Spectral Analysis of Speech Quality in VoIP for G.729A and AMR-WB Speech Coders

Abstract

The transmission of Voice over Internet protocol requires signal processing for reliable & efficient performance over the computer and the communication platform. Speech quality, as perceived by the users of VoIP telephony, is critically important to the update of the service. Signal quality is degraded by network layer problems including delay, jitter and packet loss. The work presented in this paper presents the spectral analysis of speech quality for narrowband and wideband VoIP. The simulation model of VoIP network under impairments is presented. The VoIP simulations are conducted for G.729A and AMR-WB speech coders at different packet loss rates. The comparison results of various speech analysis algorithms such as power density spectrum FFT etc. for both narrowband and wideband VoIP signals are presented in this paper. The results are also measured using PESQ for narrowband speech signal and WB-PESQ for wideband speech signal