1.1.1 Features of the system design for the markets to be served by the Digital Radio Mondiale (DRM) system
The DRM system, is a flexible digital sound broadcasting (DSB) system for use in the terrestrial broadcasting bands below 30 MHz. (Recommendation ITU-R BS.1514)
It is important to recognize that the consumer radio receiver of the near future will need to be capable of decoding any or all of several terrestrial transmissions; that is, narrowband digital (for 30 MHz RF), wider band digital (for 30 MHz RF), and analogue for the LF, MF, HF bands and the VHF/FM band. The DRM system will be an important component within the receiver. It is unlikely that a consumer radio receiver designed to receive terrestrial transmissions with a digital capability would exclude the analogue capability.
In the consumer radio receiver, the DRM system will provide the capability to receive digital radio (sound, program related data, other data, and still pictures) in all the broadcasting bands below 30 MHz. It can function in an independent manner, but, as stated above, will more likely be part of a more comprehensive receiver – much like the majority of today’s receivers that include AM and FM band analogue reception capability.
The DRM system is designed to be used in either 9 or 10 kHz channels or multiples of these channel bandwidths. Differences in detail on how much of the available bit stream for these channels is used for audio, for error protection and correction, and for data depend on the allocated band (LF, MF, or HF) and on the intended use (for example, ground wave, short distance sky wave or long distance sky wave). In other words, there are modal tradeoffs available so that the system can match the diverse needs of broadcasters worldwide. As indicated in the next section, when regulatory procedures are in place to use channels of greater bandwidth than 9/10 kHz, the DRM system’s audio quality and total bit stream capability can be greatly improved.
The DRM system employs advanced audio coding (AAC), supplemented by spectral band replication (SBR) as its main digital encoding. SBR improves perceived audio quality by a technique of higher baseband frequency enhancement using information from the lower frequencies as cues. OFDM/QAM is used for the channel coding and modulation, along with time interleaving and forward error correction (FEC) using multilevel coding (MLC) based on a convolutional code. Pilot reference symbols are used to derive channel equalization information at the receiver. The combination of these techniques results in higher quality sound with more robust reception within the intended coverage area when compared with that of currently used AM.
The system performs well under severe propagation conditions, such as those encountered under long distance multipath HF skywave propagation, as well as under easier to cope with MF groundwave propagation. In the latter case, maximum use is made of the AAC and SBR source coding algorithms, leading to much higher quality audio than that achieved by AM, since a minimal amount of error correction has to be employed. For many HF propagation conditions, the necessity to achieve a high degree of robustness reduces the audio quality compared to MF digital; nevertheless, the audio quality is still better than current AM quality.
The design permits the use of the DRM system within a single frequency network (SFN).
It also provides the capability for automatic frequency switching, which is of particular value for broadcasters who send the same signals at different transmission frequencies. For example, this is done routinely by large HF broadcasting organizations using AM to increase the probability of at least one good signal in the intended reception area. The DRM system can enable a suitable receiver to select the best frequency for a programme automatically without any effort on the part of the listener.
Figure 1 describes the general flow of the different classes of information (audio, data, etc.) from encoding on the left of the Figure to a DRM system transmitter exciter on the right. Although a receiver diagram is not included as a figure, it would represent the inverse of this diagram.
On the left are two classes of input information:
– the encoded audio and data that are combined in the main service multiplexer;
– information channels that bypass the multiplexer that are known as fast access channel (FAC) and service description channel (SDC)
The audio source encoder and the data pre-coders ensure the adaptation of the input streams onto an appropriate digital format. Their output may comprise two parts requiring two different levels of protection within the subsequent channel encoder.
The multiplex combines the protection levels of all data and audio services.
The energy dispersal provides a deterministic, selective complementing of bits in order to reduce the possibility that systematic patterns result in unwanted regularity in the transmitted signal.
The channel encoder adds redundant information as a means for error correction and defines the mapping of the digital encoded information into QAM cells. The system has the capability, if a broadcaster desires, to convey two categories of “bits”, with one category more heavily protected than the other.
Cell interleaving spreads consecutive QAM cells onto a sequence of cells, quasirandomly separated in time and frequency, in order to provide an additional element of robustness in the transmission of the audio in timefrequency dispersive channels.
The pilot generator injects information that permits a receiver to derive channel equalization information, thereby allowing for coherent demodulation of the signal.
The OFDM cell mapper collects the different classes of cells and places them on a timefrequency grid.
The OFDM signal generator transforms each ensemble of cells with the same time index to a time domain representation of the signal, containing a plurality of carriers. The complete timedomain OFDM symbol is then obtained from this time domain representation by inserting a guard interval –a cyclic repetition of a portion of the signal.
The modulator converts the digital representation of the OFDM signal into the analogue signal that will be transmitted via a transmitter/antenna over the air. This operation involves frequency upconversion, digitaltoanalogue conversion, and filtering so that the emitted signal complies with ITUR spectral requirements.
With a nonlinear highpowered transmitter, the signal is first split into its amplitude and phase components (this can advantageously be done in the digital domain), and then recombined (by the action of the transmitter itself) prior to final emission.
126.96.36.199 Audio source coding
The source coding options available for the DRM system are depicted in Fig. 2. All of these options, with the exception of the one at the top of the figure (AAC stereo), are designed to be used within the current
9/10 kHz channels for sound broadcasting below 30 MHz. The CELP option provides relatively low bitrate speech encoding and the AAC option employs a subset of standardized MPEG4 for low bit rates (that is, up to 48 kbit/s). These options can be enhanced by a bandwidthenhancement tool, such as the SBR depicted in the figure. Representative output bit rates are noted in the figure. All of this is selectable by the broadcaster.
Special care is taken so that the encoded audio can be compressed into audio superframes of constant time length (400 ms). Multiplexing and unequal error protection (UEP) of audio/speech services is effected by means of the multiplex and channel coding components.
As an example of the structure, consider the path in Fig. 2 of AAC mono plus SBR. For this, there are the following properties:
Frame length: 40 ms
AAC sampling rate: 24 kHz
SBR sampling rate: 48 kHz
AAC frequency range: 06.0 kHz
SBR frequency range: 6.015.2 kHz
SBR average bit rate: 2 kbit/s per channel.
In this case, there is a basic audio signal 6 kHz wide, which provides audio quality better than standard AM, plus the enhancement using the SBR technique that extends this to 15.2 kHz. All of this consumes approximately 22 kbit/s. The bitstream per frame contains a fraction of highly protected AAC and SBR data of fixed size, plus the majority of AAC and SBR data, less protected, of variable size. The fixedtimelength audio superframe of 400 ms is composed of several of these frames.
As noted in Fig. 1, the DRM system total multiplex consists of three channels: the MSC, the FAC and the SDC. The MSC contains the services, audio and data. The FAC provides information on the signal bandwidth and other such parameters and is also used to allow service selection information for fast scanning. The SDC gives information to a receiver on how to decode the MSC, how to find alternate sources of the same data, and gives attributes to the services within the multiplex.
The MSC multiplex may contain up to four services, any one of which can be audio or data. The gross bit rate of the MSC is dependent upon the channel bandwidth and transmission mode being used. In all cases, it is divided into 400 ms frames.
The FAC’s structure is also built around a 400 ms frame. The channel parameters are included in every FAC frame. The service parameters are carried in successive FAC frames, one service per frame. The names of the FAC channel parameters are: base/enhancement flag, identity, spectrum occupancy, interleaver depth flag, modulation mode, number of services, reconfiguration index, and reserved for future use. These use a total of 20 bits. The service parameters within the FAC are: service identifier, short identifier, CA (conditional access) indication, language, audio/data flag, and reserved for future use. These use a total of 44 bits. (Details on these parameters, including field size, are given in the system specification.)
The SDC’s frame periodicity is 1 200 ms. Without detailing the use for each of the many elements within the SDC’s fields, the names of them are: multiplex description, label, conditional access, frequency information, frequency schedule information, application information, announcement support and switching, coverage region identification, time and date information, audio information, FAC copy information, and linkage data. As well as conveying this data, the fact that the SDC is inserted periodically into the waveform is exploited to enable seamless switching between alternate frequencies.
The coding/modulation scheme used is a variety of coded orthogonal FDM (COFDM) which combines OFDM with MLC based on convolutional coding. These two main components are supplemented by cell interleaving and the provision of pilot cells for instantaneous channel estimation, which together mitigate the effects of shortterm fading, whether selective or flat.
Taken together, this combination provides excellent transmission and signal protection possibilities in the narrow 9/10 kHz channels in the longwave, mediumwave and shortwave broadcasting frequency bands. And it can also be effectively used at these broadcasting frequencies for wider channel bandwidths in the event that these are permitted from a regulatory standpoint in the future.
For OFDM, the transmitted signal is composed of a succession of symbols, each including a guard interval – a cyclic prefix which provides robustness against delay spread. Orthogonality refers to the fact that, in the case of the design of the DRM system, each symbol contains approximately 200 subcarriers spaced across the 9/10 kHz in such a way that their signals do not interfere with each other (are orthogonal). The precise number of subcarriers, and other parameter considerations, are a function of the mode used: ground wave, sky wave, and highly robust transmissions.
QAM is used for the modulation that is impressed upon each of the various subcarriers to convey the information. Two primary QAM constellations are used: 64QAM and 16QAM. A QPSK mode is also incorporated for highly robust signalling (but not for the MSC).
The interleaver time span for HF transmission is in the range of 2.4 s to cope with time and frequencyselective fading. Owing to less difficult propagation conditions, a shortened interleaver with 0.8 s time span can be applied for LF and MF frequencies.
The multilevel convolutional coding scheme will use code rates in the range between 0.5 and 0.8, with the lower rate being associated with the difficult HF propagation conditions.
1.1.3 Transmitter considerations
The DRM system exciter can be used to impress signals on both linear and nonlinear transmitters. It is expected that highpowered nonlinear transmitters will be the normal way of serving the broadcasters. This is similar to current practice which exists for doublesideband amplitude modulation.
Because of this need, over the past few years, using the DRM system and other prototypes, effort has been spent to determine how these nonlinear transmitters can be used with narrowband digital signals. The results have been encouraging, as can be seen from recent DRM system field tests.
Briefly, the incoming signal to a Class C (nonlinear amplification) transmitter needs to be split into its amplitude and phase components prior to final amplification. The former is passed via the anode circuitry, the latter through the grid circuitry. These are then combined with the appropriate time synchronization to form the output of the transmitter.
Measurements of the output spectra show the following: the energy of the digital signal is more or less evenly spread across the 9/10 kHz assigned channel; the shoulders are steep, and drop rapidly to 40 dB or so below the spectral density level within the assigned 9/10 kHz channel, and the power spectral density levels continue to decrease at a lower rate beyond 4.5/5.0 kHz from the central frequency of the assigned channel.
1.1.4 Over the air
The digital phase/amplitude information on the RF signal is corrupted to different degrees as the RF signal propagates. Some of the HF channels provide challenging situations of fairly rapid flat fading, multipath interference that produces frequencyselective fading and large path delay spreads in time, and ionospherically induced high levels of Doppler shifts and Doppler spreads.
The error protection and error correction incorporated in the DRM system design mitigates these effects to a great degree. This permits the receiver to accurately decode the transmitted digital information.
A receiver must be able to detect which particular DRM system mode is being transmitted, and handle it appropriately. This is done by way of the use of many of the field entries within the FAC and SDC.
Once the appropriate mode is identified (and is repeatedly verified), the demodulation process is the inverse of that shown in the upper half of Fig. 1, the diagram of the transmitter blocks.
Similarly, the receiver is also informed what services are present, and, for example, how source decoding of an audio service should be performed.
1.1.6 Ongoing case study in Italy since 2006: DRM daytime MW Tests for frequencies below 1 MHz
The transmission site located near Milan was used to provide for an initial field test on frequency (693 kHz). The DRM signal is being broadcast by a station in Siziano, located 20 kilometres south of Milan. The same site is used to broadcast RAI’s regular analogue MW signals.
The analogue transmitter (working on 200 kW at 900 kHz) was combined with the digital transmitter (working on 34 kW at 693 kHz) and radiated by the same antenna structure.
On the basis of acquired data for the DRM transmission we can reach the following conclusions.
The whole north-west part of Italy is completely covered with a signal strength with a level greater than the minimum one indicated in Recommendation ITU-R BS.1698 for the adopted configuration transmission parameters (38,6 dBμV/m). Moreover minimum SNR of 14,1 dB was exceeded in each measurement point, also in deep valleys. The extension of coverage area can be identified with national border (Sestriere, Ceresole Reale, Domodossola and Bormio). On the east direction the DRM signal is available up to Trieste on which seacoast the field strength is 48,5 dBμV/m with a SNR of 21,7 dB. Due to particular topography and poor ground conductivity the Brennero valley was covered only before the town of Trento. In south-east direction DRM is available up to just before Ancona. In south direction DRM reaches all Liguria coast, and a part of Tuscany coast up to Grosseto town. The cities of Genova, Savona, La Spezia and Livorno are also covered.
The whole coverage results are indicated on Map 1. The inner contour shows the coverage area in which both commercial and professional receivers were able to decode DRM signal. The outer contour shows the coverage area in which only professional receiver was able to decode DRM signal.
Measured coverage area
The service area shown on Map 2 is computed on the basis of 45 dBμV/m for towns below 1,000 living persons and of 53 dBμV/m for towns with more than 1 000 living persons.
At the moment, about 150 static measurement points were verified.
Some data analysis was done in order to identify locations where reception was not available because of local particular situations:
– in the centre town of Turin, 125 km far from the transmitter, in 1 of 12 measurement points the performance of DRM signal has been damaged by an electric feeder for public transport. At that point was recorded a SNR of 13,4 dB with a signal strength of 52,1 dBμV/m and no audio decoding;
– northern from Milan, at the beginning of Valtellina valley (93 km far from the transmitter) some topographical situations and poor ground conductivity cause low signal strength (35,7 dBμV/m) and SNR (8,5 dB). Travelling along the valley route the signal and SNR come back to increase up to Bormio city, 170 km far from the transmitter.
Predicted coverage area (according to Recommendation ITU-R P.368-7)
During day time no discernable broadcasting interference situations were recorded in the whole predicted and measured coverage area.
As can be easily noted, measured and predicted area match quite well.