Traditional voice communication between distant participants has been realized via long-distance speech for years. This mutual acoustic communication from person to person over greater distances, is the most widely used form of telecommunications. For this form of communication, there is the circuit-switched telephone network, through which voice connections can be established between two and more partners.
Voice communication requires transmission with as little delay as possible, so that the continuous data stream for the voice arrives at the receiver in its original form. In this way, the intelligibility of the transmitted speech is still ensured even if the transmission is impaired by dropouts, jitter, delay times and interfering influences such as crackling or noise.
The evaluation of speech quality can be objective and subjective. A well-known subjective evaluation is the Mean Opinion Score( MOS). Inconjunction with the E model standardized by the International Telecommunication Union( ITU) under G.107, this can also be determined in simulations. Another ITU standard for evaluating voice quality is G.114, which also takes the packet loss rate into account.
The most important quality criteria for the transmission of voice information are delay times, bit error rates, echoes and jitter. Since the ear is sensitive to sound fluctuations and speech interruptions, delay times should be approximately constant. Speech quality is not affected by the delay during transmission; only the call quality deteriorates. Bit errors, on the other hand, have the effect of cracking noises. A bit error rate of `10^-2` makes a strongly disturbing pattering noticeable, at `10^-3` the pattering dissolves into a dense sequence of pops, which at a bit error rate of `10^-4` changes into single pops. At `10^-5` only single pops are audible and bit error rates of `10^-6` do not affect speech transmission at all.