1、100 - STD.ITU-R RECMN Sfl-853-1-ENGL 1997 II Li855212 0534i!83 345 Rec. ITU-R SM.853-1 RECOMMENDATION ITU-R SM.853- 1 NECESSARY BANDWIDTH (Question ITU-R 77/1) (1992-1997) The TTU Radiocommunication Assembly, considering that the concept of “necessary bandwidth” defined in No. 146 (S1.152) of the Ra
2、dio Regulations, is useful for a) specifying the spectral properties of a given emission, or class of emission, in the simplest possible manner; b) that with regard to the efficient use of the radio-frequency (RF) spectrum, necessary bandwidths for individual classes of emission must be known, that
3、in some cases the formulae listed in Recommendation ITU-R SM. 1 138 can only be used as a guide and that the necessary bandwidth for certain classes of emissions is to be evaluated corresponding to a specified transmission standard and required quality; c) that changes in technology have resulted in
4、 additions and variations in the modulations used for radiocommunication: d) in signal characteristics (e.g. average talker level), that ihe numerical parameters used in the necessary bandwidth formulae may change with time due to changes recommends that the necessary bandwidth formulae (contained i
5、n Recommendation ITU-R SM. 1138) be supplemented with the following formulae. 1 Multi-channel frequency division multiplex - frequency modulation (FDM-FM) emissions To account for changes in average talker level, which may occur over time, the formula for the necessary bandwidth B, of multi-channel
6、FDM emission is: d x 3.76 x antilog where: M: maximum modulating frequency (Hz) d: per-channel deviation N, : number of circuits in the multiplexed message load K: unity X= -2 to +2.6 for 12 I Ne . = 0.06675 X 10-6 = 0.167 x s Bandwidth: 3.36 x lo6 Hz 4M50PON Annex 2 contains the method used for det
7、ermining the necessary bandwidth of the unmodulated pulses. 3 Digital modulation The necessary bandwidth and example K values for several digital modulations are given in Table 2. Annex 3 contains the methods used for determining the necessary bandwidths for digital modulation. 102 STD-ITU-R RECMN S
8、M.853-L-ENGL L577 m 9855212 053V285 LLB m Rec. ITU-R SM.853-1 TABLE 2 Digital modulation Modulation and conditions m 2-PSK (unfiltered) S = 2 (computed) 2-PSK (filtered, BER = 1 x lW3) S = 2 (computed) MSK (unfiltered) S = 2 (computed) I D = 0.25 R Gaussian filtered MSK (GMSK) 3 dB premodulation Gau
9、ssian filter bandwidth = 0.25 R S = 2 (computed) D = 0.25 R Digital FM (Continuous phase FSK) rectangular pulses S = 2 (computed) D = 0.35 R nI-QAM Microwave digital s = 2“ (n 2 2) Roll-off = O to 1 50% splitted Tx/Rx optimally filtered (computed)(4). (3 Necessary bandwidth formula 2RK B - - log2S 2
10、 RK Bn = +2DK BR = 2 RK B - - log2S Example K value 10.28 2.0 0.36 3.52 -0.28 0.18 0.89 See Fig. 1 ercentage fractional power containment bandwidth(l) 99 95 100 100 99 99.9 99 99.9 SeeFig. 1 BER: binary error ratio. FSK: frequency shifting key. MSK: minimum shifting key. PSK: phase shift keying. QAM
11、: quadrature amplitude modulation. (I) Recommendation ITU-R F.1191 foresees that for digitally modulated systems in the fixed service the necessary bandwidth should be defined for a percentage fractional power containment equal to 99%. (*) For this case EbMo = 7.5 dB. (3) For this case Eb/No = 9.3 d
12、B. (4) Practical filtering may give slight difference in the K value versus containment computed relationship. (3 4- and 8- QAM formats coincide with filtered 4- and 8- PSK formats. 1 STD-ITU-R RECMN SM-853-1-ENGL 3997 = 4855232 D53Li28b 054 D Rec. ITU-R SM.853-1 103 FIGURE 1 K value versus roll-off
13、 for m-QAM modulation formats Parameters: power containment factor Roll-Off Containment power (96): - 95 97 99 - - 0853-01 ANNEX 1 FDM-FM necessary bandwidth calculations lulti-channel FDM-FM emissions Recommendation ITU-R SM.1138 “Determination of necessary bandwidths including examples for their c
14、alculation and associated examples for the designation of emissions” includes under its Annex 1, Table III-B, the necessary factors for use in computing peak frequency deviation of multi-channel FDM-FM emissions. Peak frequency deviation is a critical factor in Carsons rule, B, = 2Mt 2DK, used to ca
15、lculate necessary bandwidth for frequency spectrum allocation purposes. Table III-B is reproduced as Table 3. The factors 2.6, -1 and -15 in the Table are the average power (dBmO (see Note 1) values that were found in a standard, commercial telephone, public switched network circuit. The values were
16、, in fact, based upon measurements of “talker volume” conducted in 1960 that were previously agreed to in the ex-CCIR and eventually at the World Administrative Radio Conference (Geneva, 1979) as applicable, for the purposes of necessary bandwidth calculation. NOTE 1 - “dBmO” refers to the power (dB
17、) relative to 1 mW referred to a point of zero relative transmission level. 104 Number of telephone channels, Nc 3 Nc 12 Rec. ITU-R SM.853-1 Multiplying factor() 1 Value in dB above Value in dB specified by the equipment manufacturer or station licensee, subject to administration approval 4.47 x ant
18、ilog 20 1 TABLE 3 Multi-channel FDM-FM emissions 60 I Nc 240 Nc 2 240 MULTIPLYING FACTORS FOR USE IN COMPUTING D, PEAK FREQUENCY DEVIATION, IN FDM-FM MULTI-CHANNEL EMISSIONS - + ;ologNcI 1 3.76 x antilog 3.76 x antilog -15 + i: log NC For FDM-FM systems the necessary bandwidth is: Bn = 2M+ 2DK The v
19、alue of D, or peak frequency deviation, in these formulae for B, is calculated by multiplying the r.m.s. value of per-channel deviation by the appropriate “Multiplying factor” shown below. In the case where a continuity pilot of frequency& exists above the maximum modulation frequency M, the general
20、 formula becomes: B, = 2fp + 2DK In the case where the modulation index of the main carrier produced by the pilot is less than 0.25, and the r.m.s. frequency deviation of the main carrier produced by the pilot is less than or equal to 70% of the r.m.s. value of per-channel deviation, the general for
21、mula becomes either B,= 2fp or B, = 2M+ 2DK 12 I Nc 60 3.76 x antilog r.6 + iolog Nc () In the above chart, the multipliers 3.76 and 4.47 correspond to peak factors of 11.5 dB and 13.0 dB, respectively. In 1975 and 1976, further measurements of speech signal power were made in the same circuits and
22、networks, using similar methodology, so as to allow direct comparison of results to the earlier work. The later measurements have been under study, both in industry and government, since that time, finally leading to modifications to typical domestic applications in the public switched telephone net
23、works. STD-ITU-R RECMN SM=853-L-ENGL I777 I 4B55212 0534288 727 W Rec. ITU-R SM.853-1 105 To summarize the 1975-1976 study, which included the reasons why differences from the earlier study occurred, it was found that substantial changes have aggregated over time to produce an average decrease in ac
24、tual average talker power level of approximately 4.6 dB. The changes have tended to increase the uniformity of service in the switched public networks from the viewpoint of speech volumes. These include a decrease in the proportion of toll grade battery systems, loss plan improvements, upgraded tele
25、phone sets, and an increase in direct trunking. Direct distance dialling has become normal practice with new loop and trunk design techniques. Furthermore, advanced digital data acquisition technology facilitated the measurement in 1975-1976 of speech signal power with greater precision than was pos
26、sible in 1960 when volume unit meters were used for that survey. Standard deviation of all measurement between the 1960 and 1975-1976 surveys, was reduced by an average of about one third, from 7 to 4.6 volume units. A multi-stage statistical sampling scheme (near-end and far-end) of talker power me
27、asurements on more than 10 O00 calls, originating from approximately 2 500 loops, was used. Average conversational signal power (averaged over the entire observation interval), and a new measure of speech level known as equivalent peak level (EPL) were the measures used to characterize talker signal
28、s. Loop d.c. current, class of service, switch type, and call destination were recorded as part of the 1975-1976 study. NOTE 2 - Volume unit is one way of measuring speech level with a power-level indicator calibrated in terms of dB for a steady sine-wave voltage, with 1 mW in 600 i2 taken as a refe
29、rence. The response of the indicator is not frequency weighted. Volume unit readings are averages achieved by a particular set of meter ballistic (mechanical) characteristics. The later study indicated that with (1975 and 1976) public switched network systems, there was little dependence of speech s
30、ignal power on call destination or originating class of service (residence, business, local, toll, and combined). Small differences are explained for the most part by loop characteristics. There is little, if any, variation in speech signal power attributed by study conclusions to psychological fact
31、ors such as call distance or perception of received volume. The average values measured indicate that the switched telecommunication network in 1975-1976 was essentially transparent to customers in the sense that talker signal power was not found sensitive to call distance, local or toll call classi
32、fication, or other factors that are outside local loop circumstances. In summary, it is believed that the later measurements of telephone “talker level” indicate the same, normal speech levels of the population when not using the telephone. It is concluded that little or no further change in represe
33、ntative talker level is expected in the future on the switched public telephone network where the measurements were taken. Consequently, this fact should be taken into account in an examination of the necessary bandwidth formulae for FDM-FM systems when telephone speech comprises a significant part
34、of the total FDM-FM circuit. In the B, formula as shown in Table 3, the three factors 2.6, -1 and -15 are components of the multiplying factor used to determine D, or peak frequency deviation. Practical consequences of a reduction in “talker power” on the telephone circuit feeding an FDM-FM radio ci
35、rcuit is less peak deviation. There are three independent parameters that determine the peak deviation of the FM signal, all of which are constrained by the design of the system so as to limit the maximum value of each so that D does not exceed a predetermined value (in FDM-FM systems). Those variab
36、les are: - r.m.s. value of per-channel deviation, - average power in a message channel, - number of total channels in the multiplexed message load. If the average power of speech signals can be reduced, as indicated in the 1975-1976 study, a trade-off would be possible with the two other parameters.
37、 This could be achieved by: - increasing the number of channels in the same frequency bandwidth, - increasing spectrum efficiency by reducing the bandwidth for the same number of channels, - a combination of the two methods. For example, in systems using 4 kHz voice grade message channels, it would
38、be possible to vary the ratio of data to voice traffic. A user could choose to remain within a specified RF bandwidth and then select an average message channel power level which could be used to make a trade-off between an increase in the number of message channels against their individual frequenc
39、y deviation. For a given ratio of data traffic to voice traffic, a user might consider that a too large increase in the number of 4 kHz message channels would reduce the per channel deviation to the point where signal quality would become degraded. However, an average message channel power level cou
40、ld be chosen which keeps STD-1TU-R RECMN SM=853-L-ENGL 1997 I 4855232 0534287 8b3 W 106 Rec. ITU-R SM.853-1 within the specified RF bandwidth and which can permit some increase in the number of message channels while still allowing a proportional increase in the individual channel frequency deviatio
41、ns. By this method, instead of using all the available spectrum to maximize the number of message channels, it would be possible to obtain an increase in the value of the per-channel deviation which would provide an improvement of the signal-to-noise ratio and a probable reduction of BER of the data
42、 traffic. A microwave system can therefore be designed for which the parameter for the average message power level is chosen between the level used in the current equations and one 4.6 dB lower. There are valid reasons for changing (i.e. increasing) the other available parameters in the system when
43、the average message power level (talker power) is decreased. However, a decision to carry out the change is appropriate for the carriers (or governments) concerned to make, taking account of the advantages and disadvantages of the possible trade-offs. The important point here is recognition of the c
44、hanged average “talker power” level. To take account of the decrease in average “talker power” by 4.6 dB in speech telephone circuits, typically the 2.6, -1 and -15 values are replaced by the variable X which can then range between limits comprised of the present values, and corresponding values tha
45、t are 4.6 dB lower, depending upon the total number of circuits in the FDM-FM system, and the composition of the system itself. Thus, the necessary bandwidth, B, is: d x 3.76 x antilog where: M: maximum modulating frequency (Hz) d: per-channel deviation N, : number of circuits in the multiplexed mes
46、sage load K: unity X= -2 to +2.6 for 12 I N, 60 and for Y = 2 X= -5.6 to -1.0 for 60 I N, 240 and for Y = 4 X= -19.6 to -15.0 for N, 2 240 and for Y = 10. The term enclosed in brackets is the peak deviation, D. The numerator (X+ Y log N,) of the fraction represents the average power of the composite
47、 signal delivered to the modulator input of the transmitter. The value of 3.76, as noted in Table 3, corresponds to a peak factor of 11.5 dB. To correctly select a value for X in the formula for B, it is helpful to summarize applicable conditions under which it is used in an FDM-FM system. Final cho
48、ice within the 4.6 dB range may be empirical. It is clear from the study details compiled that average “talker power levels” of -2, -5.6 and -19.6 dBmO should replace the corresponding numbers for N, 2 12 in the B, formulation with an FDM-FM system that is used to provide trunk connections for modem
49、 commercial public telephone circuits where most of the FDM-FM channels are used for speech. In smaller, private, or older FDM-FM systems, particularly those with N, 12 or those containing data (non voice) in most of the channels, the original values, as shown in Table 3, nominally apply. Typical multi-channel data circuits operate at power levels from -13 to -15 dBmO. Therefore, the composite loading limit will be determined by using the value X= -13 to -15 for systems with a large percentage of data circuits for N L 240. Individual channel signalling, as opposed to c
