Smoothing factor and voice perturbation measurements. Objective: Although gradual frequency and amplitude variations originating from intonation or vocal instability are not related to voice quality, jitter and shimmer are sensitive to such variations. These parameters are therefore calculated by subtracting the average of a group of successive periods from the middle period; the period number that is averaged is called the smoothing factor. The aim of this study was to investigate the effect of smoothing factor size on frequency and amplitude perturbation measurements.
Methodology: A prospective study was designed, incorporating recorded voice samples from ten male and ten female healthy adult volunteers; samples were analyzed with the Multi Dimensional Voice Program for Multi-Speech. We used all low smoothing factor sizes up to 21, and then skipped smoothing factor sizes in steps of ten, resulting in 28 different levels.
Results: Smoothed pitch perturbation quotient and smoothed amplitude perturbation quotient values increased as the smoothing factor size increased. Smoothing factor size was correlated with smoothed pitch perturbation quotient and smoothed amplitude perturbation quotient values in males (r = 0.589 and r = 0.698, respectively) and females (r = 0.736 and r = 0.847, respectively).
Conclusion: Our study revealed an exact relationship between smoothing factor values and perturbation measures. At low smoothing factor values, perturbation measures are sensitive to short-term variations, whereas at high smoothing factor values, they are less sensitive to short-term but more sensitive to long-term variations.