Digital modulation in SNG
Without going into heavy-duty theory, you need to appreciate that there are different types of modulation (we have already gone over the basics of this) and in digital modulation the predominant modulation is QPSK – a type of phase modulation. Other types of phase modulation are 8-PSK and 16QAM, but these are not commonly used in DSNG.
With QPSK, the carrier undergoes four changes in phase and can thus represent 4 bits of data, and every 2 bits in a QPSK modulated signal make up a ‘symbol’. We talk of a digital modulated signal in symbols per second, or mega (million) symbols per second (MSps).
We have talked about encoding in terms of compression and MPEG-2, but there is also an element of digital encoding within the digital modulation process.
This is because bits are added in the modulation of a digital signal to correct for errors in transmission – termed ‘error correction’.
The signal is inevitably degraded in transmission due to the effects of noise and interference, and to compensate for this inevitable consequence, ‘check’ bits are added to the bit stream in the modulator to enable errors to be detected at the downlink.
Error correction
Although error correction coding reduces the transmission power required, as the error correction can tolerate higher levels of unwanted signal (noise), the demand for bandwidth increases because of the greater overall amount of data being transmitted.
There are two types of error correction added to the digital stream in its transmission to correct for errors – Reed–Solomon and forward error correction.
In Reed–Solomon code, each block of data of programme signal has an additional check block of data added to compensate for any errors that the signal may suffer on its passage from the transmitter to the receiver. In a data stream used in a DSNG uplink, the Reed–Solomon code is defined as (204,188).
The second error correction process is forward error correction (FEC), where bits are added in a predetermined pattern. This is decoded at the receiver using a decoding process called Viterbi to detect any loss of information bits, and attempt to reconstruct the missing ones.
The number of bits added by this process defines the FEC ratio for the signal, and is typically 3/4 – in other words, for each 3 bits, 1 extra bit has been added. Some of the other FEC rates for QPSK that can be used are 1/2, 2/3 and 5/6 but the standard DSNG rate is generally 3/4.
The greater the degree of FEC applied, the more rugged the signal will become, but the occupied bandwidth will need to increase to cope with the error correction overhead. Occupied bandwidth is – as the name suggests – the amount of bandwidth that the signal occupies.
Transmission and symbol rates
The overall data rate, including RS, FEC and the information rate is termed the Transmission Rate. So now we can calculate the actual transmitted symbol rate – this is an important defining parameter for a digital signal.
If we assume that a typical DSNG signal is an 8 Mbps information rate signal, then this is an actual bit-rate of 8.448 Mbps – a standard data rate in the digital ‘hierarchy’. The data stream is QPSK modulated, so the information symbol rate is 4.224 MSps (2 bits per symbol), and the calculations produce a transmitted symbol rate of 6.1113 MSps.
Therefore, the 8 Mbps DSNG signal is expressed as an 8.448 Mbps signal transmitted with QPSK modulation at 3/4 FEC rate, and 204,188 Reed–Solomon coding, giving a modulated symbol rate (Modulation Rate) of 6.1113 MSps.
Sometimes the transmitted symbol rate is expressed without RS coding, which in this case is 5.632 MSps.
The RF bandwidth required is approximately the Modulation Rate multiplied by a factor – typically 1.35 – to give the occupied bandwidth. This signal will fit within a 9 MHz channel – it will actually occupy just over 8 MHz, but there is a guard band allowed, minimizing any interference with signals in adjacent channels on the satellite.
The 9 MHz channel has become a nominal standard for DSNG signals in most satellite operators’ systems.
Quality and compression
Video compression is widespread in broadcasting, and a very large proportion of SNG operations are now in the digital domain. The effect of compression on picture content varies widely depending on that content, and there is still no widely accepted scientific method of measuring picture quality – the human eye has been found to be the best guide, but it is a subjective evaluation.
However, suffice to say that for most news organizations, the advantages of the cost of utilizing compression far outweigh what are considered esoteric concerns over issues of quality.
An inevitable result of compression, whether video or audio, is delay. There are a number of computations carried out in both the compression and decompression processes, and the higher order compression processes (such as MPEG-2) take longer to be completed than the more simpler compression systems. However, as we have seen, there are ways to minimize the on-screen effects of this.