carriers are carrying synchronization and control signals to the receivers, but most of carriers conveys signal-data to the receiver.

Transcription

1 Hierarchical encoding an interesting specialty within DVB-T and DVB-H. INTRODUCTION Highly promoted by the German expert group in the DVB-T group, hierarchical encoding represents a mode, which is compromise between an request for portable or even mobile reception and the need for maintaining a relative high capacity in order to serve a stationary reception. Hierarchical encoding puts unfortunately some constraints to the DVB- T system. This article will try to dig into the advantages and the compromises one has to face when selecting the hierarchical mode. The DVB-T, C-OFDM system Before going into the details about hierarchical encoding. It could be useful to update the reader on the DVB-T system. The DVB-T system uses a C-OFDM modulation to carry the data the receiver. Before transmission the data signal, MPEG 2 transport stream is subjected to two types of error protection systems. The first is the and second is a Viterby/Trellis encoding. The is a shorten version of the original proposal, it convert 188 bytes data packets into 204 bytes packets. The broadcaster has the choice to select between various protection rates or depths of the Viterby/Trellis encoding. carriers are carrying synchronization and control signals to the receivers, but most of carriers conveys signal-data to the receiver. The modulated signal is generated from the digital information, carriers + modulation data, by synthesizing the envelope of the total signal in amplitude and phase. This process is referred to as the Inverse Fast Fourier Transformation. Each of the data conveying carriers are modulated by either QPSK, 16 QAM or 64 QAM. The size of the IFFT processing is 2 k for the 1705 carrier system and 8k for the 6817 carrier system. In the receiver the signal is processed by an FFT system in order to regenerate the digital information from the carriers. Below on fig. 2 is shown from one set of carriers (a symbol) to another set the number of carriers carrying digital information and those reserved for synchronization (Pilots). Fig. 2 Pilot insertion pattern The C-OFDM system makes use of a great amount of carriers. The UK system is using 1705 numbers and the continental part of Europe 6817 carriers. Some of the

2 1/T Symbols T Guard Interval, Tg N= 6817 Frequency The 8-k system offers four possible guard intervals, T g = 28 µs, 56 µs, 112 µs and 224 µs. The 2 k has also 4 possible values, but only ¼ of the 8 k in µs). Time Fig. 3 The OFDM signal in the 8 k version An essential point, when it comes to transmission, is the guard interval. The guard interval is a part system specification and is illustrated in fig. 4. The C-OFDM signal is built up during the time T-T g. At t = t o +T-T g all the carriers has reached their final rest state equal to one of the points in the constellation diagram. An example of a 64 constellation diagram is shown at fig. 5. From t o +T-T g to t o +T the constellation is kept constant in way frozen. This is called the guard interval, T g. For further protection this constellation information is transmitted just after the previous transmitted symbol. Thanks to the heavy interleaving used in conjunction with the of data in relation to the carrires. The information is so widely spread over the numbers of carrier that even a notch in the middle of the C-OFDM block of carrier will often not cause loss of transmission thanks to the error correction system The C-OFDM system also allows a number of transmitters on the same frequency. The transmitter can actually support each other in an SFN operation. (SFN is an abbreviation for Single Frequency Networking). stream Modulation IFFT (2k or 8k) and guard interval IF Precor rector and D/A converter, RF converter RF Fig. 4. Block diagram of DVB-T exciter using the PROTV modulator, PT 5780 The guard interval allows the receiver system to absorb reflections from e.g. buildings or even multipath reception as inside a building. Length of guard interval is a compromise between bit rate and protection against multipath reflections. The system allows the transmitters to operate on the same frequencies. All the transmitters in the SFN network must be synchronized to the same reference and they must output the same signal (same bits) to exactly same time.

3 The time and frequency reference can easily be taken care of by means of a GPS receiver system and by including an SFN adapter in the network center. The SFN adapter inserts an information packet in the transport stream. The packet called MIP (Mega Initialization Packet) carries all the timing information in the data stream to the transmitters. Fig. 4 illustrates the function of a modern high quality DVB-T modulator. The transport stream receiver s operates in the MFN mode (MFN: Multi Frequency Network), the transport stream operates as a Stream Adapter. The TSA will either remove stuffing bytes from the transport stream or insert stuffing bytes in order to adapt the bit rate coming from the multiplexer at the program provider to actual bit rate required by the selected COFDM mode. In an SFN application the detects and reads the information in the MIP and sets the modes and the delays accordingly. Fig.5 Typically Constellation display of a received signal achieve a stable reception by using the 64 QAM mode. The experience from Singapore, where 8 transmitters in an SFN configuration (SFN: Single Frequency Networking) serves the public busses, shows that 16 QAM and QPSK mode is the options available for such a service. The drawback of going for QPSK or 16 QAM is the loss in capacity over the 64 QAM. QPSK 2 bit / carrier 16-QAM 4 bit / carrier Fig. 6 illustrates the difference in capacity of the modulation modes used in DVB-T 64-QAM 6-bit / carrier To compensate or perhaps better explained to combine mobile services together with a fixed reception application the DVB-T group introduced the Hierarchical encoding. This encoding calls for a special DVB-T modulator. Such modulators are irst now commercially available. One of the problems has been the definition of performance in an SFN operation. One of the products, which is able to fully comply the DVB-T hierarchical mode is the modulator from PROTV Technologies in Denmark. This modulator can be ordered with a hierarchical option. The hierarchical encoding. The 64 QAM received signal shown in fig. 4 is measured in a receiver under ideal reception conditions. When it comes to non-ideal conditions as portable receiver or mobile reception, it can be very difficult to Fig. 7 The PROTV modulator, PT 5780

4 Fig. 8 illustrates the idea of hierarchical encoding. There are defined three possible constellation offsets α=1, 2 and 4, but the principle is the same. Two transport streams are fed to the modulator, one of the transport stream controls in which quadrant the constellation is placed (High Priority Stream) and the other defines the 16 QAM signal (Low Priority Stream) Limitation of the hierarchical encoding Based upon the experience from Singapore and also from published information in Germany and UK. There a few draw backs of the hierarchical encoding. One, is that it supports only one robust program, as the HP transport Stream is carried as QPSK modulation. In Singapore it was decided to split the transmitter network into two independent networks. One network serves the mobile application and it can carry up to 3 programs. Of speed reasons it is using 2k, the other parameters are guard interval ¼ and code rate ½. Fig. 8 Hierarchical constellation diagram. The benefit of the hierarchical encoding is that the system allows one program for mobile or portable application together with service supporting typically 3 programs to be pick up in stationary application. The second network is using 64 QAM modulation and the other parameters are 8k IFFT, Code rate 2/3 and guard interval 1/16. Hierarchical encoding will also reduce the decoder margin for the LP transport stream, typically hierarchical encoded signal will require 2 db better C/N ratio than for a non-hierarchical signal. SFN operation represents some difficulties as the LP and HP channel might come from different program providers and via different networks. The PROTV modulator is able to cope with these obstacles. Data is read from each of the MIP (LP and HP). The modulation parameters as guard interval, modulation parameters and time to transmit are read from the HP MIP. Only code rates are read from the LP and HP MIP.

5 Stream HP Hierarchical Multiplexer Modulation IFFT (2k or 8k) and guard interval IF Precor rector RF and D/A converter, converter RF Stream LP Fig.9 The block diagram of the hierarchical DVB-T modulator The capacity of the LP channel is also limited by the guard interval used for the HP channel, as the guard interval must be the same for both channels. Code rate for the two channels can differ. Information about some typical bit-rates might assist the broadcaster to make his choice. The most popular mode for stationary application is IFFT 8k, Guard interval 1/16, 64 QAM and code rate 2/3, for the 8 MHz Bandwidth system this mode supports a bit-rate of Mbit/sec. The mode used for the buses in Singapore is IFFT 2k, Guard interval 1/4, 16 QAM and code rate ½: This mode supports only 9.95 Mbit/sec. A hierarchical mode for mobile and stationary could be HP: IFFT 2k, Guard interval ¼, code rate ½, this equals 4.98 Mbit/sec. The LP channel will normally be 16 QAM and code-rate 2/3, which supports Mbit/sec. In total the hierarchical encoded channel support in total a bit-rate of Mbit/sec, but only one mobile TV program. In principle 8k would be a better choice do to the longer guard interval. However, the receiver s presently available will loose decoder margin at a speed of 80 km/h at 600 MHz. Sometime in during 2002 at least two supplies have declared that they will launch chip sets, which can operate at a higher speed also in 8k mode. Conclusion Hierarchical mode represents an interesting solution for a broadcaster, who wants to transmit on program in very robust mode and can accept a minor loss in decoder margin for his other three programs. One application for the hierarchical mode is to serve mobile and portable applications. The Australian expert group has formulated another. In Australia the are considering transmitting the standard resolution program as HP transport stream and use the LP transport stream for the high definition information or the high definition program. The latter will depend of the principle for the commercial receivers for High Definition. Anten