It is important to understand the reason behind degradation of quality of service in real time applications in order to justify handoff latency budget of less than 50ms usually mentioned in the literature.

The The International Telecommunication Union (ITU) recommends a mouth to ear latency of 200ms maximum. It is also widely accepted that when the station is using wireless medium, the handoff latency should be kept under 50ms to preserve good QoS. We will attempt to understand how these values were calculated and how they should be interpreted.

We will decompose the 200ms budget and understand the processes that contribute to the overall delay. These processes include codec operations, transmission over the Internet, buffering at the receiver side, etc…

Some of these procedures take predictable amount of time e.g. buffering at the receiver side and codec operations. The transmission of packets over the Internet on the other hand introduces variable latency. The purpose of these series is to understand how much absolute jitter is acceptable in VoIP systems.

When a communication carried between two or more parties is taking place, if the mouth to ear delay exceeds a certain amount. The conversation becomes distorted; the communicating parties may interrupt each other and start to talk at the same time.

The International Telecommunication Union (ITU) have published a recommendation (ITU G.114) in which analytical evaluation of the effect of mouth to ear latency on human conversations were carried.

According to these results, a delay more than 200ms is considered not optimal. In order to ensure the best quality for real communications, a delay of less than 200ms from mouth to ear must be maintained.

The 200ms budget is divided amongst the different procedures that take place in order to transfer voice over the interconnected equipment. In Voice over IP, basically, the voice is encoded then transmitted over the IP network to the destination, then it is decoded and played back to the user. Each phase involved in the voice transmission process consumes time that must be subtracted from the 200ms budget.

In the following articles, we will see how to estimate the latency contributed by each process involved in the VoIP communication.

The remaining of this series is organized as follows : In the next article, we will see how to estimate latency introduced by coded operations. The third article will focus on de-jitter buffers. Finally in the fourth article wee will see how to estimate the maximum latency budget attributable to wireless handoffs, which is the main goal behind this analysis. Stay tuned.

• Codec latency in VoIP communications
• De-Jittering buffers and real time network services
• Wireless handoff delay and its impact on VoIP performance and QoS
• What is the handoff delay budget in VoIP over wireless links

Labels: wireless, performance