ABSTRACT
The word ‘telecommunication’ is derived from two Greek words, which can be translated as ‘the passage of information at a distance’. Historically the only means of communication available were visual communications, e.g. beacons, smoke signals and semaphore signalling. Modern telecommunications networks enable different types of data to be passed: Voice (by telephony), Paper images (by facsimile), Computer systems (by using digital data networks), Plain text (by telex), Formatted Text (by electronic mail and related systems) and Video Systems (by Integrated Services Digital Network – ISDN). The different forms of telecommunication are distinguished in two ways: firstly by the type of device used (telephone, fax machine, telex machine, etc.) and secondly by the kind of network used to interconnect the devices (telephone network, telex network, data network, etc.) For example, telephones, fax and modems would use the Public Switched Telephone Network (PSTN) and telex machines would use a dedicated telex network. Digital data would have to use a digital network like X.25, Frame Relay, Switched Multimegabit Data Service (SMDS) or Asynchronous Transfer Mode (ATM), to name a few. All PTTs (Post, Telephone and Telegraph) and all other Licensed Operators in the UK operate a digital network. The bandwidth of PSTN is determined by various speech components. The speech input from a PSTN telephone consists of a sum of a number of different frequencies components, from about 50Hz to about 7 kHz depending on who is speaking. The speech output from a PSTN telephone also consists of a number of frequency components, but this time only from 300 to 3400Hz. Thus the telephone network cuts out all frequency components below 300Hz and above 3400Hz. Hence the bandwidth of PSTN is 3.1 kHz (3.4 kHz to 300Hz). Voice being transmitted through the PSTN is called 3.1 k or ‘commercial speech’. Music frequencies can be in the range of 30 kHz (Double Bass) to 20 kHz (Flute). The telephone service is a result of this compromise. Transferring all speech components would result in a high quality but expensive service; progressively reducing the bandwidth lowers both the quality and cost. There are several different methods to achieve this. Two main
methods used in telecommunications are Pulse Code Modulation (PCM; 64kbit/s) and Adaptive Differential Pulse Code Modula-
tion (ADPCM; 16-64 kbit/s). The basic coding technique used is PCM, but this coding technique is expensive on networks because each digitally encoded channel will require 64,000 bit/s of bandwidth. ADPCM is a predictive coding which converts the analogue signal to PCM first. Then this coding scheme will predict a new value with the reference to the previous sample point and it will look at the actual point and work out the difference between the two points: if the change is positive then the 3-bit code is prefixed with a 1. If the change is negative then the 3-bit code is prefixed with a 0. Thus the speed of ADPCM is calculated as follows: 8000 samples per second multiplied by 4 bits per sample ¼ 32,000 bit/s. ADPCM will use 32 kbit/s for voice compression. Note: Recommendation G.721 ADPCM, which will only support 32 kbit/s, is now obsolete and is replaced by G.726, which supports 5-, 4-, 3-and 2-bit sampling for speeds of 40, 32, 24 and 16kbit/s respectively.
