VoIP Voice and Fax Signal Processing

20

VoIP VOICE QUALITY

For several decades, telephone users have been experiencing public switched telephone network (PSTN) based voice communication quality. Subjectively, this experience is used as the main reference for comparing the voice quality from other voice communication systems. In PSTN-based systems, voice has good intelligibility, acceptable speaker identification, naturalness, and only minor disturbing impairments. The PSTN uses G.711 μ-law and A-law as compression. Compressed bits are sent on synchronous times-division multiplexed (TDM) interfaces. In VoIP voice communication, a similar G.711 is also used end to end as one of the compression codecs. In VoIP, instead of TDM bits, groups of compressed bytes are sent as a packet on the Internet Protocol (IP) network. The VoIP quality closely matches with PSTN quality under the correct conditions of end-to-end G.711 IP packet transmission. VoIP will also use many high-compression codecs like G.729AB, G.723.1A, and iLBC, which causes VoIP voice quality to be lower than PSTN quality even in best end-to-end packet delivery networks. The goal here is how to reach and finally to exceed PSTN quality through VoIP.

VoIP will have to take care of additional impairments to achieve the quality comparable with the PSTN quality. The four main voice impairment contributors popularly referred for VoIP telephony according to TIA-810A [TIA/EIA-810A (2000)] are delay, echo, voice compression, and packet loss. In a recent upgrade of TIA ...