The Stereophonic Zoom

1

The stereophonic zoom

The stereophonic zoomBy Michael Williams

A variable dual microphonesystem for stereophonicsound recording

IntroductionIn the field of monophonic sound recording thesound recording engineer has a large range ofdifferent types of microphone from which tochoose his particular preference for recordingany specific sound source or sound environment.All manner of frequency response curves can bechosen either to modify the timbre of the soundsource, or, if desired, to reproduce it as faithfullyas possible. The directivity of the microphone andits position in relation to the source, can also bechosen to obtain, fairly easily, the desired rela-tionship of direct to reverberant sound.

Unfortunately we do not have the same flexibilityin the choice of microphone systems for stereo-phonic sound recording. The number of availablemicrophone systems and the types of micro-phones used, is relatively limited, and, almostwithout exception, those that are available havefixed characteristics. It is undeniable that each ofthese systems may be optimum in one particularsituation, however none of them can be consid-ered as meeting all our needs for stereophonicsound recording.

In no way can we consider our present stereo-phonic sound recording microphone systems asenabling the listener to experience natural per-ception of the original sound source. The soundimage as perceived by the listener is conditionedby the choices imposed by the sound recordingengineer. Balance, sound perspective, localisa-tion, etc are determined by the sound engineerand his skill in transmitting these parameters tothe listener will remain part of the art of soundrecording for many years to come.

It is unfortunately too often the case that the po-sition of the microphone system is a compromisebetween a good stereophonic sound image andthe optimum ratio of direct to reverberant sound.In more general terms, it is evident that the dif-ferent parameters that must be taken into ac-count during the process of sound recording areinfinitely variable, each situation encountered be-ing in itself unique. The conditions that are nec-essary to achieve optimum results in a specificcontext with the present "tools of the trade" arebut rarely encountered.

Even in the face of this very limited choice, manyattempts have been made to establish which, ofthe different systems available, could be consid-ered as having the best overall performance.The futility of this type of approach becomes im-mediately evident in the light of the fact that eachspecific system has a unique set of characteris-tics and that the sound recording sources andenvironment are infinitely variable.

On the contrary, rather than reduce the choice ofsystems, an effort must made to increase thenumber of systems available. Each sound record-ing engineer must have the largest possible se-lection of systems to choose from, in order tosolve the specific problems presented by a partic-ular recording situation and, to express his ownpersonal interpretation, as freely as possible.

The ideal is of course an infinitely variable systemcovering all situations. This document describesjust such a variable dual microphone systemcalled « The Stereophonic Zoom » and explains

2

the basic theory of dual microphone systems ingeneral and how this has been interpreted to de-velop this variable system.

The "Stereophonic Zoom" enables the soundrecording engineer to get much nearer to the op-timum result in the great majority of recordingenvironments. It uses any desired first order mi-crophone directivity with it's associated frequencyresponse curve, and enables reasonably indepen-dent control of:

• Stereophonic Recording Angle,• Angular or Geometric Distortion,• Reverberation distribution,• Early reflection localisation .

The analysis of the physics and psychoacousticsinvolved in this type of system has already beenpublished in papers given by the author to variousConventions of the Audio Engineering Society:

1984 - 75th AES Convention in London - Preprint2072,"Stereophonic Zoom ; a practical approach to de-termining the characteristics of a spaced pair ofmicrophones"

1987 - 82nd AES Convention in London -Preprint 2466,"Unified Theory of Microphone Systems forStereophonic Sound Recording"

1990 - 88th AES Convention in Montreux -Preprint 2931,"Operational limits of the Variable M/S Stereo-phonic Microphone System"

1991 - 91st AES Convention in New York -Preprint 3155,"Early Reflections and Reverberant Field Distribu-tion in Dual Microphone Stereophonic SoundRecording Systems"

The basic operational simplicity of this systemis described in this document. This should enablethe sound recording engineer quickly to use theSt Zoom System for everyday sound recordingwith a minimum of study.

The operational approach

I. 1 Characteristics of thestandard listening configurationThe basic characteristics of a stereophonic soundrecording system are determined as a function ofthe relative position of the loudspeakers in relationto the listener during reproduction. It is almost uni-versally accepted as a Standard Listening Configu-ration, that the listener must be placed at the sum-mit of an equilateral triangle, the loudspeakers be-ing positioned at each extremity of the base of thetriangle and directed towards the listener (Figure1).

Figure 1 – Standard listening configuration

It is essential to attenuate reflections from the ceil-ing, floor and walls and in addition, symmetry shouldbe maintained in relation to the shape of the listen-ing room, so that any remaining reflections affectequally the sound heard from the left and rightchannels. Only then will the conditions be adequateto hear with clarity and precision the stereophonicimage generated by the specific microphone sys-tem used during the recording session.

Further improvements can be made to the listeningenvironment by following the IEC recommendationsconcerning the specification of a standard listeningroom. Another interesting solution to the problemsin designing a listening room has been developed byBob Walker of the BBC Research Dept., and pub-lished in Audio Engineering Society Preprints 3543,3853 and 4645.


3

The intention in this document is to describea variable dual microphone system that willreproduce realistic stereophony, thereby cre-ating a good "impression" of relative acousticsize for each sound source and maintainingthe continuity of the sound environment. It isimportant again to emphasise the word"impression" in this context, as we are indeedonly concerned by impressions. For instance,in recording an orchestra, we must createthe impression of depth and localisation of in-dividual instruments, as realistically as possi-ble. The acoustic signals, received by our earsin the listening room coming from the loud-speakers, bear very little relation to theacoustic information actually received by thelistener in the concert hall. In fact, it could besaid that, any recording and reproduction sys-tem which is capable of giving this"impression" is acceptable, no matter whatmeans are used to obtain it.

Intensity Difference (or Intensity Ratio) andTime Difference information can be gener-ated by two spaced directional microphones.The Intensity Difference information gener-ated by the microphones, is a function of theposition of the sound source and the angle be-tween the axes of the directivity patterns ofthe microphones. The Time Difference infor-mation, on the other hand, is a function of theposition of the sound source and the distancebetween the microphones. To obtain IntensityDifference information only, the microphonesmust be "coincident", whilst Time Differenceinformation only, will be obtained with spacedomnidirectional microphones or parallelspaced directional microphones.

I. 3. 1 Operational characteris-tics of the variable dual micro-phone systemA Variable Dual Microphone System is basi-cally two identical directional microphonesmounted in such a way as to be able to modifyboth the distance between the microphonecapsules and the angle between the axis ofdirectivity (Figure 2).

I. 2 LocalisationWithout going into all the details described inChapter II, one can say that the localisation of asound source between the loudspeakers is ob-tained :

• by varying the intensity ratio between the twoloudspeakers,

• or by creating a time difference between them,• or by a combination of both intensity and time

difference.

If the same sound is produced by each loud-speaker at the same level and at the same instant,then one will get the impression that the sound im-age is situated in the centre between the two loud-speakers (0° in Figure 1). If on the other hand thesound produced by the right channel is louder thanthe left, then the localisation of the sound imagewill be situated somewhere to the right hand loud-speaker (between 0° and 30° in Figure 1). If on theother hand the sound intensity of both channels isthe same, but there exists a small time difference(less than a millisecond) between each channelthen a similar effect is obtained i.e. if the left chan-nel is in advance of the right then the sound is lo-calised to the left and vice versa.

It is quite commonplace to create a variation ofthe Intensity Ratio on a mixing desk, by means of asimple potentiometer, normally called a "pan pot".On the other hand it is quite exceptional to find avariable delay line associated with a mixing desk tocreate Time Difference information between leftand right channels. Both of these techniques pro-duce localised sound sources between the loud-speakers, unfortunately without any informationnecessary to produce natural acoustic size andstereophonic acoustic environment. This processis very aptly called in French "Monophonie Dirigé",which translates to "Directed Monophony" or"Positioned Monophony", as opposed to NaturalStereophony. Even the word ‘”Natural” in this con-text could be considered as a misnomer, as weare really involved in the process of producing the“impression” of a natural listening experience.


4

Angle between the microphones, whereas theindividual curves represent the various combi-nations of angle and distance needed to ob-tain a specific stereophonic recording angle(shown in circles as a ± value on the dia-grams).

The distance and angle between the microphones ischosen according to the Sterophonic Recording An-gle that is desired. Figure 3 shows the various com-binations of distance and angle as a function of theStereophonic Recording Angle. The x and y axes(abscissa and ordinate) represent the Distance and

Figure 3 – SRA diagram for cardioid microphones

From now on, this diagram, which we will call the"SRA diagram" (Stereophonic Recording Angle dia-gram), is our basic guide to choosing a specific

combination distance/angle corresponding to agiven Stereophonic Recording Angle. Certain limi-tations to our choice of distance/angle will be con-sidered a little later in this chapter.


Figure 2 – Angle and distance between two microphones

5

The Stereophonic Recording Angle is defined as :

THAT SECTOR OF THE SOUND FIELD INFRONT OF THE MICROPHONE SYSTEMWHICH WILL PRODUCE A VIRTUAL SOUNDIMAGE BETWEEN THE LOUDSPEAKERS.

Let us consider as an example, the combination :20cm / 90°

(20cm between the microphone capsules / 90° an-gle between the directivity axes)

We can read off the SRA diagram (Figure 3) thatthis combination has a S.R.A. of :

± 50°

Note : The Stereophonic Recording Angle is, by con-vention in this publication, always specified as ± anangle, '+' indicating clockwise measurement and '-'indicating anticlockwise measurement. This conven-tion is adopted in order to avoid confusion with thespecification of the angle between the microphoneswhich is specified as the total angle. In the above ex-ample this means that + 50° is measured clockwisefrom the front centre axis of the microphone sys-tem, and -50° measured anticlockwise (a totalstereophonic coverage angle of 100°).

Two other possible combinations with the sameStereophonic Recording Angle of ± 50° are :

10 cm / 130° and 30 cm / 50°

But of course there are a multitude of combinationspossible for a specific SRA. Here are two examplesof other readings from the SRA diagram :

10cm / 60° has an SRA of ± 90°40cm / 110° has an SRA of ± 30°

All sound sources within this angular sector will bereproduced as virtual sound sources BETWEEN theleft and right loudspeakers. Any sound source out-side this sector will be reproduced ON either one orthe other loudspeaker. The operational procedurefor setting up the system consists basically of mea-suring the sector occupied by the sound source andusing this as the required SRA of the microphonesystem. The combination distance/angle can thenbe read from the SRA diagram.

It should be said however, that it is not obligatoryfor the SRA to be equal to the angle occupied bythe sound source. Most sound recording engi-neers in fact prefer the SRA to be slightly largerthan the sound source sector. This is equivalentto leaving a little headroom in a picture or morecorrectly in this case " sideroom " in the soundimage. The amount of sideroom is obviously amatter of individual judgement, but is rarelymore than about 10° for a small group of musi-cians. On the other hand, in the case of a muchlarger orchestra it is quite often necessary to dothe opposite and place the limits of the SRAwithin the orchestra (negative sideroom), the leftlimit being within the first violins, the right limitwithin the double basses. This allows morespace for the individual instruments (flute, clar-inet, oboe, etc…) in the middle of the orchestra.But this is a question of individual preference,and there are as many different choices asthere are sound recording engineers!

I. 3. 2 Microphone positionHow does one determine the position of the mi-crophone in the first place? This is not a ques-tion of measurement or even following a set ofrules, but more one of individual preference.However, it is simple enough to describe thosefactors that are modified with a change in micro-phone position.

The variation of distance between the soundsource and the microphone will certainly changethe level of direct sound reaching the micro-phone, however this will be perceived as achange in the ratio of direct to reverberantsound. It is this ratio that is responsible almostentirely for our perception of depth or sound per-spective within the sound image. Each individualsound engineer will have his own subjective ap-preciation of the optimum value of this ratio inrelation to the type of recording being made. Inthe case of multiple sound sources, as for in-stance with an orchestra, the position of the mi-crophone system must also take into accountthe relative acoustic levels of individual instru-ments or sections of instruments, the aim beingto obtain a good “balance” between all the differ-ent sections of the orchestra.


6

Once the microphone position has been deter-mined, one has only to measure the angular sec-tor covered by the whole of the orchestra, decidehow much "sideroom" one wishes, and then readoff the corresponding combination of distance andangle:

Stereophonic Recording Angle = Sound Source Sector + "Sideroom"

I. 3. 3 Frequency response curveand directivityA major advantage using the « Stereo Zoom » or"Variable Dual Microphone System" is that we donot need to restrict the choice of directivity pat-terns to only cardioid microphones. In Figure 4,the SRA Diagram is shown for a specific Hypocar-dioid ("Wide Angled cardioid" or " Infra-carded") Mi-crophone, and Figure 5 shows the SRA Diagramfor a specific Hypercardioid Microphone.

Figure 4 – SRA diagram for a hypocardioid microphone (back attenuation 10 dB)

Figure 5 – SRA diagram for a hypercardioid microphone (back attenuation 10 dB)


7

It is general practice to use omnidirectional micro-phones to generate "Time Difference only"stereophony. However by reducing the angle be-tween two directional microphones to 0° and byvarying only the distance between the microphonesthe same result is obtained. The fact that the twomicrophones are parallel means that, althoughthere is attenuation due to the directivity of the mi-crophone when the sound source is to one side orthe other of the microphone axis, the attenuation isthe same on both microphones. Therefore no“Intensity Difference” information is generated. Thismeans that the SRA values can be read off any ofthe SRA diagrams as long as the angle betweenthe microphones is considered to be 0°, this is

more easily seen on Figure 4 – the SRA diagramfor a hypocardioid‹d directivity pattern. For exam-ple:

35cm produces an SRA of ± 90°41cm ------------------------- ± 70°50cm ------------------------- ± 50°

These values can be verified in Figure 6, whichrepresents the SRA Diagram for OmnidirectionalMicrophones. Variation in the angle between themicrophones will obviously not produce anychange in the Stereophonic Recording Angle of aspecific microphone system, as long as the mi-crophones have a truly omnidirectional responseover the major part of the audible spectrum.

Figure 6 – SRA diagram for omnidirectional microphones

However, there are other important factors thatmust influence our choice of directivity pattern forthe microphones. The first consideration is thefrequency response curve associated with a spe-cific directivity. Omnidirectional microphones havein general a good response in the bass frequen-cies, in the case of omnidirectional electrostaticmicrophones the frequency response can go aslow as a few Hertz. However this is not the casefor pressure gradient microphones, a figure-of-

eight microphone has a 6dB/octave roll-off in thebass frequencies. The exact frequency at whichthe roll-off starts is determined by the manufac-turer, depending upon the required microphonesensitivity and signal to noise ratio. Figure 7 showsthe variation in roll-off as one passes from omnidi-rectional through to figure-of-eight directivity. Thisseries of frequency response curves related to di-rectivity assumes that all other frequency depen-dent factors are equal, especially sensitivity.


8

Figure 7 – Frequencyresponse curve anddirectivity


Omnidirectional

Hypocardioid

Cardioid

Hypercardioid

Bidirectional (Fig-of-8)

9

It is therefore quite possible to take advantage ofthe better bass response with a hypocardioid mi-crophone as against a cardioid, or on the contrary,to attenuate bass response with a hypercardioid.This choice is generally determined by the acousticenvironment of the recording (traffic or ventilationnoise, structure born vibration, etc...), or simply byone's personal preferences. However any practicalStereophonic Recording Angle (between ± 90° and± 30°) can be obtained with any chosen directivitypattern. Other reasons for choosing different direc-tivity patterns will be considered in later chapters,for example the response of the microphone sys-tem to the surrounding acoustic environment.

I. 4 Operational limitationsIn Figures 3, 4 and 5 it would seem that there is analmost infinite choice of the different combinationsof distance and angle that produce a given Stereo-phonic Recording Angle. Although this is certainlythe case concerning the angular limits of theStereophonic Recording Angle, there are howeversome other operational limitations which must betaken into account in the final choice of a micro-phone system. Two additional major characteristicsof the reproduced stereophonic sound field mustbe considered :

• Distortion of Angular Localisation within the re-produced sound field

• Variation of the ratio of direct to reverberantsound within the reproduced sound field

A more detailed analysis of these two characteris-tics will follow in paragraphs I. 4.1 and I. 4.2.TheSRA Diagrams represented in Figures 3, 4 and 5are just the first stage in the complete graphicalrepresentation of the characteristics of differentmicrophone systems. A quick look at Figures 8, 9and 10 will show the method now used to repre-sent these two additional characteristics :• The values of Angular Distortion are repre-

sented within a square box on each curve

4 5 6 7 8 9

• the two shaded areas at the top and the bot-tom of each diagram show the combinationswhere the ratio of direct to reverberant soundvaries within the stereophonic recording an-gle. In the top shaded area there is too muchreverberation in the centre zone of a particu-lar microphone system, whereas in the bot-tom shaded area the reverberation increasestowards the extremities of the SRA.

Figure 8 – SRA diagram for hypocardioid microphones (back attenuation 10dB)showing angular distortion and reverberation limits


10


Figure 9 – SRA diagram for cardioid microphonesshowing angular distortion and reverberation limits

Figure 10 – SRA diagram for hypercardioid microphones (back attenuation 10 dB)showing angular distortion and reverberation limits

11

I. 4.1 Angular distortion -geometric distortionThe characteristic of Angular Distortion or Geo-metric Distortion describes the variation of the po-sition of each particular element of the soundsource within the reproduced sound stage. It hasnothing to do with the characteristic of elec-tronic distortion of the signal. Angular Distortionrepresents the angular linearity of reproduction,that is the relative angular position of individualsound sources within the Stereophonic Recording

Angle when reproduced in front of the listener.The characteristic of Angular Distortion, for a givendual microphone system also must not be confusedwith the Angular Compression or Expansion of thereproduced sound stage in relation to the originalsound field. The sound source normally occupies asector that is a good deal larger than the 60° repro-duced in the equilateral triangle listening situation,therefore angular compression of the reproducedsound source must exist (Figure 11).

Figure 11 – Angular compression in reproduction (but without angular distortion)


12

Figure 12 – Angular expansion in reproduction (but without angular distortion)

Angular Distortion is one of the operational limita-tions to the choice of various combinations of dis-tance and angle in the SRA diagrams. It can varyconsiderably in relation to the directivity patternthat is chosen. In general it would seem normal tochoose combinations that present a minimum ofAngular Distortion, but of course certain soundrecording situations can warrant the use ofhigher values of Angular Distortion. The soundrecording engineer must have a wide knowledge

of all the characteristics of the different systemsavailable in order to be able to choose the sys-tem that will be absolutely optimum in a given sit-uation.

On each SRA diagram specific values of AngularDistortion are shown in boxes situated alongeach SRA curve. These values should give anidea of the relative quantity of Angular Distortion.To understand the significance of a specific An-gular Distortion value we need to look a bit closer

The opposite situation (shown in Figure 12) in whichthe sound source sector is less than 60°, is the sit-

uation in which expansion of the angular propor-tions of the reproduced sound image occurs, butthis situation is hardly ever encountered.


13

at the way these values were derived. In Figure13, we have a series of sound sources A, B, C, Dand E, situated at equal distances along the arcof a circle

If no Angular Distortion exists in reproduction,

Figure 13 – Five sound sources A, B, C, D and E at regular angular spacing

then the relative positions of A, B, C, D and E will bemaintained (Figure 14).• A at the left hand loudspeaker• B midway between left loudspeaker and centre• C in the centre• D midway between right loudspeaker and centre• E at the right hand loudspeaker

Figure 14 – No angular distortion with reproduction


14

Angular Distortion will modify the reproduced posi-tions of sound sources B and D, they will be shiftedtowards each loudspeaker, as illustrated in Figure15, where a shift of 5° of B and D is shown. The

sound sources A, C and E will not be affected bythis shift and will remain in the same relative po-sition (A and E being, by definition, the limit of theSRA, C being the centre of the system).

Figure 15 – Angular Distortion of 5°

It is this Angular Shift of the sound sources B andD that is used in the SRA diagrams, to representthe maximum Angular Distortion of a given sys-tem. We can read off the SRA diagrams specificvalues of Angular Distortion for given combinationsof distance/angle from Figures 8, 9 and 10. Hereare a few examples taken from Figure 9 (SRA dia-gram for cardioid microphones):

15 cm 110° >>>>>>>>5° (SRA of ± 50°)28 cm 60° >>>>>>>>4° (SRA of ± 50°)0 cm 90° >>>>>>>>6° (SRA of ± 90°)37 cm 0° >>>>>>>>9° (SRA of ± 80°)

This characteristic of Angular Distortion is per-ceived as a compression or crushing of the ex-tremities of the reproduced sound field towards

the loudspeakers. This can also be interpreted asa stretching of the centre of the reproducedsound image, sometimes called "hole in the cen-tre". This is in fact a misnomer, as the centre ofthe sound image still exists but is distributed overa wider sector.

It should be noted that, in Figure 9 for example,the Angular Distortion is minimal towards thecentre of the diagram. In other words, for a givenSRA one obtains minimum Angular Distortion byusing a combination of Intensity Difference(Angle) and Time Difference (Distance)stereophony in about equal proportions. AngularDistortion increases by a few degrees, with sys-tems using only Intensity Difference (Distance =0cm), and even more so with systems using


15

mostly Time Difference (Angle between micro-phones = 0° or with spaced omnidirectional micro-phones).

It is also possible to see from these SRA diagrams,how the specific directivity that is chosen for the mi-crophones has an effect on the quantity of AngularDistortion at various combinations of distance/an-gle. The most extreme case being omnidirectionalmicrophones spaced at 40 cm to 50 cm, where An-gular Distortion is about 9° to 10°.

I.4.2 Variation of the ratio of directto reverberant sound within theStereophonic Recording AngleThere are two distinct zones, shown in Figures 8, 9and 10 by shaded areas, where this problem exists.In the shaded area at the top of each diagram, theratio of direct to reverberant sound decreases to-wards the CENTRE of the SRA, whereas, in theshaded area at the bottom of each diagram, this ra-tio decreases towards the EXTREMITIES of the SRA.

This effect is obviously more apparent in reverberantsurroundings, and can to some extent be neglectedin a relatively "dead" studio environment. In outsiderecording however, the reverberation level is re-placed by environmental ambience and these zonesmay again become apparent.

The exact angle, at which change in reverberationbecomes unacceptable, is again a subjective factorand can vary from one person to another. The limitsshown on the diagrams can again only be an approxi-mation.

I. 5 Recording procedureWe are now ready to put into practice the VariableDual Microphone System in a stereophonic record-ing situation. Good angular localisation of the repro-duced stereophonic image plays an important role instereophonic sound recording, it is certainly not theonly factor that must be considered, but it is one ofthe most important. In what order should one anal-yse each of the parameters of the microphone sys-tem in relation to the characteristics of the soundsource and its acoustic environment?• Frequency Response Curve of the microphones• Position of the microphone system• Localisation within the Stereophonic Sound Image

Here again, each sound recording engineer mustuse his own knowledge and judgement to adapt tocircumstances, and of course, each person mayhave individual preferences. It would also be a mis-take to try and standardise a given procedure,just as much as it would be a mistake to try andpass on one's own individual way of working - thisleads sooner or later to a complete sclerosis ofthe "art of sound recording".

The following procedure should therefore be con-sidered just as a starting point (if this is needed),in developing your own working methods:

1) Frequency Response CurveEach microphone has a typical frequency charac-teristic which determines the quality of timbrereproduction of the sound source. This character-istic, especially with respect to the bass response,must be considered in relation to the directivity ofthe microphones to be used as shown in Figure 7.

2) Position of the microphone systemHere we are looking for the optimum ratio of di-rect to reverberant sound, as well as a good bal-ance between individual elements of the soundsource. If these two factors are not compatible, itis here that perhaps the first compromise has tobe considered.

3) Stereophonic Sound ImageThe Stereophonic Recording Angle is chosen soas to take into account the sector occupied by thesound source, plus any "sideroom" that is consid-ered necessary.

4) Angular DistortionIn general, there is but one combination of dis-tance / angle for the desired SRA, if minimum An-gular Distortion is required. Otherwise the only re-striction in the choice of other combinations ofdistance / angle is that the shaded areas shouldbe avoided, unless specific effects are considerednecessary.

5) Mono/Stereo computabilityThe different combinations of distance / angleavailable, enable one to opt for Intensity Differ-ence Stereophony (having good mono/stereocompatibility) or on the other hand, to increasethe Time Difference contribution to obtain a bet-ter feeling of "space", but we all have differentideas on this subject!


16

If a specific characteristic is causing difficulty, itmay be necessary to change the order of priorityand start this sequence again. This process contin-ues until optimum results are obtained or at least,the best possible compromise.

I. 6 Measuring instrumentsThe sector, occupied by the sound source plus anysideroom that is desired, must be determined rea-sonably accurately, so that one can read off fromthe SRA diagrams, the combinations of distanceand angle for the required Stereophonic RecordingAngle. At present no measuring instrument existsthat is adapted to this type of measurement, exceptperhaps the sextant. This situation is easily recti-fied, if you are prepared to take up saw, hammerand nails, and follow a simple "Do It Yourself" plan.

The optical part of the instrument that will be de-scribed is a 180° optical viewer ("Judas hole"), nor-mally mounted as a door viewer, to check who isoutside before opening up. In France one of thefirms making this viewer is called "BLOSCOP", how-

ever, there must exist a number of firms makingthis type of device in other parts of the world. It isworth buying the luxury model with maximum di-ameter optics, in order to obtain a good clear im-age and a real 180° viewing field. You could alsouse the "Fish Eye" lens from your camera as aviewer, but this is a rather expensive solution tothe problem.

The details of the complete viewer are shown inFigure 16, and as you can see, the only materialsthat are needed, are a few long and short nails,and some offcuts of plywood. Nails have beenplaced every 10°. The long nails being placed at0°, ± 30° and ± 60°.

Measuring the sector covered by the soundsource is now a simple matter looking at thesound source through this viewer and reading offthe angle corresponding to the left and right lim-its of the sound source. Figure 17 shows such aview through the viewer - you can see why wehave nicknamed the instrument "the crocodile”

Figure 16 – “The Crocodile”


17

Figure 17 – “The crocodile’s eye view !”

But it is not yet time to leave the workshop, we stillneed to make a support for the microphone sys-tem, so that it is possible to easily adjust both dis-tance and angle between the microphones.

A metal bar in Duralumin or other similar metal ofabout 300mm long, 25mm wide and 4mm thickmust be drilled with 10mm diameter holes every

20mm. The centre hole can be drilled at 8.5mmdiameter and threaded with a 3/8" Whitworthtap if this is available, the bar can then bescrewed directly onto the top of a microphonestand. It is now a simple matter to fix two micro-phone clips at various positions along the bar, us-ing obviously short 3/8" Whitworth fixing screwsto hold the base of the microphone clips. Thissame thread is often used with photographicequipment, so see whether your local photo shopcan help with the necessary fixing screws.

Measurement of the angle between the micro-phones can be made with a large protractor. Butif you can have access to a metal turning lathe,then it would be even better to make a pair ofgraduated rotating supports for the micro-phones. The distance between the microphones ismeasured of course, with a ruler or tape mea-sure.

However there is a very simple way of setting upthe distance and angle between the microphoneswhich requires only a sheet of A4 paper and relieson the ancient Japanese art of “Origami” (but in avery much simplified form) ! Here are a few illus-trations :

Figure 18 – Microphone measurement by “origami”


18

The corner of an A4 page folded as in Figure 18bwill give us 45° and 135°, opened out flat and thenfolded again as in Figure 18d will give 67.5° and112.5°. This means we can now measure accu-rately angles of 45°, 67.5°, 90°, 112.5° and 135°,and other angles can be estimated “by eye”. An A4page (21 cm x 29.7 cm) folded vertically as in Fig-ure 18f will give us a width of 10.5 cm, openedout flat and then folded horizontally as in Figure18g will give us 14.85 cm. So we can measure10.5 cm, 14.85 cm, 21 cm and 29.7 cm. Againother distances can again be estimated.

I. 7.1 Preliminary listening testsBefore trying to analyse the Variable Dual Micro-phone System in a real recording situation, it isworthwhile carrying out a few simple listeningtests, which will help in your understanding of thebasic characteristics of the system. Record some-one talking whilst moving along the circumferenceof the circle of about 1.5 metres in radius. It is im-portant for them to indicate their position in rela-tion to the front axis of the microphone system atabout every 10°. The object of the exercise beingto try and determine that sector of the angularmovement that is reproduced as a stereophonicimage between the loudspeakers. The limit of theStereophonic Recording Angle is obviously the an-gle at which apparent movement ceases i.e. whenthe reproduced sound remains fixed at one or theother of the loudspeakers.

The average voice has a relatively limited fre-quency range (150Hz to 8kHz), so in order to testthe performance of the recording system at fre-quencies above 8kHz, one must use a differentsound source much richer in the higher frequencyrange ("maracas" or something similar), again indi-cating every 10° the position of the sound sourcein relation to the axis of the system. However itshould not come as any surprise to observe thatthe range of frequencies covered by the humanvoice will be enough to establish perfectly reliablelocalisation characteristics. In this way the majorpart of the characteristics of the microphone sys-tem can be analysed.

Here are a few examples of interesting combina-tions of distance and angle that can be tried usingcardioid microphones to start with :

a) 0 cm 90° } Coincident cardioidb) 0 cm 180° } Intensity differencec) 0 cm 130° } Stereophonyd) 0 cm 60° }

e) 30 cm 90° } Spaced cardioidf) 20 cm 70° } Intensity and time differenceg) 10 cm 60° } Stereophonyh) 35 cm 10° }

Note carefully your impressions during these lis-tening tests, with special reference to:

• Perceived Stereophonic Recording Angle• Impression of Angular Distortion• Evolution of the ratio of direct to reverberant

sound within the SRA• Precision of the reproduced sound image• Spatial impression

Not only will this type of critical listening test helpin your knowledge of the various combinationsthat can be used in the Variable Dual Micro-phone System, but it can help considerably in thebasic training of the ear to analyse the differentcharacteristics of a stereophonic recording.

I. 7.2 Experimental recording andlistening testsFor the sake of this specific type of test it is inter-esting to divide the sound sources into four cate-gories:

1. Straight line moving sources.A very common type of sound: a passing train,car, motorbike, a plane taking off or landing, foot-steps passing from one side to another, etc.....

2. Continuous straight line sources.The sound source is spread out in a straight linefrom left to right: waves on a beach.

3. Sound sources occupying an angular sectorless than 180°.This is the most common situation encounteredin everyday recording: an orchestra, a group ofpeople in conversation, a firework display at acertain distance, chickens in a field or chickencoop !!!....

4. Totally surrounding environmental sounds:This is probably the most difficult type of sound


19

to reproduce as the reproduction sound stage islimited to only 60°. Birds in a forest, noises andconversation in the middle of a market, ambiencein a railway station, etc,........

Each of these types of sound source has specificsound recording difficulties and requires a differentapproach to obtain satisfactory reproduction. Thedifferent SRA diagrams (Figures 8, 9 and 10) en-able us to experiment different possibilities of SRA.The correspondence of each combination triedmust be carefully noted in relation to each record-ing for future reference. A series of listening testscarried out after such recordings, together withdiscussion of the results, will often suggest a newseries of recordings that are necessary to find theoptimum characteristic of SRA for a particular situ-ation. Experimentation with different combinationsof distance and angle for the same SRA value, canalso help in finding the optimum solution for record-ing that particular sound source and in that specificsound source environment.

The "quality" of stereophonic reproduction can alsovary considerably according to the type of soundsource used. Musical instruments, percussion in-struments such as drums, xylophone or marimba,with little decay time, should be compared with oth-ers such as bells or gongs, with a long decay time.The amount of reverberation can also influence theperception of stereophonic "quality", or is it just thatour perception of the timbre of the instrument haschanged in this case ?

I. 8 Listening with headphonesUnfortunately, it is quite often necessary to monitora stereophonic recording using headphones, eventhough the recording will eventually be heard onloudspeakers. The characteristics of the sound im-age (SRA, Angular Distortion, etc.) heard on head-phones will differ considerably from that repro-duced by loudspeaker stereophony. There is nosimple solution to this problem, except to use theheadphones to monitor only the technical quality ofthe signal together with the ratio of direct to rever-berant sound, and rely on the angular measure-ments made with the "crocodile" as a basis onwhich to choose the Stereophonic Recording Angle.This problem with headphones is as simple to un-

derstand as it is complex to solve. When using apair of headphones, one hears only the left chan-nel with the left ear and vice versa, this is nor-mally called Binaural Stereophony. On the otherhand, when listening to loudspeaker stereophony,the left ear hears both the left and the right loud-speaker, the latter being heard with a certain at-tenuation. Similarly the right ear hears both rightand left LS signals, the left LS signal being slightlyattenuated.

This acoustic interference of the left loudspeakersource on the sound received by the right handear, and right source on the left ear, is called“Acoustic Crosstalk”. In order to obtain a satisfac-tory stereophonic sound image with loudspeak-ers, it is necessary to about double the IntensityDifference and Time Difference in relation to thequantity necessary to produce a good sound im-age with headphones. This means that a goodstereo image on headphones will be too small inthe loudspeaker base or sound stage - the soundimage will be situated towards the centre of theloudspeaker base line. Whereas a good loud-speaker stereo image, will be much too wide forheadphones - the sound will seem to be just onthe left and right ear, with little in between.

There is, however, a way of passing from one lis-tening situation to the other. It is necessary totransform the analogue audio signal to digital in-formation, and by the techniques of Digital SignalProcessing reduce the effect of crosstalk. Butthis process is at an early stage in its develop-ment and we will have to wait a few years beforeit becomes available as a standard correction tothe type of listening system used.

I. 9 Analysis of the listening testsThe original text of this document was first pub-lished in 1989 in France, as a series of articlespublished every two months for the magazinecalled "Actualité de la Scénographie". At the timewe thought it a worthwhile experiment to try andpersuade the readers to carry out the recordingand listening tests without having any informationconcerning what they were supposed to hear, theactual analysis of the listening tests following twomonths later in the next edition.


20

This version of the Stereophonic Zoom©2002 Rycote Microphone Windshields Ltdand Human Computer Interface


Our reasoning behind this method of presentationwas very simple. It is very difficult to create a neu-tral listening situation, our capacity to analyse asound image is too often influenced by intellectualconsiderations or simply by the visual environment.In other words, if we know what we want to hear orwhat we should hear, then we are quite capable ofpersuading ourselves that that is exactly what weare hearing!

This is a problem in everyday working, even withthe most experienced sound recording engineersand is even worse in the student learning situation.In his eagerness to succeed, the student will adoptany visual or intellectual clues that are available,completely missing the one and only factor that isimportant, which is his own listening experience.It is the duty of the instructor to eliminate all thevisual clues and create the conditions necessary tostimulate real listening.

This article is an extract from the first chapterof the original publication “St Zoom” by MichaelWilliams and should enable the sound recordingengineer quickly to use the St Zoom System foreveryday sound recording with a minimum ofstudy.

However it must not be forgotten that a deeperunderstanding of the operational method and itsjustification must take into account a detailedmathematical, physical, and psychoacousticalanalysis of the characteristics of the stereo-phonic microphone system. This subject is devel-

oped in the second chapter of the book “StZoom”, the third chapter being an attempt toreply to the numerous questions that ariseconcerning the basic theory involved in the de-velopment of the system, and the applicationof this system of stereophonic sound record-ing to the multitude of situations encounteredin everyday recording.

The complete document can be obtained fromMichael Williams, “Sounds of Scotland”,BP50, 94364 BRY SUR MARNECEDEX, Franceor by email: [email protected]

Michael Williams started his professional ca-reer at the BBC Television Studios in London in1960. In 1965 he moved to France to work for''Societe Audax'' in Paris developing loudspeak-ers for professional sound and television broad-casting, and later worked for the ''ConservatoireNational des Arts et Metiers'' in the adult educa-tion television service. In 1980 he became afree-lance instructor in Audio Engineering andSound Recording Practice, working for most ofthe major French national television and soundbroadcasting companies, as well as many train-ing schools and institutions. He is an activemember of the Audio Engineering Society, andhas published many papers on Stereo and Multi-channel Recording Systems over the pasttwenty years. He is at present the AES Publica-tions Sales Representative in Europe.

The Stereophonic Zoom

Documents

Transcript of The Stereophonic Zoom