Effects of the Voice over Internet Protocol on perturbation analysis of normal and pathological phonation.

OBJECTIVE In this study, a Voice over Internet Protocol (VoIP) communication based on G.729 protocol was simulated to determine the effects of this system on acoustic perturbation parameters of normal and pathological voice signals. PATIENTS AND METHODS Fifty recordings of normal voice and 48 recordings of pathological voice affected by laryngeal paralysis were transmitted through a VoIP communication system. The acoustic analysis programs of CSpeech and MDVP were used to determine the percent jitter and percent shimmer from the voice samples before and after VoIP transmission. The effects of three frequently used audio compression protocols (MP3, WMA, and FLAC) on the perturbation measures were also studied. RESULTS It was found that VoIP transmission disrupts the waveform and increases the percent jitter and percent shimmer of voice samples. However, after VoIP transmission, significant discrimination between normal and pathological voices affected by laryngeal paralysis was still possible. It was found that the lossless compression method FLAC does not exert any influence on the perturbation measures. The lossy compression methods MP3 and WMA increase percent jitter and percent shimmer values. CONCLUSION This study validates the feasibility of these transmission and compression protocols in developing remote voice signal data collection and assessment systems.