Design of a voice quality assessment system based on multidimensional parameters

The goal of this work is to develop an ‘assessing’ rather than ‘measuring’ system of voice quality based on multidimensional clinical parameters. The system which can run fluently on a PC with Windows, consists of the signal acquisition module, the signal processing module, the results analyzing and exporting module, and video capture module. By analyzing both the speech and electroglottograph signals, parameters of fundamental frequency, amplitude, energy, perturbation and vocal fold vibration etc. are extracted to establish 5 statistical standard models for assessing the voice quality: the hoarse voice, the harsh voice, the breathy voice, the regulation of vocal folds vibration and the time of glottal closure. In the meantime, the system supports the function of capturing the laryngeal endoscope images basically. It brings more convenience to otolaryngologists on improving diagnosis efficiency.