NICAM: NICAM stands for Near Instantaneously Companded Audio Multiplex. The analogue audio signal is digitised and is compressed when in digital format. The digital information can then be transmitted in a multiplex format so that mono channels or a stereo channel can be transmitted in the one frame. At the receiving end, the signal is then digitally expanded and converted back to analogue. There are two basic reasons for companding the audio. The first is the reduction of bits required, the second reason is the improved signal to noise ratio in the received audio. The analogue audio is sampled at a rate of 32 KHz. This is just over twice the maximum audio frequency of 15 KHz. There are 14 bits in the initial audio sample. This 14 bit binary word is digitally compressed to 10 bits. The NICAM system was developed by the BBC in the seventies. The UK version is often referred to as NICAM 728. The number 728 refers to the transmitted bit rate of 728 Kilobits per second. One 728 bit frame is transmitted every millisecond. The breakdown of the bit rate is shown below. 8 bits Frame Alignment Word 8 Kbit/s 5 bits Control Information 5 Kbit/s 11 bits Additional Data 11 Kbit/s 704 bits Sound, Parity or Data 704 Kbit/s 728 Kbit/s The first step in the compression procedure is to separate the digital samples into blocks of thirty two samples each. The second step is to code the 14 bit binary samples using a 10 bit 2s Complement code to an accuracy defined by the magnitude of the largest sample word in the block. The 2s Complement code is an alternative way of representing a number using ones and zeroes. One bit in the number represents the whether the number is positive or negative. This is its sign. The rest of the bits represent the magnitude of the number. This type of representation is called " Signed Number Representation". In this system, positive numbers are represented as simple unsigned binary. There is no change. For a negative number, the situation is different. The binary number to be converted is inverted on a bit by bit basis. A 1 will become a 0 and vice versa. Then 1 is added to it. The most significant bit, MSB, of a negative 2s complement number will be 1. The following example will make it clearer. Example: To express -7 in 2s Complement, invert 0111,(+7 = 0111). This gives 1000. Add 1, this gives the final value of 1001. Therefore in 2s Complement code -7=1001. In the third step a Scale Factor is then generated. The Scale Factor tells the receiver which degree of compression is used. The scale factor word is three bits wide. The Coding Range is defined by the Scale Factor. There are five coding ranges. These are relative to a maximum amplitude of 1. Range 1: 1 to 0.5 Range 2: 0.5 to 0.25 Range 3: 0.25 to 0.125 Range 4: 0.125 to 0.0625 Range 5: 0.0625 to 0 As can be seen from the diagram, if the block of samples are in coding range 1, the four least significant bits are discarded. For range 2, the three least significant bits are discarded as is the second most significant bit. In each case, the number of bits in the sample is always 10. In this way, the 14 bit samples are first coded into 2s Complement and then into a 10 bit compressed sample. A further level of information is provided for the receiver by the protection range. This is also related to the Scale Factor. The relationship is shown in the table below. Coding Ranges Protection Ranges Scale Factor r1 r2 r3 1 1 1 1 1 2 2 1 1 0 3 3 1 0 1 4 4 0 1 1 5 5 1 0 0 5 6 0 1 0 5 7 0 0 1 5 7 0 0 0 A parity bit is added to each sample. This is produced by checking the six most significant bits. The sound and parity bits in the NICAM frame only comprise 704 bits of the 728. The other bits are taken up by the Frame Alignment Word, 8 bits; the Control Information, 5 bits; Additional Data, 11 bits. The Frame Alignment word is an eight bit word, 01001110, that is transmitted at the start of every frame. Its function is to synchronise the receiver with the datastream. The control information is transmitted as a five bit block. The first bit, c0 is the frame flag bit. It is high for the first eight frames and low for the next eight. The next three bits, c1, c2, c3 identify the application of the contents of the sound block. The correct term for them is the application control bits. Bit c3 is used to indicate the need for further processing by the receiver. It is this bit that would be used to switch in the decryption circuitry in the receiver. If the bit is high and the receiver does not have the necessary decryption circuitry, then the audio output will be switched off. The fifth bit c4 is the reserve audio switching flag. This bit is high when the FM subcarrier is carrying the same channel as the NICAM. Application Control Bits Table Data Contents Of Sound Blocks c1 c2 c3 0 0 0 Stereo signal. Alternate samples. 0 1 0 Two Mono channels, M1 and M2, alternate frames 1 0 0 One Mono audio and one data channel, alternate frames. 1 1 0 One 704 bit data channel There are eleven additional data bits. The function of these bits has not been defined. It would be easy to use them as a service identifier or for transmitting other data. The modulation format used to transmit NICAM, terrestrially, is Differentially Encoded Quadrature Phase Shift Keying or DQPSK for short. This is a rather elegant modulation system in that it reduces the bandwidth required to transmit the data. Each phase change represents a bit pair or two bits of data. The following diagram shows the rest state of the carrier phase. The rest states are ninety degrees apart. The carrier phase remains in one of the rest states until a bit pair causes it to change phase by the predetermined amount. The phase change caused by each bit pair is shown in the table. Carrier Rest States Bit Pair - Phase Change Table 1 Bit Pair Phase Change A B In Degrees | 4 --------- 2 0 0 0 | 0 1 -90 1 0 -270 3 1 1 -180 With the carrier in rest state 1, a bit pair of 10 will cause a phase change of -270 Degrees. This would put the carrier phase in rest state 4. Applying a further bit pair of 11, would cause a phase shift by - 180 Degrees to rest state 2. Applying a further bit pair of 01 would cause a phase shift of -90 Degrees to rest state 3. Strangely, the negative shift is clockwise. The modulation format is clear and unambiguous. A bit pair can always be recovered by comparing the present phase of the carrier with the immediately previous phase. Before the 728 bit frame is converted to bit pairs, the datastream is scrambled for spectrum shaping purposes. This is carried out to ensure that the data looks like noise and so causes minimum interference to the video or other audio carriers. A pseudo random sequence generator is EXORed with the data stream. The PRSG is a nine stage type. The initialisation word is 111111111. The Frame Alignment Word is not scrambled. The first bit that is scrambled is the bit immediately after the Frame Alignment Word and the last bit that is scrambled is the bit immediately before the FAW. The descrambling of the frame must be carried out at the receiver before the multiplex is split up. Since the information is in digital format, it is easy to encrypt. The simplest method is to encrypt the Frame Alignment Word, FAW. This would ensure that a NICAM descrambler that didn't have a valid FAW couldn't lock up the frame. The NICAM digital signal is EXORed with a PRSG bitstream for noise reduction purposes. This PRSG has a fixed seed. If a variable seed is used, a simple though efficient encryption system could be effected. The keys for the encryption system could be held in the descramblers or transmitted in the unassigned bits in the NICAM signal. The NICAM format could easily be adapted to a satellite television application. The first and most obvious modification to the standard is the carrier frequency. The carrier frequency for NICAM 728 as used in the UK is 6.552 MHz. This frequency is obtained by multiplying the transmitted bit rate of 728 Kilobits per second by 9. By multiplying the bit rate by 10, a carrier frequency of 7.28 MHz would be obtained. The bandwidth of the NICAM carrier on the UK terrestrial system is 700 KHz. The beauty of this modification is that only necessary modifications to a normal NICAM demodulator are the carrier crystal frequency and the input bandpass filter.