..

RESEARCH DEPARTMENT

DETERMINATION OF THE EFFECTIVE MODULATION DEPTH OF MONOPHONfC

PROGRAMME TRANSMITTED ON THE ZENITH-G.E. STEREOPHONIC SYSTEM

Report No. l-050

( 11162/ 13)

THE BRITISH BROADCASTING CORPORATION

ENGINEERING DIVISION

RESEARCH DEPARTMENT

DETERMINATION OF THE EFFECTIVE MODULATION DEPTH OF MONOPHONIC

PROGRAMME TRANSMITTED ON THE ZE~ITH-G.E. STEREOPHONIC SYSTEM

Report No. L-050

( 1962/13 )

;l"Y9-c~~tWL '_ . -D. E. L. Shorter, B. Sc. (Eng.), A.M. I. E. E. (W. Proctor Wi 1 son)

Th1s Report 1. the propert7 or the British Broadcasting Corporation and .&1 Dot be reproduced 1n any torm without the written permission ot the Corporation.

Report No . 1-050

DETERMINATION OFTHE EFFECTIVE MODULATION DEPTH OF MONOPHONIC

PROGRAMME TRANSMITTED ON THE ZENITH - G.E. STEREOPHONIC SYSTEM

Section Ti tle Page

SUMMARY • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• 1

1 GENERAL................................ 1

2 METHOD OF MEASUREMENT • • • • • • • • • • • • • • • • • • • • • • • •• 3

3 PROGRAMME MATERIAL • • • • • • • • • • • • • • • • • • • • • • • • •• 3

4 RESITLTS................................ 4

5 SITBJECTIVE CHECK BY PROGRAMME METER • • • • • • • • • • • • • • • • •• 4

6 CONCLUSIONS.............................. 7

ACKNOWLEDGEMENTS • • • • • • • • • • • • • • • • • • • • • • • • • •• 8

8 REFEREN CES • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• 8

TABLE 1 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• 9

May, 1962 Report No. L-050

( 1962/13)

DETERMINATION OFTHE EFFECTIVE MODULATION DEPTH OF MONOPHONIC

PROGRAMME TRANSMITTED ON THE ZENITH-G.E.STEREOPHONIC SYSTEM

SUMMARY

In the Zenith-G.E. stereophonic broadcasting system, a small proportion of the available modulation capacity is taken up by a pilot signal used in the demodula­tion process at the receiver. The remainder is divided between the sum of the left­and right-hand stereophonic signals, transmitted in the same way as a monophonic programme, and the difference between these signals, transmitted by modulation of a subcarrier. An investigation has been carried out to determine how far the level of the sum signal is lowered by the presence of the difference signal. Measurements made with an audio-frequency circuit designed to simulate the combination of the sum and difference signals in the transmitter yield a figure which, for the commonest types of programme, varies from 2 dB to 4 dB; in calculating the signal received by the monophonic listener, this figure must be added to the losses from other causes.

1. GENERAL

In the Zenith-G.E. stereophonic broadcasting system, the signal which frequency-modulates the carrier consists of three* components:

(1) An audio-frequency signal equal to the sum of the left- and right-hand stereophonic signals.

(2) A suppressed sub carrier amplitude-modulated with the difference between the left- and right-hand stereophonic signals.

(3) A pilot signal of half the subcarrier frequency, utilised in a stereo­phonic receiver to reconstitute the subcarrier.

The present investigation is concerned only with the utilisation of the available carrier deviation remaining after allowance has been made for component (3), i.e. with the deviation resulting from components (1) and (2) taken together.

Let A and B be the instantaneous values of the left- and right-hand signals respectively. Then component (1) of the transmitter input consists of an audio­frequency signal of instantaneous amplitude A + B, while the envelope amplitude of the subcarrier sideband signal which constitutes component (2) of the transmitter input is A-B.

*A fourth component in the form of a frequency-modulated subearrier cODTeylng a ftstore-casting" programme may also be providedD

2

Fig. 1 shows the way in which these modulation components are combined. If

D is the instantaneous sum of the audio-frequency component (1) and the envelope

ampli tude of component (2), both taken in the same sense, then

whatever the signs or relative magnitudes of A and B. FUrther,

if D = 2A

while if D = 2B

It may be noted that the last result could also have been obtained by a

process of time division multiplex transmission, switching to signals 2A and 2 B alter­

nately;l the data obtained in the present investigation are therefore applicable, to

systems of thi s type.

If, as is usual, the incoming programme is controlled so that nei ther A nor

B exceeds a prescribed maximum figure, the maximum value of D is automatic ally

restricted to twice this figure and the transmitter is lined up accordingly.

The audio-frequency output of a monophonic receiver is unaffected by the

presence of the subcarrier signal and is proportional only to A + B. Since, however,

A + B is in general less than D, the monophonic receiver yi elds a lower audio­

frequency output for the same total modulation than could have been obtained for the

same peak excursion D if the A - B signal had been absent. In the extreme case where

ENVELOPE OF SUPERIMPOSED SUB -CARRIER SIGNAL

COMPONENT (2)

~ / \

/ \

/ n~-' ! MAIN A.F. MODULATION / 1 A '[ 8 ~

COMPONENT (Il / ---I. /

I I /

\ / \~IA+8 :, \ I D= 2A I ,_

~ \~~IIIBI~ " \

'/ \+!

TIt.1~_

"

-~A t 8) 1

1 D=28

• (IBI> IAI) A+8

Fig. I - Addition of audio-frequency modulation A + B and sub-carrier signal of envelope ampl itude A - B to give overall modulation D

3

one of the stereophonic signals, for example B, is zero, this deficit amounts to 6 dB, for any instant when the signal 2A is taking up the full available modulation depth at the transmitter, the audio-frequency output from a monophonic receiver corresponds to a signal of A; only in the special case when A = B at the peaks of modulation is the loss to the monophonic listener zero. The loss which occurs on programme depends on the degree of correlation between the left- and right-hand signals and in particular the relationship between them when modulation is a maximum; this report describes an attempt to determine the figure experimentally for various types of broadcast material. It should be noted that this loss is in addition to the reduction in modulation -usually given as 1 dB - due to the presence of the pilot signal.

2. METHOD OF MEASUREMENT

A circuit was constructed in which voltages proportional to lA + BI and lA - BI were produced by full-wave rectification, without smoothing, of the outputs of a sum-and-difference network. Provision was made for applying either the I A + BI voltage alone or the total voltage I DI = lA + BI + I A - BI to the rectifier of a conventional peak programme-meter circuit, so that the crest values of D and of A + B could be separately studied and their ratio determined. Pre-emphasis of 50 ~s was applied to both A and B signals entering the system.

The overall accuracy of the measuring system was investigated by applying tone from a single source to one or both pairs of input terminals and a check was also made with two equal but uncorrelated white noise sources; the maximum error in the determination of the ratio of D to A + B was within ± ~ dB.

3. PROGRAMME MATERIAL

Most of the test material was taken from recordings made in the course of B.B.C. stereophonic experiments but a few items appearing on an E.M.I. demonstration recording were also included. All but two of the test passages were recorded with coincident microphones only. The exceptions were two items in which alternative recordings had been made simultaneously, one with coincident microphones and the other wi th microphones spaced 10 ft (3 m) apart; the latter arrangement produces the "hole­in-the-middle" type of stereophony which is favoured by some recording organisations.

For operational reasons, the level of programme signals delivered to a transmitter has no clearly defined upper limit and cannot be specified except in terms of the frequency with which, or the proportion of time for which, a stated amp:.i tude is exceeded; the required information can be obtained in the former case by counting peaks and in the latter, by a process of integration. For the present purpose, the latter alternative was adopted and an integrating device,2 consisting of a series of counters operated by thyratrons, was employed. This equipment was operated by the anode current of the peak programme meter previously referred to, the counters indica­ting the proportion of the programme time for which each of a number of predetermined current values had been exceeded. From the counter readings, a curve showing the cumulative probability distribution of the programme meter reading throughout the excerpt was plotted. Each test item was played twice in succession to obtain corres­ponding distribution curves for the D signal and the A + B signal respectively.

4

40 RESULTS

Fig. 2 shows the distribution curves obtained. '!he ordinates are programme meter readings plotted to a d~oibel soale with an arbitrary 3ero in eaoh oase but for brevity will be referred to loosely as "levels"; the absoissae give the peroentage time for whioh a gi verf level was exoeeded.

In general, the distribution curve for A + B (dashed line) lies, as expeoted, at a lower level than the oorresponding distribution ourve for D (full line). In Test 1, for example, the levels exo.eeded by signal D and signal A + B for 1% of the time are respeoti vely +22 ' 8 dB and +19' 6 dB. Suppose that a stereophonio transmitter is modulated with this programme and is so lined up that the level of +22·8 dB repre­

sents the nominal maximum amplitude for signal D; It is assumed that for the 1% of the time that this level is exoeeded, the resulting overload is tolerated or al ternati vely, that a limi ter is used to prevent the excess signal from reaching the modulator. If now the A - B signal were removed, thus giving monophonic transmission only, the A + B

signal input to the transmitter (and henoe the output from a monophonio reoei'Ver) could be inoreased by 22·8 - 19<6, or 3' 2 dB, without the nominal maximum modulation being exoeeded for more than 1% of the time. Where the di stri bu tion curves for D and A + B are not parallel, this differenoe between the levels of the A + B signal obtainable wi th stereophoni 0 and monophonic transmi ssion respecti vely depends to some extent on the peroentage of the time for which it is desired to achieve full modulation. Table 1 shows, for each programme excerpt, the separation between the two distribution curves at the levels which are exceeded for 1% of the time. For other percentage figures, the separation oan be read from the curves.

In general, it will be seen by the curve spacing that the values of Il/(A + B) obtained differ, as might b e expected, aooording to the spatial distribution of the sound souroes, the ratio approaching unity in the case of a single instrument in the centre of the stage (Test 4) but being greater when a number of instruments are spread out across the stage, especially when performers on opposite sides of the stage do not pl~ simultaneously (Tests 2 and 6).

In the case of the play "Sherlock Holmes" (Test 3) the D curve diverges from the A + B cmrve somewhat abruptly at the highest levels, suggesting that the most powerful sound SOurces are those furthest from the centre of the stage; this situation is not surpri sing sinoe, for example, the characters only shout at one another when th~ are on opposite sides of the scene. A similar connection between level and position is to be seen in the "Mastersingers" overture (Test 7) in which the loudest passages are those involving percussion instruments on the extreme right of the stage.

The difference between the results obtained with coincident and spaced microphones, exemplified in Tests 9a, 9b, lOa and lOb, is in the direction to be expeoted, sinoe the degree o f correlation between the A and B signals associated with a given sound source will in general be reduoed by the time delays resulting from the wide miorophone spaoing.

5. SUBJECTIVE CHECK BY PROGRAMME METER

As a check on the foregoing resuLts, arrangements were made to observe the

5

+28 +28

+24 +24

+20

0 Cl: W N + ,.

1< '" Cl: !:: +12 .. a:

+22·8 V"? TEST 1 )- f-

V V + 19·6 1--:"'" V' ...... .., ..,

V ,/

//'

V V /

+20

0 Cl:

~ +,.

1< '" Cl: !:: +12 .. a:

TEST 2 /V

V""

V //

..... /

/ / / /

V / .. .. ' ~ +B / /"

..

.; +8

V /"

+4

0,.

+28

+24

+20

0 ffi N +16

1< '" Cl: !:: .. Cl:

'" .. '

0 Cl:

~ > Cl: .. a: !:: .. a:

'" .. ' ~

+12

+8

+4

o ..

+28

+24

+20

+16

+12

+8

+4

~/ +4

V / : / ' .-

/ V /

05 90 80 70 60 50 40 30 20 10 2 I 05 0-20·' 0-05 0-01 ° os 05 90 80 70 60 50 40 30 2e 10 5 2 I 0·5 0-2 (}I 0-05 0,01

TIME FOR WHICH LEVEL SHOWN IS EXCEEDED,ok TIME FOR WHICH LEVEL SHOWN IS EXCEEDED,ok

+28

/ / .- +24

/ / TEST3 I

V / /

/ l /

V // V ~/

/ /

I/V V/ /

+20

0 0:

~ +,. > 0: .: 0: !:: +12 .. 0: .:

i +8

+4

-- --TEST 4 .....:;:; ~ r::-.....--

/-:; Vi--' /'

V/ /V

~ (,I

V ~/ /

V Vv V

05 90 80 70 60 50 40 30 20 10 5 2 I 0-5 0-2 0-' 0-05 0· 01

V

0 .. eo 70 60 50 40 30 20 10 95 90

TIME FOR WHICH LEVEL SHOWN IS EXCEEDED,°k. TIME FOR WHI CH LEVEL SHOWN IS EXCEEDED,°k,.

+28

+24

TEST 5 I-' 1-'--- +20

......---TEST 6 V V

............ 1-'-' ......

// /V / ...

......

/' ,..-/ /'

/ /

~ ~ +16 1< .: a: ....

+12 iii a:

'" .. " ~

+B

V/ "" ...-'

V/

V V //

,/

............

V V l)/

./'

./

/ +4

V ....

95 90 80 70 60 50 40 30 20 10 5 2 I 01 0'1 0-' 0-05 O·(M os 90 80 70 60 50 40 30 20 10 5 2 I O-S c>z CH o-os 0·'"

Tlt.IIE FOR WHiCH LEVEL SHOWN IS EXCEEDED.% TIME FOR WHICH LEVEL SHOWN IS EXCEEDED,"

Fig. 2 - Level distribution curves showing relative levels of D and A + B (Tests I to 6)

D

A + B ------

6

o 0:

'"

+28

+24

+20

..... + 16

1< .. ~ + 12 0: .. ~ ... +8

+. /V

TEST 7

Vy ~ ....

V;~'

I ~V ;'

y~

1-'/ --v

V V .--...-V ...-...-,

, 98 05 OD 80 70 6()S04D 30 20 10 2 I OS 02 CO{ 005 001

+26

+24

+20

0 0:

::! + ,.

1< .. 0: ... +12 iD 0: .. ~ +8

+4

,

+28

+20

:i '" N +16

1< .. 0:

~ +12 0: .. ~ .. +8

TIME FOR WHICH L EVEL SHOWN IS EXCEEDED,°k,

,....-r-- r--TEST 9a

~ I-' .-

// V ,,' L,,/

r--f' /

V .... " .......

V V

-~" .... "' .......

98 9S 90 80 7060 50 40 30 20 10 5 2 I OS 0201005 001

TIME FO R WHICH L EVEL SHOWN 15 EXCEEDED,%

1,..-:1-

TEST lOa V V

vV' /

./ I-" ..... r-

vV " ......- ) .....

........V V VV

/

/ V ....

'. PS 90 80 706050 40 lO 20 10 5 2 I OS 0201 0-05.00 . , TIME FOR WHICH LEVEL SHOWN IS EXCEEDED .'-

+ 28

~

+24 f-

V r- -+20

0 0:

~ +,.

V I-" 1-'-

TEST 8

lL"'" ,I--'

/ / ..... LL

1< .. 0: VI t: +12 ID 0: .. ID' ~

+9 1//

+4 '/ / /

os OD 80 10 60 50 40 30 20 10 2 I 05 02 o-r 0-05 0 Of

TIME FOR WHICH LEVEL SHOWN IS [XCEEDEO.%

+ 28

+24

+20

0 0:

::! +,. >-0: .. 0: t: +12 ID 0: .. i

+9

V r-f--TEST 9b V

/ ...- ~--V V /f' //

V V //

V v ~.Y V vV

........ I--~ .-

+4

, 98 OS 90 80 70 6() so 40 30 20 IQ 2 I 0·5 0·2 0-' 0-05 001

TIME FOR WHICH LEVEL SHOWN IS EXCEED£O .%

+28

+24 l-t-

+20

V V TEST lOb

V ' -,-0 0:

~ +,.

/V , V ,'"

1< .. a: I- +12 iii a: .. ID' ~

+9

V V v/

......-

V V V

./

~/

+4

os 00 80 70 60 so 4030 20 IQ 2 I o-s O'ZI>IH5 0-01

TIME FOR WHICH LEVEL SHOWN IS EXCEEDED .\,

Fig. 2 - Level distribution curves showing relative levels of D and A + B (Tests 7 to lOb)

D

A+B------

7

level of the A + B signal from some of the test items after the programme had been hand-controlled by an experienced Studio Manager using successively:

(a) a double programme meter, with the two pointers indicating respectively

the quantities 2A and 2B

(b) a single programme meter indicating the quantity A + B.

Condition (a) represents the normal method of controlling a stereophonic transmission

on the Zenith-G.E. system and is also the method of control employed in the recording of stereophonic programmes. Condi tion , b) is the procedure which would be adopted if

the A + B signal were used to provide monophonic transmission only. Provision was made to record the A + 3 signal from the controlled programme in the two conditions

for subsequent subjective comparison. This procedure proved to be unnecessary

however; the required setting of the control fader was found to be constant throughout the excerpt, not only in condition (a), a result which was to be expected since the

recorded programme had already been controlled in a similar manner, but also constant­although at a different value - in condition (b).* To find the amount by which the

level of the A + B signal could be raised if monophonic programme only were to be transmitted, it was sufficient therefore to subtract the values of attenuation in the

control attenuator in the two cases. The resulting figures are shown in the last column of Table 1 under the heading "Studio Manager's estimate"; their precision is

necessarily limited by the finite steps - 1 dB - in the control attenuator. The test was extended to include an additional orchestral item - Tests lla and llb in

Table 1 - not covered by the level distribution measurements; two recordings, made

under widely differing acoustic conditions, were utilised.

It will be seen that the figures given under "Studio Manager's estimate"

agree, within the limits of experimental error already indicated, with the correspond­ing figures in the preceding column. For the B.B.C. orchestra in Tests lla and 11b

the results obtained by the Studio Manager were similar to those for the E.M.I. orchestra in Test 8.

6. CONCLUSIONS

In the Zenith-G. E. system of sterophonic broadcasting, the audio- frequency

signal at the output of a monophonic receiver is reduced - as in other multiplex stereophonic systems - below the level obtainable in monophonic transmission. In

assessing this loss in signal, allowance has to be made, not only for the presence

of the pilot signal, but for the sidebands of the subcarrier which transmits the

difference between the left- and right-hand stereophonic signals. The loss due to the pilot signal is usually given as 1 dB. The additional loss due to the presence

of the subcarrier sidebands varies according to the relative levels of the sum and difference signals; in an extreme case it could amount to 6 dB but for the commonest

*A slight sImplification has been made here o Had the test items been intended tor use in a monophonic programme only, it Is possible that a few solo passages might for aesthetic reasons have been slightly raised in level; this Is a matter for the discretion of the individual Studio Manager. However, none of the passages concerned would have modulated the transm1tter fully, So that as far as the general level of programme Is coneerned the conclusions to be drawn from the present experiment would not be affeoted by such a change.

B

types of programme is likely to be from 2 dB to 4 dB. It should be noted that in oompari son with multiplex systems of the Crosby type, in whioh the loss of signal to the monophonio listener is a fixed quantity, the Zenith-G.E. system possesses the advantage that the level of the monophonio signal at any time is as high as the na.ture of the programme will penni t.

'7. ACKNOWLEDGEMENTS

The experiments desoribed in this report were carried out by Messrs. H.D. Harwood, W.!. Manson and A.A. Harris, with the assistanoe of Mr. R. S.C. Gundry of O. & M. (Sound).

8. REFERENCES

BRR

1. "A Summary of the Present Posi tion of Stereop):lOnio Broadoasting", Research Department Report No. A-051, Serial No. 1959 / 14.

2. "Peak Counting and Integra.ting Circuits for the Statistioal Analysis of Programme Levels", Researoh Department Technioal Memorandum No. L.1oo4, 195).

TABLE 1

Test .

No. Descri ption

1 Monologu e by actor walking about.

2 Scene from play "The Bullet". Two

characters at respectively three­

cparters left and three-q1larters right.

3 Scene from play "Sherlock Holmes".

Two to six characters ; considerable

movement.

4 Pi ano Solo. Sound source extending

from half-left to half-right.

5 Brass band. Soloist on right alter-

nating wi th accompanim ent on l eft .

6 Dance band. Alternate beats taken

by instruments on left and right.

7 Symphony orchestra.

overture.

8 Symphony orchestra.

the Si rdar",

"Mastersingers"

"Procession of

9a Male chorus "Here's to the Maiden".

Performers distributed across stage,

soloist half-left, singing most of the

time, pi ano on extreme right,

9b Same item as 9a, but microphones

spaced 10 feet (3 m) apart.

lOa Male Chorus, "Some Folks like to sigh".

Layout as for 9a and 9b, but soloist

singing only occasionally.

lOb Same item as 10a, but microphones

spaced apart.

lla Symphony Orchestra. Excerpt,

Symphony No.4, Tschaikowsky, St.

Andrew's Hall, Glasgow. (empty)

llb Same item and same orchestra as lla,

but recorded in Glasgow Studio 1.

9

Duration 20 lOgi0 D/(A + B)

Ori gin of te st At levels Studio

exceeded for Manager's passage 1% of time estimate

E. M.1. l' 55" 3 ' 2 dB

B.B.C. 3' 20 " 4'7 dB 4 dB

B.B.C. 17 ' 00" 2 '6 dB

B. B. C. 4' 15" 1 '1 dB 2 dB

E. M. 1. 2' 20" 3 8 5 dB 3 dB

E. M. 1. 2' 35" 4'3 dB 4 dB

B.B.C. 9 ' 42" 1'7 dB

E. )A:. I. 4' 00" 2'6 dB 3 dB

B. B. C. l' 53" 2'2 dB

B. B. C. l' 53" 3'5 dB

B. B. C. l' 36" 4 '4 dB 5 dB

B. B. C. l' 36 " 3'7 dB

B.B.C. ca. 3' 3 dB

B.B.C. ca. 3' 3 dB