Voice Signals Produced With Jitter Through a Stochastic One-mass Mechanical Model

Area of Science:

• Acoustics and phonetics
• Biomedical engineering
• Voice science

Background:

• Jitter, a perturbation in vocal fold cycle length, arises from quasiperiodic vocal fold oscillations.
• Jitter is characterized by random deviations in glottal cycle lengths and is often expressed as a percentage of the glottal pulse duration.
• Elevated jitter levels can signify pathological voice conditions such as vocal fold nodules or paralysis.

Purpose of the Study:

• To develop a stochastic model for vocal fold jitter.
• To utilize a one-mass mechanical model of vocal folds with assumed right-left symmetry and horizontal motion.
• To investigate the relationship between vocal fold stiffness as a stochastic process and jitter generation.

Main Methods:

• A stochastic process was employed to model the stiffness of each vocal fold.
• A one-mass mechanical model of the vocal folds was constructed, assuming complete symmetry and horizontal motion.
• The model's ability to produce jitter was assessed.

Main Results:

• Probability density functions of fundamental frequency were generated for varying jitter levels.
• Synthesized voice samples were created to demonstrate the effects of different jitter levels.
• The model's output was compared against established voice analysis software (Praat).

Conclusions:

• The proposed stochastic model successfully generates vocal fold jitter.
• The model provides a framework for understanding jitter and its relation to voice pathologies.
• Validation using Praat software confirmed the accuracy of jitter measurements in synthesized voice signals.