far as they act as stationary boundary conditions. As soon as the glottal flow excites the ventricular folds to move periodically, the process is destabilised resulting in a decrease of voice quality, represented by lower CPP values.