Skip to main content

Full text of "BBC Research and Development Report Number 1978-20"

See other formats


BBC RD 1978/20 



13(33 # 

THE QU€£N S AWAHD 

RESEARCH DEPARTMENT REPORT 



ilRROW-BAiD F.M. SYSTEM FOR TELE¥ISION LINIS: 

a feasibility study 



R. Storey, B.Sc. 



Research Department, Engineering Division 

THE BRITISH BROADCASTING CORPORATION July 1978 



BBC RD 1978/20 

UDC 621.397.232.6 



NARROW-BAND F.M. SYSTEM FOR TELEVISION LINKS: 

A FEASIBILITY STUDY 

R. Storey, B.Sc. 



Summary 

The use of f.m. television links at u.h.f. (Band V) depends, at present, upon 
the availability of a number of adjacent 8 MHz broadcast television channels to 
accommodate the 16 MHz bandwidth occupied by the transmitted signal. 

Because of the expansion of the u.h.f. transmitter network, fewer spare sets of 
contiguous channels are expected to be available. To overcome this difficulty, it was 
proposed that the f.m. television signal could be modified to work within a bandwidth 
of only 8 MHz. To evaluate the proposal, an experimental narrow-bandwidth f.m. 
system was constructed. 

It was found that f.m. television signals could indeed be transmitted successfully 
in a nominal bandwidth of 8 MHz, instead of the nominal 1 6 MHz used at present. Asa 
consequence of the bandwidth reduction, both linearity and signal-to-noise ratio were 
slightly impaired, nevertheless it is thought that a link of this kind should be advantageous 
under certain conditions. 



Issued under the authority of 




^,^7ru^^ 



-^^Tppl 



Research Department, Engineering Division, 
BRITISH BROADCASTING CORPORATION 

July 1978 
(EL-138) 



Head of Research Department 



NARROW-BAND F.M. SYSTEM FOR TELEVISION LINKS: 
A FEASIBILITY STUDY 

Section Title Page 

Summary Title Page 

1. Introduction 1 

2. Theoretical considerations 1 

2.1 Description of the basic spectrum 1 

2.2 Spectrum shaping 3 

2.3 The sound channel 3 

3. Experimental arrangement 3 

4. Description of tests 4 

5. Discussion of results 5 

5.1 Differential gain and phase distortion 5 

5.2 Tests with augumented puise-and-bar signal. 6 

5.3 Noise performance . 6 

5.4 The sound channel 7 

6. Conclusions 7 

7. References 7 

Appendix 8 



(EL-138) 



NARROW-BAND F.M. SYSTEM FOR TELEVISION LINKS: A FEASIBILITY STUDY 

R. Storey, B.Sc. 



1. Introduction 

The f.m. television links used for outside broadcasts at 
present operate at s.h.f. for point-to-point use, and at u.h.f. 
Band V for point-to-point or mobile use. The Band V links 
need spectrum space in which to work, and it is becoming 
increasingly difficult to find sufficient space at some link 
sites because of interference from the expaning u.h.f. trans- 
mitter network and the corresponding risk of interference 
to domestic reception of u.h.f. broadcasts from the link. 
Some means of carrying the television signal in a reduced 
bandwidth would therefore be very useful, hopefully leading 
to an increase in the number of sites where links may be used 
in the future. 

The purpose of the work described in this Report was 
to investigate the feasibility of transmitting an f.m. television 
signal with a vestigial sideband spectrum in a bandwidth of 
only 8 MHz, instead of the 16 MHz required for the present 
double sideband system, and to identify any additional 
system requirements that might arise as a consequence of 
this bandwidth reduction. 

Three further Reports will consider the effect 

of narrow band operation on interference between f.m. 
links and a.m. broadcast signals, and the susceptibility of 
d.s.b. and v.s.b. signals to multipath conditions.* 

A theoretical appraisal was made of several possible 
spectrum-shaping schemes, and the conclusions were verified 
in some preliminary experiments, before arriving at the 
spectrum shape described below. A system was then evolved 
by measuring the impairments introduced and minimising 
these by the various means to be described. 



2. Theoretical considerations 

2.1 Description of the basic spectrum 

Fig.1 shows thespectrumofa70 MHzcarrier, frequency- 
modulated by a standard-level a.c. -coupled "non-linearity 
test" waveform. The deviation sensitivity is 4 MHz/volt 
after CCIR 625 line pre-emphasis. The test waveform used 
(shown later in Fig. 8) consists of a "staircase" on each 
line of the television video signal with a sine-wave at 
colour subcarrier frequency superimposed on each of the 
six steps of the staircase. 

It can be seen from Fig. 1 that the f.m. spectrum 
consists of a central group of components, carrying mainly 

* It should be pointed out here, that as a result of the work 
described in Reference 4, it was concluded that, with the system 
of Band V broadcast channel allocations which exists in the U.K., 
not enough new link sites are made available to justify the extra 
expense of the v.s.b. system dexribed in this report. In a 
different system of channel allocations, however, a v.s.b. system of 
this type could significantly increase numbers of available link 
sites. 







.s; 


1(1 










a 




"S 






7n 






X 




E 










:i(i 


o 


















40 


fe 












.^ 


Kn 














to 




-a 


60 






(i 













62 66 70 74 78 

FREQUENCY, MHz 

Fig. 1 - Spectrum of the f.m. television signal when the 
modulation is a staircase waveform (with colour subcarrier). 

low-frequency picture information, and two asymmetrical 
groups of components spaced at ±4.43 MHz from the carrier- 
rest frequency (70 MHz) carrying predominantly high- 
frequency and colour information. Low-frequency compo- 
nents in the modulation waveform have a large modulation 
index, i3 (the ratio of peak frequency-deviation to modul- 
ation-frequency), and therefore produce many closely- 
spaced sidebands at the centre of the f.m. spectrum, whereas 
high-frequency components have a low modulation index 
and produce two groups of first-order sidebands at the 
modulating frequency from the carrier rest frequency; the 
higher-order sidebands are of relatively low level. Because 
the higher-frequency components of the modulating signal 
producesidebandsabove and below the carrier rest frequency, 
it would appear that a saving in band-width could be effected 
by eliminating one set of the sidebands. 

^<1 j8>i ^3<^ 




Q._Qn I . . . ■ 

E ""^ 60 62 64 66 68 70 72 74 76 78 80 

D 

frequency, MHz 

Fig. 2 - Sketch of the f.m. spectrum of Fig. 1 showing the 
division into three parts. 



(EL-138) 



test 
signa 
input 



bandwidth of conventional system ( Ri16MHz) 



3 



Q. 

E 
a 







bandwidth of v.s.b 
, system ( ^SMHz) , 



J L 



carrier rest 
frequency 



J L 



J L 



60 62 64 66 68 70 72 74 76 78 80 
frequency, MHz 

Fig. 3 - Idealised frequency characteristic of the i.f. filter. 



I _^ 



pre- 70 MHz f.m. 

# emphasis modulator 



frequency 
counter 



'^B 




BB^Bi 



6dB 



group delay 
i.f. filter corrector 
(i) (2) 



limiter 



/ 



-ye 



[> 



lOdB 



-»- 



4> 



-»- 



6dB 



/ 



K^ps 



70 MHz f.m. de- 

demodulator emphasis 



video 
equaliser 




phase 
corrector 



distortion 
analyser 



->■ 



^ 



O 



->- 



(-1) 64-72 MHz b.p.f. for v.s.b link 
62-78 MHz b.p.f. for d.s.b link 

(2) Not used for d.s.b. link 



(EL-138) 



Fig. 4 - Experimental Arrangement 
-2- 



2.2 Spectrum shaping 



3. Experimental arrangement 



Thethree main groups of spectral components produced 
by the non-linearity test signal are further identified in Fig. 2. 
This shows one group centred around the carrier rest 
frequency and extending over 4 MHz (the peak-to-peak 
deviation); the two other groups of spectral components 
are located one on each side. A television picture signal will, 
in general, contain a continuum of frequencies, and under 
these conditions the sideband structure is not readily 
divided into such well defined-groups. 

A filter which confines the spectrum to within the 
frequency-range (of 4 MHz) through which the carrier is 
deviated will introduce severe non-linear distortion of the 
modulation. This is because a modification of the relative 
amplitudes of the sidebands in this frequency range by such 
a filter causes the deviation sensitivity of the system to vary 
with the instantaneous level of the low-frequency (luminance) 
information in the modulating video signal; thus non-linear 
distortion results. 



A filter with an asymmetrical passband — passing all 
the inner sidebands and the outer sidebands on one side 
of the carrier — will produce what can be regarded as a 
vestigial sideband (v.s.b.) signal. If the spectrum is shaped 
in this way, the effect is to reduce the deviation for the 
higher video frequencies. The change in deviation which 
results in this case is much less dependent upon levels in 
the modulating signal, and the distortion introduced is 
principally linear, affecting the amplitude/frequency response 
of the system. Non-linear distortion does occur in the 
part of the spectrum near the skirts of the VSB filter, 
but this is not important because the energy in this region 
is low. This is dealt with in greater detail in the Appendix. 



Practical tests were conducted using the experimental 
arrangement shown in Fig. 4. The modulator and de- 
modulator were standard commercial units operating with 
a carrier rest frequency of 70 MHz (the intermediate 
frequency used in most link equipment); the filters were 
therefore designed for use at 70 MHz. 

The demodulator unit had limiter stages at i.f., but an 
extra limiter was constructed for use immediately before 
the demodulator. Limiting (and thus performance) were 
found to be much better with this extra limiter included. 

Bandpass filters for use at 70 MHz were constructed as 
separate high-pass and low-pass sections for instrumental 
convenience. An i.f. group-delay equaliser was also con- 
structed for use with the v.s.b. filter; it enabled the group- 
delay variation with frequency introduced by this filter to 
be reduced by about 30%. The amplitude and group-delay 
characteristics of the narrow-band filter (with and without 
the group-delay equaliser) are shown in Fig. 5. It will be 
seen, with reference to the photograph of the complete f.m. 
spectrum in Fig. 1, that the narrow-band filter passes the 
principal spectral components around the carrier rest 
frequency of 70 MHz, plus the sidebands extending down 
to 64 MHz, including the important group of sidebands 
centred on (70 - 4.43) MHz. Sidebands above 72 MHz 
and below 64 MHz are attenuated. The spectrum obtained 
after the signal has passed through the narrow-band filter 
is shown in Fig. 6. 

The required video equaliser response was found by 
measuring the video amplitude/frequency response of the 
vestigial sideband f.m. link shown in Fig. 7. A simple 
bridged-T network was designed to fit this curve and video 



It was decided, to employ a v.s.b. filter with a 
characteristic approximating to the ideal rectangular shape 
shown in Fig. 3, and to correct for the resultant linear 
distortion of the video signal using an equaliser. It was 
decided that the upper portion of the spectrum should be 
removed because the modulation polarity was such that the 
synchronising pulses caused deviation towards the higher 
frequencies; it was thought that any non-linearity intro- 
duced by filter group-delay at the pass-band edge would be 
less significant if only the synchronising pulses were 
affected. A further 2 MHz was removed from the lower 
end of the spectrum to give a channel width of 8 MHz. 

2.3 The sound channel 

Existing links use a 7.5 MHz f.m. sound subcarrier 
added to the composite video signal. As a consequence of 
the reduced bandwidth of the v.s.b. system, the sound sub- 
carrier frequency must be reduced, and 6 MHz was chosen 
as a convenient frequency. Intermodulation products, 
principally those at f - f and 2f - f (where f = sound 
carrier frequency and f = colour subcarrier frequency) 
will of course be increased as a result of the poorer linearity 
of the v.s.b. system. 




60 



65 70 

frequency, MHz 



Fig. 5 - Frequency characteristics of the 64 - 72 MHz 
bandpass filter. 

—— Amplitude 

_ Group delay 

_ Group delay with partial correction 



(EL-138) 







■^ 


u 






o 




■D 










71 














E 










3( 




















41 






Q 








> 


151 


ro 




^ 




CQ 




X> 


61 














o. 




" 


71 



^;= 



ij- 







66 70 . 

FREQUENCY, MHz 



Fig. 6 - Spectrum of signal after v.s.b. filtering 

group-delay correction was achieved using a Cintel 'B' 
group-delay corrector. 

Provision was also made for adding a 6 MHz f.m. sound 
subcarrier to the input video signal before modulation and 
for retrieving it after the f.m. demodulator. 

4. Description of tests 



The initial setting-up and testing of the narrow-band 
f.m. system was performed using two non-linearity television 
test signals, one having a step ("staircase") waveform with 
colour subcarrier on each step as shown in Fig. 8, and the 
other comprising CCIR test waveforms using one line of 
staircase alternating with three lines of either black level or 
white level. The performance of the link was measured in 
terms of the differential gain and differential phase distortion 
of the colour subcarrier at the video output, for various 
values of carrier deviation. The effect of partially equalising 
the group-delay response of the narrow-band filter was also 
investigated. 




Fig. 8 - Staircase waveform at the output of the waveform 

generator 



In order to evaluate performance under noisy con- 
ditions, tests were conducted with noise injected into the 
system at the input to the i.f. filter. Injection of noise at a 
high level was adopted in preference to attenuation of the 
carrier, so as to ensure adequate limiting even at low carrier- 
to-noise ratios. 

The output signal-to-noise ratio was then measured for 
a number of different carrier-to-noise ratios at the input to 
the limiter. Carrier-to-noise ratio was determined by using 
a bolometer power-meter to measure, alternately, carrier 
power (in the absence of any significant noise) and noise 
power (in the absence of carrier) at the input to the limiter. 
Noise measurements were also made using a d.s.b. 16 MHz- 
bandwidth i.f. filter, keeping the input signal unchanged 
as for a normal f.m. transmission with 4 IVIHz peak-to-peak 
deviation; in this case the carrier-to-noise ratio was 
determined with reference to 16 MHz bandwidth instead of 
8 MHz. 




0-2 



0-3 0-4 



0'6 0-8 1 2 3 

frequency, MHz 

Fig. 7 - Frequency characteristic of the v.s.b. link before equalisation. 



(EL-138) 



■4- 



a 



.E 6 
o 
en 
_5 

O 






a\ 






^X 



o _!. — «-' 
__ 0— -x''''a 



i 


k 




«>8 


- 




(0 






CD 

fe7 


- 


o 


0) 
X) 




A 


nrfci 


- 




u> 




/ 


x: 5 


_ 


/ 


Q. 




^) 


o4 


i 


/x 


^ -, 


J 


3 


CD 


■^ r 




ISz 


- °\x V^ 




•1 

n 


1 1 1 


( 



69-8 700 



70-2 70-4 706 
frequency, MHz 



70-8 



69-8 700 



70-2 70-4 70-6 
frequency, MHz 



70-8 



F/fir. S 

(a) Differential gain distortion as a function of the frequency corresponding to black level 

(b) Differential phase distortion as a function of the frequency corresponding to black level 

o Points obtained for CCIR "bar on" waveform, x Points obtained for CCIR "bar off" waveform. A Points obtained for staircase waveform. 



The behaviour of the system with an added sound 
subcarrier was also studied. The 6 MHz f.m. sound sub- 
carrier was added to the vision signal and impairments to 
the picture were assessed subjectively, for various sound 
subcarrier amplitudes, with both the v.s.b. and the full 
bandwidth system. The sound channel signal-to-noise 
ratio was also measured relative to the peak modulation 
level {± 50 kHz deviation) using the subcarrier amplitude 
recommended for existing links and a comparision was 
made between the v.s.b. and full bandwidth systems as 
above. 

Since the video circuits of the commercial f.m. 
modulator were a.c. - coupled, the effects of clamping 
the video input waveform could not be investigated 
directly. The effect of a black-level clamp was simulated 
by adjusting the carrier rest frequency for each video test 
waveform, such that black-level always corresponded to a 
given frequency. Graphs of differential gain and phase 
distortions as a function of this frequency, are given in 
Fig. 9. 



5. Discussion of results 



5.1 Differential gain and phase distortion 

Results for the measurements of differential gain and 
differential phase distortion of the colour subcarrier are 
given in Table 1. It is evident from these results that 
group-delay correction of the i.f. filter even to the limited 
extent employed in these experiments produces a worth- 
while reduction in differential phase distortion. It is 
possible that a more elaborate i.f. group-delay equaliser, 
giving a greater degree of correction, might be justified, 
but this was not pursued. 

A deviation of 4 MHz peak-to-peak (i.e. 4 MHz 
peak-to-peak for a 1 volt composite video signal after 
CCIR pre-emphasisM was adopted for subsequent tests. 
The worst-case distortions, at this deviation, were 4% for 
differential gain and 372° for differential phase, using 
partial group-delay correction of the i.f. filter. 



TABLE 1 



Differential gain and ptiase distortion as a function of deviation 



pk-pk 


With Group Delay Corrector 


Without Group Delay Corrector 














Deviation 


CCIR Bar off 


CCIR Baron 


Staircase 


CCIR Bar off 


CCIR Baron 


Staircase 


8 MHz 


15% 4° 


7% 6° 


5% 13° 


20% 12° 


10% 10° 


3% 9° 


6 MHz 


8% 372° 


.2% 5° 


2% 8° 


7% 10° 


4% 7° 


5% 4° 


4 MHz 


4% 2° 


2% , 172° 


- 3%° 


3% 5° 


1% 3° 


2% 272° 


2 MHz 


- V° 


1% 1° 


- 2° 


1% 272° 


172% 172° 


2% 1° 


1 MHz 


- - 


- 72° 


- %° 


1° 


- 172° 


- 72° 



(EL-138) 



,^ 



Fig. 10 - Staircase waveform at the output of tfie v.s.b. Iinl< 



Since the levels of non-linear distortion are dependent 
upon the instantaneous level of the input signal, black-level 
clamping should allow the best differential gain and phase 
performance to be achieved independently of signal content. 
This was investigated using the simulated clamping technique 
mentioned in Section 4 and the results given in Fig. 9 shovv 
that clamping the input waveform so that black-level 
corresponds to 70.15 MHz should give differential gain and 
phase distortions of VAX and V/2° respectively for all three 
test signals. 

The non-linearity test waveform obtained at the out- 
put of the system is shown in Fig. 10. 

5.2 Tests with augmented pulse-and-bar signal 

Tests with the augumented pulse-and-bar gave the wave- 
forms shown in Fig. 1 1 for the 2T pulse, and that in Fig. 12 
for the 1 0T pulse at the output of the f.m. system. The 2T 
pulse/bar ratio was 96% for the positive pulse and 100% for 
the negative pulse. Chrominance/luminance gain and delay 
inequality were reduced to negligible proportions by the 
video equalisation and phase correction, as can be seen from 
the IDT pulse shown in Fig. 12. 




Fig. 12 - 10T pulse response of v.s.b. Iinl<. 



5.3. Noise performance 

The results of the tests with added noise are shown in 
Fig. 13. Both of the graphs show the threshold character- 
istic of f.m. systems which occurs at a carrier-to-noise ratio 
of about 10 dB. 

The v.s.b. system appears to have a signal-to-noise ratio 
about 4 dB lower at its output than the d.s.b. system 
operating with the same carrier-to-noise ratio at the limiter 
input. It should be appreciated, however, that, for the 
same field-strength, the v.s.b. receiver will have a 3 dB 
higher carrier-to-noise ratio at its limiter input than a d.s.b. 
receiver because of its lower i.f. bandwidth. Thus the v.s.b. 
system should give a signal-to-noise ratio only 1 dB less 
than that of the d.s.b. system when operating at the same 
field strength. 



CO 50 


I 


1 1 1 

d.s.b^ 


1 
/v.s.b. 


-^ 


■o 










o 




















o 










m40 


- 






- 


Ul 










o 










c 










^30 


- 






- 


L- 










o 




















g?.o 


/ / 






- 


en 










(/) 










j^ 










S -10 


If 






- 


Q. 




1 1 1 


1 





Fig. 11 - Positive 2T pulse response of v.s.b. link. 



10 20 30 40 50 60 

carrier-to-noise ratio, dB at input to limiter 

Fig. 13 - Unweighted signal-to-noise ratio as a function of 
carrier-to-noise ratio. 



(EL-138) 



- 6 - 



TABLE 2 

Subjective grades for pictures received over botti ttie VSB 

and tfie DSB systems witti various sound subcarrier levels. 

(6-point impairment scale) 



p-p sound 


VSB 


DSB 


subcarrier 


system 


system 


level 






+ 2dB 


4 


472 


- 2dB 


3 


3 


- 8dB 


2 


1 


- 14dB 


1 


1 



* dB relative to 1v video signal (normal d.s.b. value — 14 dB) 

5.4 The sound channel 

Table 2 shows the impairment grades obtained in a 
brief subjective test using two observers judging pictures 
according to a 6-point scale*, when the 6 MHz f.m. sound 
subcarrier (unmodulated) was added to the video input 
signal. It can be seen that when the subcarrier was added at 
the same level as the 7.5 MHz subcarrier used in existing 
links, (i.e. a peak-to-peak amplitude 14 dB below 1 volt 
peak-to-peak of video), there was no noticeable impairment 
to the picture. The v.s.b. system is, of course, more 
susceptible to intermodulation distortion, illustrated by the 
subjective grades obtained for higher sound-subcarrier levels. 

The sound channel signal-to-noise ratio obtained when 
using the v.s.b. system, with the sound subcarrier at the 
same amplitude as for d.s.b., was 8 dB worse than that 
obtained with the full bandwidth system. This increase in 
the noise appeared to be caused by interference to the 
sound channel by the vision signal. 

6. Conclusions 

An experimental study of a simulated narrow-band f.m. 
television link has shown that it should be quite feasible to 
send f.m. television signals in a bandwidth of only 8 MHz, 
instead of the 16 MHz used at present. Linearity of 
the narrow-band link is likely to be slightly worse than 
that of the wideband link. The experimental system in 
which clamping of the video signal was simulated introduced 
172% differential gain and 1^/2° differential phase distortion, 
whereas the wide-band system has very low differential 
gain (less than 1%) and 72° differential phase distortion. 
The results obtained suggest, however, that a more thorough 
correction of the group-delay response of the i.f. filter 
might improve linearity. 

The performance of the narrow-band system under 
noisy conditions appeared to be almost as good as that of 



' The 6-point impairment ^a\e used was tliat given in CCIR Report 
405-1 (New Delhi, 1970) and is as follows: 1. Imperceptible; 
2. Just perceptible; 3. Definitely perceptible but not disturbing; 
4. Somewhat objectionable; 5. Definitely objectionable; 
6. Unusable. 



the wideband f.m. system when operating with 4 MHz 
peak-to-peak deviation (the value normally used in practice). 

It was found possible to transmit the normal sound 
subcarrier in the vision channel, although its frequency had 
to be lowered to 6 MHz because of the reduced bandwidth 
of the v.s.b. system. Under these circumstances, the sound 
channel signal-to-noise ratio was found to be reduced by 
8 dB. Moreover, the margin of protection against visible 
interference to the picture from the sound signal was 
reduced by about 6 dB. 

It seems likely that the principal problem, increasing 
over the next few years, will be that of interference from 
u.h.f. television broadcasts into the link; interference to 
broadcast reception by the link is likely to be less serious. 
One possibility which could be considered therefore, would 
be to use a conventional, wide-band f.m. transmitter for the 
link but to retain the option of using wide-band or v.s.b. 
narrow-band receiver techniques. A receiver with optional 
wide or narrow bandwidth i.f. filters, plus the appropriate 
video equaliser for use with the narrow-band i.f. filters, could 
then be used in either mode depending on the interference 
being received from Band V broadcast transmitters. Other 
reports in this series give further consideration to this 
possibility. 

Nevertheless, the use of a wide-band transmission 
would depend on there being no likelihood of the link 
transmitter causing interference to broadcast reception. 
The present link transmitters give an output power in the 
region of 2 watts which seems unlikely to create a very 
extensive interference field. Should future developments 
lead to links being provided with greater transmitter powers, 
however, it is possible that the link transmitter bandwidth 
will have to be restricted. The necessary filter for this 
would have to be placed after any non-linear amplifier and 
this might mean that a different filter would be needed for 
each u.h.f. channel; there could therefore be some incon- 
venience, together with a significant increase in cost, if 
this option were to be provided. 

7. References 

1. CCIR: Recommendations 405-1 Xllth Plenary 
Assembly (New Delhi, 1970). Vol. VI, Part 1, 
p. 148, Curve B (625-lines). 

2. GILCHRIST, N.H.C., 1977. Narrow-band f.m. 
system for television links: test of performance 
under conditions of multipath propagation. BBC 

Research Department Report No. 1978/22. 

3. GILCHRIST, N.H.C, LYNER, A.G, 1977. Narrow- 
band f.m. system for television links: interference 
between f.m. and a.m. television signals. BBC 

Research Department Report No. 1978/21. 

4. LAVEN, P.A., CORNELL, D.R., 1977. Narrow- 
band f.m. system for television links. BBC 
Research Department Report in course of prepara- 
tion. 



(EL-138) 



-7 



Appendix 



Using the rotating vector approach of Fig. 14, it can be 
shown that the effect of shaping the f.m. spectrum is to 
reduce the amplitude of the high-frequency components of 
the demodulated waveform. Low-frequency components 
of the modulating waveform are transmitted and received 
as double-sideband signals and are therefore unaltered. 
The higher-frequency components, however, are converted 
into single-sideband signals by the spectrum shaping filter. 

For high modulating-frequencies, the f.m. signal can 
be represented approximately by a vector rotating at the 
carrier rest frequency (co^) and two sideband vectors 
rotating at {co^ ± co^) as shown in Fig. 4(a). The higher- 
order sidebands are of relatively small amplitude, and have 
been omitted for simplicity. The phase deviation (<j)) is 
proportional to the level of the modulating signal. The 
effect of shaping the spectrum of a signal carrying the 



higher modulating-frequencies using the filter characteristic 
shown in Fig. 3 is to remove the vector at (w^ + co^^), 
thereby reducing the phase deviation and, since the resultant 
R is no longer of constant length, to introduce some 
amplitude modulating (see Fig. 14(b)). Non-linear distortion 
of the phase deviation characteristic is also introduced. 
Subsequent limiting removes the amplitude modulation, 
effectively by redistributing the sideband power in such a 
way as to give a constant amplitude (see Fig. 14(c)), but 
does not alter the phase deviation {(p') of the v.s.b. signal. 
The new sidebands are similar, but not identical, to the 
sidebands present in the original double-sideband f.m. 
signal, and are at a lower amplitude. 

Thus the change from d.s.b. to v.s.b. operation with 
increasing modulation frequency gives the sloping amplitude 
characteristic of the unequalised v.s.b. link shown in Fig. 7. 




resultant 
length = /?) 



sideband 
amplitude = a 




ojo-CiJm 



resultant 
(length ?!/?) 




ultant 
gth ^/?) 



e = 



CoJ 



(b) 



(c) 



Fig. 14 

(a) Vector diagram for f.m. with a small modulation index. 

(b) Vector diagram of an f.m. signal with one sideband removed. 

(c) Vector diagram of the single sideband signal after limiting. 



CS/VY 
(EL-138) 



-8- 



Printed by BBC RESEARCH DEPARTMENT, Kingswood Warren, Tadworth, Surrey, KT20 6NP