TECHNICAL ADVANCES IN DIGITAL AUDIO RADIO BROADCASTING

Transcription

1 TECHNICAL ADVANCES IN DIGITAL AUDIO RADIO BROADCASTING Ram Prasath.S and Sundar.K M.tech, Research Scholar, Paavai College of Engineering, Tamilnadu * Corresponding Author ABSTRACT: The worldwide installed analog broadcast networks suffer from problems in quality, power consumption, and waste of frequency spectrum. They will be replaced by digital networks in the next century. For analog networks, the planning parameters for frequency, power, and site assignment are well known. For digital audio broadcasting (DAB) networks, the research of the relevant criteria has just started. This paper presents the derivation of protection ratios and field strength levels under worst case conditions for Eureka 147 orthogonal frequency division multiplex (OFDM)-based digital audio broadcast systems from the bit error rate (BER) of the unprotected propagation channel. The channel is characterized by a three-dimensional (3-D) ray optical wave-propagation model. The peculiarities of the resulting multipath field strength delay spectra in single-frequency networks (SFN s) for the calculation of the signal-to-noise ratio (SNR) and the BER at the receiver are discussed. Different terrain types and frequencies are considered. A statistical analysis of the relationship between the SNR and BER leads to the determination of the co-channel protection ratio. An expansion thereof results in the adjacentchannel protection ratios. The minimum required field strength is derived from the protection ratios and the BER s. The influence of different spatial and temporal probabilities is investigated. Examples are shown for verification at 230 and 1472 MHz. KEYWORDS: Bit error rate, channel characterization, digital audio broadcasting, protection ratio. 44

2 INTRODUCTION: International Journal of Comprehensive Leading Research In Science ( ISSN: ) The assignment of sites, frequencies, and power for radio and TV stations in broadcast networks is based on signal to noise and interference ratios [1]. Sophisticated algorithms have been developed to handle this multidimensional problem. The latest published algorithms for network optimization, based on graph theory or simulated annealing [2],[3], lead to problems well known from other multidimensional tasks in engineering. It is experienced, especially in densely populated areas, that the assignment of power is primarily governed by the interference from transmitters in the same and neighboring channels. It is one of the goals and ideas of digital audio broadcasting (DAB) to reduce interference in single-frequency networks (SFN s) and to change and improve the power assignment[4].[5] Different system specifications for DAB are proposed in the United States, Japan, and Europe. This paper refers to the European Eureka 147 orthogonal frequency division multiplex (OFDM)- based DAB standard [6]. The receivers in DAB systems compress all signals arriving via different propagation paths within a given time interval, the guard interval, by correlation during the digital signal processing and perform a combining of the signals[7]. In this paper, the protection ratios are derived for DAB networks via the bit error rate (BER) of the unprotected channel from a three-dimensional (3-D) wavepropagation model. 3-D wave-propagation models calculate the multipath signals by ray optical methods For each multipath signal, the 3-D wave-propagation model calculates the following parameters: 1) Complex field strength of all multipath signals; 2) Time delay with respect to the first arriving signal; 3) Incidence angles (azimuth) and (elevation); 4) Polarization state. These parameters completely characterize each of this paper, the time delays and amplitudes are primarily of importance as both constitute the field strength delay spectrum (FDS) of the transmitter. Both are calculated deterministically[8]. The FDS depends on the receiver site and is different for each transmitter because of the different propagation paths in the 3-D environment together with the derivation of the BER s for the DAB typical 4-DPSK modulation. II. BIT ERROR RATE EVALUATION The BER is the most reliable measure for DAB coverage predictions, as it directly determines the quality of digital transmissions. In digital communication systems, it is determined by the 45

3 SNR, modulation type, interference and, to a large extent the FDS [8], [9]. In this section, the dependencies of the BER on the SNR, BER(SNR), and FDS are investigated. The resulting functional BER(SNR) characteristics are then approximated by heuristic analytical functions, which offer by far faster evaluations. These approximate functions are then applied to a representative test area. From the FDS, the investigation in the characteristics and dependencies of the BER is started. Fig. 1 shows an example for a computed FDS at a carrier frequency MHz in a hilly terrain with 967 multipath signals. The total FDS at the receiver site results by the superposition of the FDS s from all transmitters of an SFN. The total FDS determines, in the absence of noise, the BER of the unprotected propagation channel. The influence of fluctuating amplitudes is negligible for the following investigations. The receiver movement is only millimeters during the guard interval (64 and 256 s, respectively, depending on the carrier frequency) for velocities up to 250 km/h. Field strength amplitudes are rather stable over these temporal and spatial intervals. The phases of the signals in delayed rays, i.e., the reflected, scattered, or diffracted rays, are randomly distributed over 2π [10]. These phases vary in location and time. The Doppler effect is 46

4 not taken into account. The signal components of the total FDS are divided into three classes according to their time of arrival as follows. 1) Reference signal: The first signal, often arriving in a line-of-sight propagation path, from any one of the transmitters, determines the reference in the total FDS. 2) Delayed signals, arriving within the guard interval class. 2) For these signals, the phases are assumed to be randomly distributed, while the amplitudes are considered to be deterministic. 3) Delayed signals, arriving outside the guard interval class 3: For these signals, the phase is also randomly distributed, and the amplitude is deterministic, but their information is not related to the reference.. Additional Gaussian noise is added as a fourth class. The computation of the BER for the unprotected channel is based on the total pdf resulting from the convolution of the pdf s from each class. 47

5 For the derivation of the symbol error rate (SER) and thus the BER, fr Mi (r Mi) is integrated over the decision area Ri. As shown in fig. 4, for the case that symbol #i is transmitted 48

6 49 International Journal of Comprehensive Leading Research In Science ( ISSN: )

7 BER on a pixel by pixel basis for the whole DAB coverage area as a function of the local total FDS. In (1) (4), the noise level.the calculations are precise, but time consuming for larger areas. Therefore, it is desirable during the design of a DAB network to skip the above BER calculations for larger areas by means of a heuristic approach. In the following, a proposal for a time-saving method is derived. The key is the nonlinear influence of the SNR on the BER, which is subsequently investigated. Since the FDS s and consequently the BER s depend on the receiver site and the terrain over which the signals propagate. The effect on the resulting BER s and SER s is observed and expressed as a function of SNR by BER(SNR) and SER(SNR). III. THE COCHANNEL PROTECTION RATIO FOR DAB The required BER of the unprotected propagation channel, which still provides sufficient signal quality.. The dependencies of the SER and BER on the SNR are nonlinear. Independent of the terrain and the frequencies investigated above the dependency of the SER from the SNR has as a characteristic in common: for SNR exceeding 20 db the SER is not further improved. This allows the following conclusion. If an SNR of 20 db can be guaranteed, the DAB signal at the receiver output has the best possible quality. I.e., if the value of the SNR decreases below 20 db, the BER starts to increase. This results in a worst case assumption of a required co channel protection ratio of 20 db. IV. THE ADJACENT-CHANNEL PROTECTION RATIOS Similar considerations for the derivation of the cochannel protection ratio of 20 db are used to derive the adjacent channel protection ratios for DAB against other broadcast services. These investigations are based on the spectra of the different broadcast services. Here, the real spectra of the interfering services are considered. A. Protection Ratio for Other Interfering DAB Services The DAB spectrum given in [6] can easily be simulated on a computer. At no carrier frequency of the DAB signal, which has to be protected, the interfering signal from another DAB transmission may be less than 20 db down. The critical range is always at the edges of the 1.5-MHz-wide DAB spectrum. The spectral power density is rapidly decreasing with the frequency distance Δf from the interfering carrier. The resulting adjacent-channel protection ratios for interfering DAB services, depending on the frequency distance Δf, are given in 50

8 Table III. This worst case estimation for neighbor-channel protection ratios shows that a channel spacing of 50 khz is sufficient. If an analog TV channel is reused for four DAB transmissions, a spacing of 200 khz can be realized. This corresponds to an adjacent channel protection ratio of - 40 db. B. Protection Ratio for Interfering TV The considerations for protection ratios for interfering TV systems are also based on their spectral power densities. The results are shown in Table VI for DAB operating in a frequency band above TV. The frequency separation Δf is defined as the distance between the second audio carrier frequency of the TV signal and the first carrier frequency of the DAB signal. The different TV standards PAL, D-SECAM, and LSECAM have to be treated separately because of their different audio signal modulation type. For DAB operated in a frequency band above TV, the audio carriers of the TV signal are considered to be the most critical. Since both PAL and D-SECAM have FM-modulated audio signals, they have the same values for the protection ratios given in Table IV. L-SECAM uses AM and therefore needs a greater protection ratio. If DAB is operated in a frequency band below TV, the values given in Table V apply. They are based on the filter characteristics of TV transmitters defined by CCIR [13] for PAL and SECAM in a frequency separation of Δf=15MHz below the AM carrier frequency of the picture signal. 51

9 52

10 C. Protection Ratio for Interfering FM The final decision whether DAB should replace VHF-FM audio broadcasting in the frequency range from 88 to 108 MHz has not yet been made. The decision has to be supported by the calculation of the protection ratios for DAB in the neighborhood of FM, as both may be operated in parallel for many years. The frequency separation to FM is counted from the first DAB carrier frequency to the FM carrier frequency. A simulation of a VHF-FM audio broadcasting signal is used to investigate its spectral characteristics V. THE MINIMUM REQUIRED FIELD STRENGTH The initial calculations for the co channel protection ratio assume white noise. In the following, the minimum required field strength for a DAB receiver at 1472 MHz is calculated based on these assumptions. In these calculations, typical values for the receiver hardware components are used. The resulting total DAB noise temperature is determined as follows: Antenna cable loss : 1 db NF down converter : 5 db Gain down converter : 15 db NF receiver : 9 db Resulting total noise temperature: K. The noise figure for the receiver meets the specifications for existing and upcoming DAB receivers of the fourth generation. The noise figure for the down converter is specified by several suppliers to be less than 6 db, and the gain is announced to be up to 15 db. With this noise temperature, the spectral noise power density results in dbm/hz. VI CONCLUSIONS For the first time, the most important system parameters for DAB network planning are derived in a worst case estimation from the field strength delay spectra of 3-D wavepropagation models. Measured as well as calculated field strength, delay spectra vary significantly with the type of terrain. The functions of SER(SNR) are nonlinear and depend strongly on terrain types and carrier frequencies. A statistical analysis of approximately 2000 SER(SNR) functions encourages the worst case estimation of 20 db for the cochannel protection ratio for DAB. The main research result can be expressed by two figures Co channel protection ratio 20 db.minimum field strength 54.7 db V/m. If these criteria are satisfied, the received signal will have the best possible quality. The necessary correction 53

11 factors for the 50% field strength prediction values of 3-D wave-propagation models for other DAB specific spatial and temporal probabilities are calculated. Two typical figures obtained by this investigation are Correction factor for a temporal or spatial probability of 90% : 9 db Correction factor for a temporal or spatial probability of 99%:18 db. The correction factor has to be applied on the 50% field strength prediction values separately for time and space. These values have shown to be independent of the investigated terrain types The methods presented in this paper are used for the planning of a DAB network in the Rhine Valley and for two additional -band DAB networks in and around Stuttgart, Germany, with 12 and 14 transmitters, respectively. REFERENCES [1] Planning Parameters and Methods for Terrestrial Television Broadcasting in the VHF/UHF Bands, EBU-Tech E, May [2] A. Gamst, Some lower bounds for a class of frequency assignment problem, IEEE Trans. Veh. Technol., vol. VT-35, pp. 8 14, Feb [3] A. Quellmalz, A. Kn almann, and B. M uller, Efficient frequency assignment With simulated annealing, in 9th Int. Conf. Antennas and Propagation (ICAP 95), vol. 2, pp [4] W. A. Stacey and S. L. Edwards, Digital radio primer, in Proc. 2nd Int. Symp. Digital Audio Broadcasting 1994, Toronto, Canada, pp [5] G. Chouinard, R. Voyer, and R. Paiement, Coverage concepts for digital audio broadcasting at 1.5 GHz, in Proc. 2nd Int. Symp. Digital Audio Broadcasting 1994, Toronto, Canada, pp [6] Radio Broadcast Systems; Digital Audio Broadcasting (DAB) to Mobile, Portable and Fixed Receivers, EBU/ETSI JTC Draft prets , Valbonne, France, Mar [7] A. Lau and W. F. Williams, Service planning for terrestrial digital audio broadcasting, EBU Tech. Rev., no. 252, pp. 4 26, [8] Th. K urner, T. C. Becker, and D. J. Cichon, Analytical BER determination for digital radio systems, in COST 231 TD (94) 67, Prague, Czechslovakia, Apr [9] Th. K urner, D. J. Cichon, and W. Wiesbeck, Evaluation and verification of the VHF/UHF propagation channel based on a 3D-wave propagation model, IEEE Trans. Antennas Propagat., vol. 44. pp , Mar

12 [10] Th. K urner, M. Lebherz, W. Wiesbeck, and W. Krank, Signal characteristics of the VHF/UHF propagation channel, in Proc. IEEE AP-S Int. Symp. URSI Radio Science Meeting, Ontario, Canada, June 24 28, 1991, pp Access This Article in Online Website: QR Code Subject: Department Of ComputerEngineering How to cite this article: Ram Prasath.S and Sundar.K, Technical advances in digital audio radio broadcasting Vol 2 Issue 7, pp :