VoIP, Voice over Internet Protocol, is an emerging new technology that involves the routing of voice conversations over the Internet or any other IP – based network. It does not utilize traditional dedicated, circuit – switched voice transmission lines. Voice data flows over a general – purpose packet switched network Problems associated with VoIP includes echoes, jitter, latency, packet loss Echoes are waves that have been reflected at points on the transmission medium with enough amplitude and time difference to be acknowledged as one different from the original wave transmission. There are two forms of echoes on voice networks, Hybrid and Acoustic Echo. Hybrid echo is an electrical signal reflection that occurs at the 4 –wire to 2 – wire conversion point in a PSTN network. It is caused by impedance matching. It enters a VoIP network whenever there is a connection between PSTN and VoIP networks. Acoustic echo on the other hand is non – linear and is usually caused by poor acoustic isolation between the speaker and microphone of a device. It can gain entry into a VoIP network from any source. The major difference between the two types of echoes is that Hybrid echo is a property of the line connection and it remains mostly constant throughout the call while acoustic echo varies in latency depending on the environment of the echo source.
Figure 1 :: Echo in a VoIP Network
It is usually greater in VoIP systems than in public switched telephone network (PSTN) because of the inherent delay in VoIP systems. The human ear has the ability to detect echo waves when the delay from the original wave is equal or close to 10ms. The intensity of the echo wave should also be beat least less than 25 or 30dB of the original wave. PSTN networks do not need echo cancellation on their networks because the echo is not delayed unlike VoIP networks. In an attempt to reduce the effect of echoes on voice calls, low receiver volume was introduced. This led to callers having to shout into the mouthpiece for the user at the other end to repeat what was said. It served as a means of dealing with echoes before digital processing became possible. The voice signals heard by a listener were attenuated by the hardware. The echo goes through the attenuator twice and provided an avenue for echo reduction. It was however not satisfactory and abandoned when digital echo cancellation became possible.
David Mandelstam said that Digital echo cancellation is based on subtracting from the received signal a correction based on the response of the system to a short of sound, called the finite impulse response (FIR). The FIR is simply the echo one would hear from a short ping. Latency is technically defined as the time it takes a packet to travel from its source to its destination. It is also known as delay. This problem is most profound in satellite connection because of the large distances the packets have to travel through. It is a common problem in slow and congested connections. According to , callers notice roundtrip delay of 150ms or more. The human ear is also able to detect close to 250ms latencies. Initial works to curb this problem is to consider the devices used to route packets on the networks. Another solution is to reserve bandwidth for voice packets. This way, routers can know that the packets are in real time. Also, higher priority can also be given to voice packets. Jitter refers to a measure of the variance over time of the latency across a network. The nature of packet – switched networks involves the splitting of information into packets and each of these packets can travel via different paths to its destination. In general, it is usually a problem with slow links or congestion. According to , jitter between the source and destination should be less than 100ms. If jitter is less than this value, it can be solved. The primary solution found for...