Lighting techniques and associated - CiteSeerX

181

Lighting techniques and associatedequipment for outdoor colour televisionwith particular reference to football stadium lighting

I. F. DAVIES, BSe, CEng, FIEE, FInstP, FIllumES, M.G.A. JACKSON, FIllumES,and B. C. ROGERS, MIllumES, MAPLE.

The authors are with Philips Electrical Ltd. Thepaper was received in March 1, 1972 and presentedat a meeting of the Society on April 12,1972.

Summary The general requirements of visual performance and comfort instadium floodlighting are outlined; the essential differences between the problemsposed by colour television transmissions and black-and-white television arepointed out. Horizontal and vertical illuminances and uniformity are discussed,together with glare problems. Side-mounted and tower systems are consideredwith particular reference to colour television requirements. Design aspects of thelighting equipment are discussed in relation to the colour television cameraequipment. The need for computer planning is mentioned.

1 Introduction

Television transmissions commenced in Great Britairover 30 years ago and outdoor broadcasting usingthis medium has been commonplace for nearly aquarter of a century. Until recent times monochromesignals only have been available, and although muchlighting experience has been accumulated duringthese years, it has been associated with illuminancesof not much more than 500 lux, which is satisfactoryfor black-and-white transmissions.

The introduction of colour television (CTV) trans-missions has demanded increases in illuminance upto 3000 lux; levels as high as 5000 lux have beenreported.Even when using monochrome techniques it wasessential to study the requirements of the spectator,the viewer and the television signal transmittingsystem since these are intimately related. Thus thelighting requirements of outdoor CTV cannot bestudied in isolation; but must be considered in rela-tion to these three factors, which the paper considers.

2 Colour and eqpipraent

At first colour television broadcasting was restrict-ed to studio work and a small number of outdoor

events, the latter usually being televised during day-time. >

Outdoor colour television broadcasting has beenfound to require even better lighting than that for

interior use. Furthermore, due note must be takenof the following:Colorimetric aspects are important for the colourtelevision system, and must take care of the trans-mission and receiver characteristics to ensure

acceptable quality on the television screen. Elec-tronic and optical characteristics define the limita-tions of both cameras and receivers in respect ofsuch items as noise, time lag, picture content andquality.The lighting system has to ensure an acceptablescene for the spectators as well as for the colourtelevision system.When considering the colour television system themistake should not be made of assuming that itsoperation is similar to that of the human eye: whereasthe human eye can handle a luminance range of1: 10 000, the effective range for a CTV system is1: 20, which may reach 1: 40 under extremely goodconditions. a

Further, colour rendering in CTV is limited tocolours within the primary colour triangle, shown inFig. I, and excludes those between the triangle andthe CBE locus. Because of this, it is inevitable &dquo;;hatcolour and luminous shifts will occur.

The real scene colours are converted in the tele-vision receiver into those colours which would beobserved if the actual scene were illuminated bydaylight C (Fig. 1), and the white of the real scene isalways shifted in the receiver to white C no matterby what light source the actual scene is illuminated.

If the scene is illuminated by two different lightsources at the same tin~:e--~hich may happen undertwilight where there is some daylight as well as

at PENNSYLVANIA STATE UNIV on September 18, 2016lrt.sagepub.comDownloaded from

http://lrt.sagepub.com/

182

Fig. l. CIE colour with positions ofreceiver primary colours indicated. Theenclosed area (-o-. contains nearly allnatural colours. 9

electric light-then the camera can adapt to only oneof the sources and colour shifts will occur due to thecontribution of the other source.

The colour television system, like the human eye, islimited by its sensitivity, and in practice this is re-presented by the signal-to-noise ratio of the lumin-ance signal received by the camera-which com-prises different parts of the red, green and blue videosignals--and time-lag in the various camera andother electronic channels. The former, if excessive,results ln ’~no~v’ effects on the television screen, andthe latter in the appearance of coloured ’ghosts’behind moving objects in the scene. Most ghostsoccur due to time-lag in the red channel.

3 The CTV receiver

It is fortunate that practically all natural colours liewithin the primary colour triangle, although the purespectral colours and certain colours from fluores-cent materials do not.

In order to reproduce natural colours with goodquality on the television screen, the three receiverprimary colours should have intensities correspond-ing to the familiar spectral response curves shownin Fig. and thus ideally the camera should deliverto the receiver signals having the same spectralas these curves.

4 The CTV camera

41 1 PrinciplesThe energy from the scene being televised must beconverted into electronic video signals in the cameraand for this over 90% of all cameras employ ’Plum-bicon’ television camera tubes.

In order to achieve the three spectral responsecurves required, the ’Plumbicon’ characteristics,together with those of optical filters or mirrors, arethe variables which have to be controlled.

Fig.2.

(a) Ideal spectral response curves for receiverprimaries for high fidelity colour matching of allcolours within Rub triangle of 1.(b) Positive only response curves with cameraoptical system of 3 and XQ 1020 ’Plumbicon’ Icamera tubes.

(c) Luminance distribution with positive onlycurves of Fig. 2(b) compared the V(A) curve.

~ Spectral response curves of matrixedcam era.

(e) Lurleinance distributiort curves of2(d) compared with the V(Â) curve.

Fig. 3 shows the principle of a three-tube CTVcamera. MR and MB are dichroic mirrors whicheach reflect a part of the spectrum, the light of whicheventually passes into the red and blue ’Plumbicon’tubes PR and PB, and transmit another part whichpasses into the green ’Plumbicon’ tube PG’ Thespectral transmission of these three separatedchannels, combined with the sensitivity of the ’Plum-



183

Fig. 3..~’rinciple of CTV camera. Light splitinto three bearrts R, G and B by dichroic mirrors.They Pass into the IPlumbicon1 tubes. P, P, andP B and are converted into three electric signals.

Fig. 4. absolute spectral sensitivity of’Plumbicon’ tube with extended red response,compared with response of normal ’F’Lumbieorc’tube.

Fig. 5. Perceptive visual luminance (Lyei) range between border ofabsolute blackness (Luminosity B = 0) and border of ~as~c~~~o~°t glare(Luminosity B = 1) ~ ---~~ and luminance (Lrel) range onthe TV screen between black (B = 0) and highest intensity white (B = 1)(-) .

bicon’ tubes shown in Fig. 4, results in the spectralsensitivity curves of the three channels shown inFig. 2b.These are entirely positive curves and do not containthe negative elements in the ideal curves of Fig. 2a.This results in the luminance distribution shown inFig. 2c where it is compared with the ideal V(À)curve. In these circumstances, not only would colourshifts occur but also luminance deviations.

Nearly all the older CTV cameras were of thisoptical type, but recently it has been possible toapproximate to the ideal curves by adding and sub-tracting electronically from the three colour signals.Such a process is called ’linear matri~ir~g’.If the three colour signals are adjusted in accordancewith the following matrix:

B G R-0. 02R -0. 02R -~-1. 12R-0.01G -~-1. 23~ -0.16G+1.03B -~. 218 +0.04B

the sensitivity curves and luminance distributionbecome as shown in Figs 2d and 2e, respectively;these curves approximate closely to the ideal con-ditions shown in Fig> 2a. A small disadvantage ofemploying linear matrixing is the slight reduction inthe signal-to-noise ratio which occurs.

Recently a new type of ’Plumbicon’ camera tube-the XQ1023-has become available and this has im-proved the situation further, particularly in the redregion of the spectrum (see Fig. 4).

4.2 Luminance rangeWhereas the eye has a luminance range of 1: 10 00,the television receiver screen luminance range liesbetween that when the receiver is not operating anda maximum level beyond which details cannot be dis-tinguished : these extremes are in the ratio of only1 : 40. Fig. 5 expresses this graphically: for a givenluminance difference in the real scene, the visual

luminosity difference on the television screen is con-siderably greater than that for direct observation bythe human eye. In practice it is found that the tele-vision system not only has a limited luminance range,



184

but amplifies the sensation of luminance differencesby a factor of about 2.

In outside CTV broadcasting the upper level of theluminance range is normally that associated withplayers in white shirts and the lower limit with theless well lit parts of the scene--usually darklydressed spectators in the less well illuminated areas.

Fig. 6 shows the luminance range of a typical realscene as reproduced by a CTV system. In practiceit is found that in order to distinguish 50°/0 of detailin the spectator areas, the average illuminance onthe stands must be at least 20% of the average verti-cal illuminance of the whole scene. Further, it isfound that the preferred ratio between vertical illu-minance on the players and horizontal illuminanceon the field is between 1: 1 and 2 : l.

Fig. 6. Luminance range of the real scene asreproduced by the CTV system ( fronthighest intensity white (B = .1 ) (e.g. shirts ofplayers with 0. 6 reflectance) doumwards toborderline of blackness (B = 0) (e.g. details inless well illuminated parts of the stands).Dependent upon the vertical illuminance ratioEs (illuminance on sta~cds)/E~ (illuminance onthe players), greater or less details of thespectators is reproduced. The field luminancewill achieve a position in the reproducedluminance range in relation to the players

‘

luminance dependent upon the different lightingconditions indicated.

5.) Cojioor adag~anon, transfoirmat2on and shift

Apart from a scene illuminated by daylight C, thecolours of a scene must be transformed according tothe actual distribution of the light solazcc-sused, and this is carried out by varying the amplifi-cation of the three video signals derived from thered, green and blue channels of the camera.

Because of this it is not possible for the camera toadapt to different light sources at the same time, andso it is important, in order to avoid colour shiftsduring reproduction in the receiver, to ensure thatthe spectral characteristics of all the light sourcesused are the same and remain so during transmission.Furthermore, when two light sources are used atthe same time, the mixture of the light should beidentical over the whole area illuminated.

If two different light sources must be used together,then they should be chosen such that the amplificationfactors for the video signals are similar for bothtypes of source. If this is not done, then either colourshifts will occur or, where one source such as day-light is varying continuously in colour, continuous re-adjustment of the video signal amplifiers will benecessary.

6 Colour television system sensitivityColour television system sensitivity is limited bylag effects and signal-to-noise ratio of the videosignal and the camera amplifiers.

~.1 1 LagIn cameras fitted with ’Plumbicon’ tubes the lag canbe diminished considerably if the top-white signal ofeach video channel is not less than 100 nA and notmore than 300 nA; the first limit is the most critical.

Since the top-white signal intensities of the threecolour channels depend upon the light source usedfor illuminating the scene, different illuminances willbe required for different light sources to ensure thatthis minimum level is achieved in each of the threecolour channels.

Fig. 6 indicates that a top -white luminosity of unityis achieved when the average reflectance of an objectis D. 6. The average vertical illuminance required toreach this minimum signal current will vary withdifferent light sources, and Table 1 shows how this

parameter varies for different types of source. Thered channel is always the most critical; lag effectswill occur mostly in parts of the visual scene wherethere are great luminance differences, such as whena player moves against a dark background.

6.2 Signal-to-noise ratio

The signal-to-noise ratio of the video signal dependsmainly on the luminance signal which is derived froma proportion of each of the three individual coloursignals. The fractions of the three signals depend onthe degree of approximation to the ideal spectralresponse curves. For a camera with ’positive onlyresponse curves the luminance is expressed by:

L = 0.30R + 0. 59G + 0. IIB

which shows that the luminance signal depends in themain on the green channel, and thus the green noise isnearly twice that of the red noise and six times thatof the blue.

For a given signal-to-noise ratio the difference be-tween the required illi-iminarices is small when theeffects of different types of light source are com-pared.it -will be from Table 2 that the increase isilluminance necessary to fulfii signal-to-noise ratioand lag requirements is lowest for Na, Tl, In metalhalide lamps, which were designed specifically tooperate with ’Plumbicon’ camera tubes. Fig. 7 showsthe spectral energy distribution of the light fromthis type of lamp compared with the spectral responsecurves for ’positive only’ ’Plumbicon’ CTV camera.

7 ~l.a~~-uP requirementsTables I and 2 ’apply to a TV camera lens apertureof f/4, but the same picture quality could be achievedat half the recommended illuminances using an f/2. 8



185

Table 1. Average iluminance required to avoid lag effects for different light sources(aperture f/4 on camera).

Table 2. illuminance to achieve lag and signal-to-noise ratio(SNR) requirements.

Fig. 7, Relative radiant e~zer~gy for 2 klamps compared with positive spectralresponse curves of CTV cameras.

aperture, or at a quarter of these values with an f/2aperture. However, increase in the aperture reducesthe depth of focus of the optical system, and this inturn limits the close-up performance of the camera.

In practice, two types of camera optics are frequentlyemployed:

(a) Optics having a variable focal length rangingfrom 20 mm to 200 mm; the aperture is increasedautomatically with increasing focal length resulting ina constant light value on the camera tubes. e

(b) Optics with variable focal length ranging from18 mm to 180 mm with which increased focal lengthscan be obtained by using range extenders. One type ofrange extender will permit the focal length to bevaried from 27 mm to 270 mm with a loss of only oneaperture; another type permits variations from36 mm to 360 mm with the loss of a. further aperture.

Most television camera optics have a maximumaperture of f/2 and thus if the second type of rangeextender is used, the aperture will be reduced to f/4and the illuminances given in the previous table willbe required. If the first range extender is used, theaperture is reduced to only f/2. 8 and only half of therecommended illuminances will be needed.

8 Light sources, control gear and luminaries

The cheapest and most effective light source forcolour television is daylight, but as daylight fades,electric light becomes necessary.



186

Daylight varies considerably in intensity and inspectral composition. Its colour temperature canvary from 3500K to I2 U04 K or more for a clear bluenorth sky. In the evening the daylight colour tempera-ture is normally about 4000 K to 6000 K.

To keep colour television camera adjustments to aminimum, the lamps employed should have a similarcolour temperature and produce a similar cameraresponse to the daylight they replace.For similar reasons it is not advisable to use lampswith different spectral compositions to light the samescene, since it then becomes very difficult to ensurea constant spectral composition in both vertical andhorizontal planes over the whole area; even quitesmall differences in spectral composition can causeserious colour shifts in the colour television picture.However, no practical problems have been experiencedwhen lamp types having slightly different colourappearances have been used on the same installation,provided that their spectral energy distributions aresimilar and/or they produce similar CTV channelsignals.Besides blending with evening daylight, the lampsmust produce acceptable colour rendering for boththe colour television cameras and the spectators.It should be remembered that visual and colour tele-vision colour rendering have nothing to do with eachother. Because a light source has a relatively lowCIE general colour rendering index (Ra) it does notnecessarily mean that it is unsuitable for colour tele-vision work. For instance a light source may havea spectrum that is deficient in deep red radiation, butthis does not matter for colour television work be-cause the colour television camera is not sensitive tothese wavelengths. Again, a light source may not re-produce purple colours well, but this also is unim-portant because purples rarely occur in nature.

In order to keep the lighting electrical loan to a mini-mum the lamps should have a high luminous efficacy(lm/W) or, what is more important, produce a highcamera response per installed kilowatt. Because theformer is high it does not follow that the latter isalso high. This occurs only when the spectral re-sponse of the CTV camera matches the spectralenergy output of the lamp--see Fig. 7, Section 6. 2.

In addition to the properties so far mentioned thelamp should have an acceptable life. A minimum ofabout 300 hours is preferable; this covers about threeplaying seasons. If the lamp is required to light anarea such as a football stadium to CTV illuminances,then to avoid an excessive number of luminaires the

lamp should have an output of at least 150000 1m.

The source of light within a lamp--the discharge tubeor filament-should be reasonably compact so that theluminaire need not be excessively large, relative toits light output, m order control!, the effec-tively.Summarizing, the lamps used for CTV work ~h~auld:

(a) have acceptable colour rendering properties,both visually and in relation to the colour televisionsystems,(b) have a high camera response,(c) have high luminous efficacy,(d) emit light which blends well with evening day-light,(e) have a reasonably long life,(f) have a suitable light output, and(g) be reasonably compact.

9 Types of lamp available and their relative merits

9.1 1 High wattage tungsten filament incandescentlamps

These lamps have excellent colour rendering pro-perties, but their colour temperature of 2800 K israther low for colour television, and varies with thesupply voltage. Stroboscopic effects are virtuallyabsent on 50 Hz supplies.The major disadvantage of incandescent lamps istheir very low efficacy . up to 18 lm/~l> This meansthat a large electrical load is required for a givenilluminance and picture quality. The current surgeon switching can be as much as sixteen times theoperating current and can damage transformers andother equipment. However, such problems can beovercome by switching banks of luminaires in se-quence rather than in bulk, or by switching on areduced voltage before switching to full voltage.

9.2 Tungsten halogen incandescent filament lampsThe properties of tungsten halogen lamps are similarto those of ordinary tungsten lamps except that theirefficacy and colour temperature are higher, thoughstill on the low side for CTV lighting-up to 26 ln~/~land 3200 K. The small size of the lamps enablesrelatively small lightweight luminaries to be used.

The lamps can be obtained in ratings from 500 W to10 kW. e

9 a 3 Carboy arc lampsThese lamps are often used by television and filmcompanies for location work. They operate frommobile generators. Their spectrum is continuous andhas a colour temperature of about 3800 K whichblends moderately well with evening daylight.

9.4 Xenon lampsXenon discharge lamps have a continuous spectrumand excellent colour rendering properties (Ra8 =90) with a colour temperature of about 6000 K. Un-fortunately their efficacy is low (25 lrnj~) and thecontrol gear required is expensive and complicated.On the other hand xenon discharge lamps have theadvantage that they will re-ignite when hot with theaid of an ignitor, and their relatively small lightsource size makes optical control relatively simple.

9.5 High pressure mercury lamps (Type MB)The blue-green biased white light from these lampsmakes them unsuitable for CTV use, although theyhave been used in the past in conjunction with incan-descent lamps for lighting stadia for black-and-whitetelevision transmissions.

9. 6 Colomr ~~~~ ~~t~~ ]!1JJr’BesSillrQ;¡ ~~~ ~~~ lamps(Type 3£BFi

These lamps can be employed for CTV if they areused in conjunction with incandescent lamps in theratio of about three incandescent lamps to sevenMBF. This is not recommended, however, because ofthe problems of ensuring that the light from bothsources is thoroughly mixed, and because of otherproblems mentioned below.

In common with most high pressure discharge lamps,MBF lamps are extinguished if the electrical supplyis interrupted momentarily, and they will not re-ignite until they have cooled sufficiently which . ..



187

usually takes 5-10 minutes. Such interruptions inelectrical supply are extremely rare and are notnormally a problem.The physical size of MBF lamps makes preciseoptical control of their light difficult, which meansthat they cannot sensibly be employed in narrowbeam projector luminaires.

9. ? High pressure sodium lamps (Types SON,S~DN/’~~

The golden-white light from these lamps lacks blue,as can be seen in Fig. 9b. This makes them unsuit-able for CTV lighting, although it has been suggestedthat if used in conjunction with metal halide lamps inthe ratio of about one high pressure sodium to fourmetal halide lamps the colour rendering would bereasonable. This method is not advocated for thereasons given, but if the light mixing problem couldbe completely overcome the system would have cer-tain advantages over a system employing only metalhalide lamps: that the high pressure sodium lampswould boost the red end of the spectrum where theCTV cameras are least sensitive, and should theelectricity supply be interrupted momentarily, thehigh pressure sodium lamps would re-ignite in lessthan one minute so allowing play to be resumedalthough good picture quality would not be restoreduntil all the metal halide lamps were again fullyoperative. This can take from 5-15 minutes, unlessas on certain types, extra-high voltage ignitors orignitors which maintain lamp ionization are employed,when this time is reduced.

All the high pressure sodium lamps commerciallyavailable at present have efficacies from about 95-.1.

115 lm/W, correlated colour temperatures of about2100 K and a colour rendering index Ra8 of 21approximately.The colour rendering and colour temperature maybe improved in the future by increasing the partialvapour pressure of the sodium in the lamps’ dis-charge tube, but it is unlikely that the lamp will everbe entirely suitable for CTV work because, as Fig. 8shows, it is unlikely that the colour temperature willexceed 2700 K, which is rather low. Such colourimprovements would also be at the cost of efficacywhich would fall to about 60 lrnjW-0

Figs.9a-9e show how with increase in partial sodiumvapour pressure the spectrum broadens from the

Fig. 8. Variations of colour temperature (Tc)and colour index (Ra) of a highpressure sodium lamp with variation in thepartial vapour pressure of the sodium.

F~g. 9. The effect an increase of partialsodium vapour pressure P N on the spectralenergy output of a sodium vapour dischargelamp



188

narrow line spectrum doublet produced by low pres-sure ( IO&dquo;~ torr) sodium lamps to a very broad band,with a gap due to self-absorption when the pressurereaches 700 torr.

9. ~ Metal halide lamps (Types MBI)This family of lamps is the most suitable for CTVpurposes. The sources are relatively small makingprecise optical control possible, the lamps have highluminous efficacies (50-110 lm/W), and good colourrendering properties (Ra 60 to Ra 90).

The basic construction of metal lamps and thematerials used in them are similar to those for highpressure mercury lamps except that, in addition tothe mercury, several metal halides are contained inthe discharge tube. Auxiliary electrodes are notnormally used for starting which is effected by athermal relay or an electronic ignitor. The latter isthe quickest and most reliable method of startingunder all practical installation conditions.

Metal halide lamps have been made with ovoid outerenvelopes coated with diffusing material, cleartubular outer envelopes, and without any envelopesat all. The application of most lamps withoutenvelopes is limited because they can only be used inluminaires specially designed to ensure that thelamps operate at the correct temperature.The metal halides are selected in relation to the

performance required from the lamp, i.e. whethermaximum luminous efficacy is the objective or maxi-mum colour rendering index. Before discussing thisfurther, the basic principles of the operation of theselamps will be described.

9a ~ Basic principles of operationWhen a metal halide lamp is switched on, a mercuryvapour discharge is initiated which quickly heats thedischarge tube and vapourizes the halides. Thesedissociate at the centre of the discharge where thetemperature is in the region of 5500°C, the metal

halides becoming metal + halide. Nearer the tubewall the temperature is lower-approximately 1000°C-and in this region dissociation does not occur, butalready dissociated atoms recombine. Because ofthis phenomenon, pure metal vapour, other thanmercury, virtually never comes in contact with thequartz discharge tube wall, which prevents metalssuch as sodium attacking it.

In most lamps the metals employed have a verymuch lower excitation potential than either the mer-cury or the halide used (Fig. 10). This results inalmost all the light coming from the excited metalsand virtually none from the mercury or halide.

10 Practical metal halide lampsMetal halide lamps can be divided into three groupsaccording to the way they achieve their main objec-tive. <

The first group is designed to achieve high efficacywith acceptable colour rendering. Metal halideswhich generate a small number of spectral linesefficiently are used; the lamps are designed to emitlines whose wavelength and intensity meet therequirements.

The second group achieved good colour rendering atthe expense of reliability and/or lamp efficacy. Thisis achieved by employing metal halides which pro-duce many spectral lines or even a continuum.

The third group consists of a combination of theabove two groups.

10. 1 Metal halide lamps with line spectra

Fig. l1a shows the line spectrum of a metal halidelamp. *

Lamps in this group typically employ metal halidessuch as sodium iodide, thallium iodide and indiumiodide. Excitation produces radiation mainly at theapproximate wavelengths 590 nm (yellow) due to the

10. lonisation and excitation energy levels of some metalvapours employed in metal halide lamps.



189

Fig, 11. Spectral energy distributions of various metal halide lamps, (a) with In, Na, Ti; (b) with Dy, Tl, In;with Sn; (d) with Se, Na, Th.

z

sodium, 540 nm (green) and 670 nm (red) due to thethallium, and 400 nm and 450 nm (blue) due to theindium.

Lamps of this type produce white light most efficientlysince there is practically no ultraviolet and relativelylittle infrared radiation produced. The light is dis-tributed over the spectrum in such a balanced waythat the colour rendering is attractive and the lighthas a pleasant white colour. The colour renderingindex is 65, the efficacy can exceed 100 lm/W, and the

chromaticity point is reasonably close to the blackbody locus (see Table 3).

The lumen maintenance of this type of lamp is good:for example, the output of a 400 W lamp falls by only3 -4 per cent in the first 1000 hours.

Another advantage is that the main spectral lines lienear the maximum sensitivity of each of the threeCTV camera colour channels (Fig. 7), making thistype of lamp particularly suitable for television

Table 3. Comparison of lamp types



190

lighting, since lower illuminances installed loads arerequired to produce a given camera response thanthose needed for other types of metal halide lamps.

ltl. 2 Metal halide lamps with a continuous spectrum.

Two types will be mentioned, of which the firstemploys dysprosium iodide, indium iodide and thalliumiodide. The spectral energy distribution is shown inFig.llb. The dysprosium spectrum consists of agreat number of lines giving a colour rendering indexof 85, but at a reduced luminous efficacy of 85 lan/W.The chromaticity point is x = O. 31~ y = t3. 36, andT K = 6500 K. The lumen depreciation of this type oflamp is greater than that of the previous type. Workis still being carried out to improve this propertyand also to reduce the colour temperature which israther high for CTV purposes.

The second type of lamp in this group incorporatesiodine and tin bromide. The spectrum, ~’ig.11c isbasically continuous, and the colour rendering of thelamp is excellent (ray 90; x = 0. 35,y ~ 0. 3 5;TK = 5000 K). Unfortunately, this is achieved at theexpense of reduced efficacy -45 lm/W for a 400 Wlamp.

10.3 Metal halide lamps with a spectrum consist-ing of lines superimposed on a continuum

Lamps employing metal halide additives such asscandium, thorium and sodium iodides produce a con-tinuum with superimposed lines as shown in Fig.lld; such lamps have Ra = 70, efficacy = 80 ir~a~~,chromaticity point x = 0. 3‘~, y ~ (~. 39, and’I‘K = 4500 K.

i.0e ~ Compact source metal halide lampsSome of the lamps mentioned previously have beenconstructed in the form of compact source lamps.This permits more precise optical control of thelight but this is only an advantage for very narrowbeam floodlight projectors. The increased pressureat which these lamps operate tends to improve thecolour rendering slightly but at the expense of effi-cacy, which is about 70-80 lm/W.

11 l Lighting equipmentIn the main, the luminaire requirements for CTV arethe same as for any other floodlighting application,namely that the floodlights should be strongly con-structed, and made from materials which will with-stand continuous exposure to the weather. The lastis particularly important where the luminaires areused in industrial and coastal areas, or where the

prevailing winds are likely to provide a polluted or asalt-laden atmosphere.The minimum number of floodlights that can be usedin a football stadium depends, amongst other factors,on their beam spread and the need for good illu-romance uniformity to avoid undesirable effects suchas relatively dark patches on the pitch if one lampfails. Twenty-four floodlights are normally con-sidered the minimum for lighting a stadium.

If luminaire with low lumen outputs are used, thenumber required soon reaches the hundreds. Thusfor an installation to provide a vertical illuminanceof 1400 lux, using 80 000 1m 1 kW MBI lamps, 500 ormore floodlights would be required.The greater the number of luminaires, the higher willbe the installation costs, and the higher will be the

cost of lamp replacement, maintenance and cleaning.The last two costs are virtually the same for 1 kWand 10 kW floodlights but with 10 kW lamps only one-tenth the number is required.

Taking all costs into account, between forty and ahundred floodlights per stadium would be recom-mended. For a vertical illuminance of about 1400lux this requires lamps giving 1000 000 Im down to3C~t7 OOt~ lm per luminaire, i.e. 10 kW to 3 kW MBIlamps.Since CTV generally demands large numbers offloodlights, particular attention should be paid intheir design to the ease of installation-mechanicaland electrical-lamp replacement, maintenance, andaiming adjustment (a daylight aiming facility is anadvantage) .If possible the design should allow these operationsto be carried out with one hand, bearing in mind thatfloodlights are often installed at great heights.Where a number of floodlights are being installed onan exposed tower or stand roof, their weight andwind resistance must be considered. It is the pro-jected frontal area and weight of all the floodlights onone tower or stand that is important, and not thoseof a single unit. The weight of the control gear mustalso be taken into account. Fortunately the gear canusually be located remote from the floodlights, eitherat the base of a tower or inside the stand where its

weight is less critical and where it is more pro-tected from the weather and more accessible formaintenance. <

Mirrors, sometimes with the addition of refracting ordiffusing front glasses, are used to control the light.When the angle of incidence of a floodlight’s beam tothe ground is large, stray light above the beam cancause excessive glare. This problem can be over-come if the light above the beam is restricted by theuse of louvres or special reflector designs. Twosuch designs are illustrated in Figs. 12a and 12b;both prevent the lamp being seen directly at anglesabove the main beam. At Everton and Liverpoolfootball stadia floodlights of the type shown in Fig.12a and housing one 10 kW MBI lamp each are used.These luminaires give the light distributions shown inFig. 13a and 13b.

3. ~. m 1 Lamp control gearThe control gear required for lamps suitable for CTVlighting is basically no different from any other.However, some of the metal halide lamps with higharc voltage need to operate from voltages greaterthan the normal mains supply, and step-up leakageflux transformers therefore used to producesecondary voltages up to about 1100 V, the inductanceof the transformer limiting the lamp current.

If a suitable capacitor is placed in series with thelamp, -which is then operated from such a trains-forxner, the lamp can be made to operate at approxi-mately constant wattage in spite of fluctuations ofmains supply voltage and changes of lamp character-istics during life.

The constant wattage effect is achieved because theoverall impedance (Z) in series with the lamp iscapacitive, the capacitive reactance Xc being greaterthan the transformer’s inductive reactance XL’ Be-cause a section of the core of the transformer ispartly saturated, its inductive reactance varies with



191

Fig. 12. Luminaire optical desigits which produce a sharp above the main beam;(a) for a 10 kW MEIIH lamp; (b) with a parabolic and shield.

Fig. 13. Intensity distribution of the luminaire shown in F’ig. 12(a) when fitted t4,,ith a 10 kWMBI lamp. (a) Intensity distribution; (b) isocandela diagram.

current in such a way that an increase of lampcurrent produces a decrease in inductive reactanceand an increase in total impedance, i.e., Z :~-- XC - XL.The increase of impedance has the effect of limitingthe change of lamp current and hence lamp wattage.The type of circuit shown in Fig, 14a is used toa 10 kW kiBI The circuit voltageis less than 900 V, which means that standard 600/1000 V cable can be used on the secondary wiring.Unlike the majority of the other circuits, where alarge input current occurs at switch-on, constantwattage circuits have the advantage that almost per-fect power factor is maintained during lamp startingand running up. This means that the input currentrises to full value slowly. An electronic ignitor isused as a starting aid; this is common practice withthe majority of metal halide lamp circuits.

Most metal halide lamps will ignite with voltagepulses of less than 2000 V, which means that 600/1000 V standard p.v.c., silicon rubber, butyl, or hyper-3:~~ insulated cables can be used. Mineral insulatedcables or glass fibre varnish impregnated cablesshould not be used since they tend to break down d-deto ignition pulses which occur during the Mmp’sstarting process.

Often the cable length between the control gear orignitor and lamp is limited;. depending on the type ofignitor employed as indicated in Figs. 14b and 14c.This limit is necessary because the ignition pulsesare attenuated by the stray capacitance Cg of thecables. The actual limiting distances range from 5 mto about 200 m, depending on the type of lamp, cableand ignitor.



192

14 (a). Constant wattage control gear circuitfor a 10 kW MBIfH metal halide lamp. (b) and (e)The length of cables B-La and I-La is limited bytheir total capacitance to earth and otherconductors.

12 Stroboscopic effects

Except for tungsten filament lamps, all dischargelamps operated on 50 Hz or 60 Hz supplies producepulsating light with frequencies of 100 Hz or 120 Hzrespectively. Stroboscopic effects may result whenmoving objects are viewed. The effect can virtuallybe eliminated by ensuring that luminaires on threedifferent phases of the supply illuminate the samearea.

13 Lighting techniquesThe introduction of colour television has called for a

re-appraisal of the visual requirements in stadiumlighting.

13- 1 Early systemsHitherto when designing floodlighting schemes forspectator viewing and for black-and-white television,the main criterion was the illuminance on the hori-zontal plane (Eh). Recommendations for lightingstandards were, however, expressed in kW loadingrelated to ground spectator capacity and limitingof elevation from the centre of the pitch to thefloodlights, and were therefore a guide only. Verticalilluminance was not ignored, the relationship betweenvertical and horizontal illuminances depending on theaiming plan adopted for the floodlights. This aimingplan took into account the amount of glare consideredto be acceptable.The type of scheme chosen, e.g. a tower system or amultiple group (side-mounted) system, governs thepossibilities available to the designer. Side-mountedsystems, particularly those where the floodlights are

installed in an almost continuous line (the linesystem), are more likely to cause glare than towersystems unless particular care is taken regardingmounting height and the angle of elevation from thecentre of the pitch to the floodlights; good optical con-trol of the light from each projector is the mostimportant requirement.The use of these systems designed basically to pro-vide uniform horizontal illuminance can result in avariation of vertical illuminance across the pitch.With a four-corner system, the average verticalilluminance is generally about 0. 5 Eh av., dependingupon the siting of the towers, and is a minimum at thetouchlines and a maximum along the centre lines.Where side-mounted systems are used the verticalilluminance over the near half of the pitch from a lineof floodlights is approximately equal to the averagehorizontal illuminance. Beyond the half-way line andcontinuing out to the far touchline the vertical illu-minance may ultimately drop to approximately 0. 25Eh av. Whilst it might be desirable to maintain amore uniform vertical illuminance across the playingarea in any one direction, it is necessary for thevertical illuminance to vary around an object in orderto produce good modelling.

1.3. 2 BBC requirements for colour television trans-missions

(a) A minimum incident illuminance of 800 lux isrequired for the whole playing area. For the purposeof this specification illuminance shall be measured1-1. 5m above ground level on a plane perpendicularto the line of sight from the main camera position.(b) For televising events requiring the use of longfocus close-up lenses, the average illuminanceshould be not less than 1400 lux. In practice theselenses are nearly always wanted, and an installationnot providing 1400 lux will be in some way inhibiting.(c) The average illuminance on spectator areasthat are likely to be seen by the cameras should benot less than 35 per cent of that on the playing area.* *’

(d) The maximum and minimum illuminances on the

playing area should be within 20 per cent of theaverage illuminance. Illuminance fall-off should notexceed 20 per cent over any four metres within the

playing area.(e) The light sources should be installed to avoidlens flare.(f) The correlated colour temperature of the lightsources should lie within the range 3000 K to 5500 K.Sources with a correlated colour temperature above4000 K are preferred.(g) The colour rendering properties of the sourceused shall be suitable for television systems andshall not vary over the whole lit area.

Lighting that falls short of the above requirementswill limit the effective coverage of events such ashorse jumping and rugby football as the illuminancewould be insufficient for close-up shorts, and thedepth of field would be restricted for cameras placednear the touchlines so that players in the d7L~&dquo;fere,-~’it

planes would be out of focus. Also a good technicalquality would be difficult to achieve and maintain.

The above requirements are the result of a recentrevision, but differ only slightly from those estab-lished by the Outside Broadcasts Department in 1968.

* At the 800 lux level approximately 275 lux; at the1400 lux level approximately 480 lux.



193

Illuminance. should be measured by aiming the photo-cell of the incident lightmeter towards the maincamera position from each point in the playing orspectator area (previously the recommendation was15° from the vertical aimed towards the cameras).These requirements are not intended as a precisespecification; evaluation of a given installation or a.proposed scheme still requires appraisal by tele-vision and lighting engineers and it is suggested thatstadia authorities seek advice from the OutsideBroadcasts Department.

13. 3 Union des Associations Européennes deFootball

UEFA have recently issued their draft require-ments for the lighting of football grounds and includea minimum illuminance of 1500 lux for colour tele-

vision, ’measured in a plane perpendicular to thevisual line of the camera’.

~3q 4 Colour television systemsFrom the BBC requirements, it will be noted thatmuch higher illuminances are required for colourtelevision transmissions than are needed for black-and-white transmissions and in particular that thespecification refers to illuminance in a plane per-pendicular to the line of sight from the main camerapositions measured 1-1. 5 m above ground level. Itis necessary therefore when designing a floodlightinglayout to make a large number of accurate calculation:in this plane, in addition to those for the horizontalplane. The angle of the plane related to the maincamera viewing axis (generally referred to as’vertical’) will alter, in respect of the field, as thetarget moves across the playing area. It is not suf-ficient to specify vertical illuminances in the fourdirections parallel to the sides and ends of the fieldas will be seen later. A reasonable uniformity inboth vertical and horizontal illuminance is required,and from Section 13. 1 it will be seen that a side-mounted system is more likely to give the requiredresult than a four-corner tower system.

Assuming that sufficient height is available and thatthere are no structural problems, floodlights can bemounted along the stand roofs on each side, eitherspaced in groups or in a nearly continuous line. fnorder to provide all over the pitch the high values of’vertical’ illuminance required by the colour tele-vision camera a number of floodlight projectors mustbe trained on aiming points located near to the fartouchline. Viewing angles between the far spectatorsand the line of floodlights must be carefully checkedto avoid intolerable glare. The projectors should bedesigned so that the light source itself is screenedfrom direct view at these angles.

Floodlights installed along the edge off a stand roofcan sometimes come almost vertically above thetouchline. Whilst the horizontal illuminance may be

quite high, in the ’vertical’ plane would be in-adequate. Smaller floodlights should then be installedbelow the. roof and back inside the stand to build upthe illuminance to the required value.

Where circumstances do not permit side-mounting(e.g., where there is insufficient height available atthe stand roof, or where there are structural pro-blems) a four-corner tower system can be used pro-vided some additional side lighting is installed on thestand where the main camera is located. The needfor this arrangement is apparent from Fig. 15. It has

Fig. Z5. Vertical illuminance and CTV camerafour-tower system. The areas enclosed byeach semi-circle indicate the zones where novertical illuminance component is received bythe ec~rraera frorn towers 1 and 3.

been emphasized that the main criterion is ’vertical’illuminance in the direction of the main camera

position. A four-corner tower scheme may presentareas where the ’vertical’ illuminance is inadequatein the camera direction. For example in the area ’A’,only the floodlights on pole 2 contribute illuminationfor the camera. The vertical illuminance in the otherdirections may be quite high, and an average valuecalculated from any four directions in respect ofselected positions over the pitch would give a mis-leading figure. The deficiency can be remedied byinstalling additional floodlights along the stand roofwhere the camera is situated. The mounting heightin these circumstances could be about 15 m for theseadditional floodlights. Such an arrangement willresult in an asymmetric installation, and providedthat the differences in vertical illuminance from the

spectators’ point of view in the end stands is not verygreat, the visual result will be quite satisfactory.

Uniformity of illuminance is usually expressed asEmín/Emax, e.g. 1: i< 5, but this ratio alone is notsufficient. The variations in illuminance between thelimits should be gradual and the difference over aspecified distance should therefore be given. This isparticularly important where colour television isconcerned as sharp variations may affect picturequality. The BBC requirements that the fall-off in illuminance should not exceed 20 per cent overany 4 m in the playing area. A uniformity ratio ofapproximately 1:1,5 together with a gradient ofabout 3-5% per metre has found to be satisfac- r

tory in practice.The BBC specification also calls for an illuminanceof 35% of that in the playing area for the front specta-tor areas (i.e. approximately 460 lux). This is neces-sary to maintain a luminance contrast range withinthe luminosity range of the camera. Should the peri-pheral areas be insufficiently illuminated, picturequality will be poor due to lack of background detailand ’lag’ effects.



194

The lighting of them near-spectator areas makes aszgnificant contribution to the luminance pattern ofthe scene. Visual performance and visual comfortdepend upon the luminance ratios between the task(or target), its immediate background, and the generalenvironment.

In following the flight of a ball the observer’s eyeswill be subject to changes in luminance which maycover a considerable range. His field of view willinclude in rapid succession the playing pitch, thestands, and the floodlights. His adaptation level maybe affected to the point that visual performance be-gins to deteriorate and fatigue may result. Adaptatio:from a high to a low level of luminance is not in-stantaneous as it is from low luminance to high. Anexample of the former would be when a spectatorchanges his direction of view from the playing fieldto the stands where these are not adequately lighted.The problem is more acute when a spectator followsa high ball and in doing so looks in the direction ofthe floodlights whose luminance is more than ahundred times the mean adaptation luminance of thescene. The adaptation state of the eye is raised to acorrespondingly high level and subsequent re-adaptation to the field luminance level may takeseveral seconds; after-images of the floodlights maybe seen. Fig. 16 illustrates this point and shows forexample, that the process of re-adaptation from aluminance value of, say 8000 ad~’~n~ to 100 cd/n2takes about five seconds.

The luminance of the floodlights must therefore becontrolled and the arrangement of the projectorswhether on towers or along the stand roofs also hasconsiderable relevance. The state of adaptation ofthe eye will be less affected if the floodlights arebroken up into small groups. If, therefore, the angle

Fig. 16. Adaptation process as a fz~r~~~ie~~z of time.(Data apply to 50-75 per cent of observers).

subtended by a group of floodlights as seen by anobserver is small, i.e. less than say 2°, and if thedistances between groups represent significantlygreater angles, interference with adaptation will beminimum.

A side-mounted multiple group system, or a linesystem, can be designed within these limits. With aline system the limitations applies to the width of thevertical band of light. The comparatively largebatteries of floodlights from tower arrangementsare outside the angular limits given above and inter-ference with adaptation can therefore be considerable.

~.3. 6 Glare control

If the floodlights are trained to produce high valuesof vertical illuminance on the far touchlines, therewill be considerable risk of glare from the maincamera position. Fig. 17 shows the result of a testcarried out during a study on interior lighting usinggroups of observers to determine the dark and brightlimits of a luminous object for a particular set oftest conditions involving background luminance andobject size. For a background luminance of 100cd/M2 (for example the playing field), the range ofluminance of small targets extends from approxi-mately 0. 5 cd/n~~ up to 4000 cd/m2..The luminance values of floodlight projectors usedfor stadium lighting will normally exceed 4000 cd/in~:the values in the spill light region for typical flood-lights could be 10000 cd/M2 or more. If the pro-jectors are side-mounted and arranged in groups, orin a line, the risk of interference with adaptation will

Fig. 17. Maximum luminances for comfortableseeing obtained with small targets againstuniform background



195

be lessened and the glare will be a minimum. Obser-vations of practical installations have indicated thatwith angles of elevation to the light sources of 15°,and with sources that subtend an angle of less than 2&dquo;,a luminance ratio (light source to field) of 100: 1 issatisfactory.For angles of elevation of 20’ a luminance ratio of2Ct0 : 1 is permissible, whilst for batteries of flood-lights installed on towers subtending angles greaterthan 2°, a luminance ratio of 50: 1 would appear to bethe limit; this is illustrated in Fig. 18. Here thethreshold source luminance is related to size ofsource for some values of background luminance. Forthe data obtained by Arndt e~ al, the vertical angle tothe source was 20’, and the horizontal angle betweenthe source and the viewing direction was 0°. Hopkin-son and Bradley’s data were based on direct sourceobservation, including large luminous fields. Thesecurves and their possible extensions to smallersizes of source would appear to support the limitssuggested above.

In side-mounted schemes the floodlights installedalong the stand roofs, although mounted at a com-paratively low height, are seen against a lightedbackground (near-spectator areas and stand interior),whilst with tower installations the floodlights areviewed against a dark sky. The contrast between thelight sources and the background is then a maximumand glare is consequently considerable. It is interest-

ing that conditions met with in the past have been

Fig. 18. Relationship of discomfort glarethreshold luminance to source size for a rangebackground luminances.

reversed in that with installations designed to meetcolour television requirements, the side-mountedarrangements rather than tower systems presentless risk of glare (Fig. 19).

~’ig. 19..~iv~~~c~at F.C. showing Kemlyn Road stand with 10 kW !I~IB~/~I at high level and 2 kWl~’I~I/I~~~acadtight~ below the stand roof.



196

Good modelling is necessary for recognition of de-tails but dense hard shadows should be avoided. Theeffect will vary according to the positions of viewand the layout of the installation. A side-mountedmultiple group system or a line system, will produceacceptable modelling. Since the projectors are dis-tributed in small groups, or in a near-continuous

line, along the sides of a stadium, shadows will tendto be softened and reduced. In a four-tower system,however, the shadows are likely to be comparativelyharsh and there will be a greater variation in modell-ing effects over the area. Four deep shadows will beseen on the field of play.It should be remembered also that the relativeluminance range for the colour television camerabetween black and white is less than for the eye, andthat for a given luminance difference as observed bya spectator at the stadium, a greater difference isseen on the television screen (representing anincrease by a factor of about 2). Shadows cast by theplayers on the field are thus accentuated and harshshadows should therefore be avoided.

The ratio of the illumination vector to the scalarilluminance has been proposed as an index of modell-ing for lighted interiors. This concept might well beextended to exterior lighting-particularly where athree-dimensional quality is called for as in thelighting of sports stadia.

14 Computer-aided lighting designThe calculations necessary when designing flood-lighting layouts are tedious, and very complex wherevertical illuminance is concerned. In planning thelighting of sports stadia both horizontal and ’vertical’illuminance must be taken into account.

A large number of calculations are required in bothplanes (approximately 600) and use is made of a griddividing the area into small squares. For horizontalilluminance the grid is taken to be at ground level,whilst for ’vertical’ illuminance the grid is at aheight of 1.2 m above the pitch. A system of co-ordinates is used to define the mounting positions ofthe floodlights and their aiming points both in planand elevation.

A layout is first prepared by the designer to deter-mine the number of floodlights required and theirbasic aiming points. From this plan data are pre-pared for a computer giving the position of eachfloodlight and its type (floodlights of different beampattern and using more than one type of lamp areoften required in one scheme). The required illu-minances are calculated for both horizontal and’vertical’ planes on the grid which has been chosento enable a detailed pattern of the illuminance to beproduced. The designer can then study the patternand make any adjustments that may be necessary tohis plan. The more skilled a:id expe %iencedthe designer, the closer will his first aiming plan beto the desired result.

The computer can be programmed to give a completelist of the floodlights and lamps used together withdetains of each floodlight position and aiming point.The detailed illuminance calculations can include the

figures for average illuminance in each plane, totalinstalled luminous flux and the efficiency of thearrangement. The use of a computer in planningfloodlighting schemes for colour television is

essential, not only for the obvious advantage of speed,but also for the degree of accuracy that can beobtained.

AcknowledgmentsThe authors wish to acknowledge help received fromtheir colleagues in the preparation of this paper andin particular Mr H. C3-0 Westermann.

BibliographyDraft international recommendations for lighting atoutdoor sports grounds, Commission Internationalede l’Eclairage (1971).

Ellenbaas, W., High pressure mercury vapour lampsand their applications, Philips Technical Library.

Beijer, L. B., Broerse, P. H., Davies, I. F., and Holmes,T., Brighter, whiter and mightier light sources forpublic lighting, Publ. Ltg., 35, 173 (1970).IES Technical Report No. 7, Lighting for sports, TheIlluminating Engineering Society, London (1965).

Dorey, G. M., Pierce, M. W., and Price, W. A., Footballfloodlighting, Trans. Illum. Eng. Soc. (London), 29, 29(1964).

Regulations and recommendations for lighting offootball grounds. Union des Associations Euro-peennes de Football, Berne (June 1971).

Beeson, E. J. G., and Aldworth, R. C., The developmentof new discharge sources for colour television andtheir use for other floodlighting applications, Publ.Ltg, 36, 154 (1971).

Ackerman, K. R., Lighting for colour television, Ltg.Res. & Tech., 1, 123 (1969).

Hopkinson, R. G., A proposed luminance basis for alighting Code, Trans. Illum. Eng. Soc. (London), 30,63 (1965).

Schreuder, D. A., The lighting of vehicular traffictunnels, Clever-Hume Press, London (1965).

Bodman, H. W., Quality of interior lighting based onluminance, Trans. Illum. Eng. Soc. (London), 32, 22(1967).

Lynes, J. A., Burt, W., Jackson, G. K., and Cuttle, C.,The flow of light into buildings, Trans. Illum. Eng.Soc. (London), 31, 65 (1966).Jackson M. G. A., Rogers, B. C., and Lewis, R. A.,Colour television transmission—outdoor lighting.Light Ltg, 62, 265 (1969).

Aldworth, R. C., Football field floodlighting for colourtelevision, Light Ltg, 64, 154 (1971), and 64,188 (1971).BBC requirements for television lighting at stadiafor colour outside broadcasts, BPW/AAA Issue 3(August 1971). Advances in ’Plumbicon’ camera tube design,Electronic Applications, 30, 18 (1971).

Lord, A. V., Advances in colour television cameras,Electronics and Power, 17, 337 (September 1971).

Patnam, R. E., Wiggin, J. F., Clark, C. N., and WilliamsH. G., Discharge lamps and colour television, Jnl.S.M.P.T.E. 78, 700 (1969). van der Meer, A. J., Outdoor lighting for colour tele-vision, Int. Ltg Rev., 19, 62 (1968).



197

van Doorn, A. G., The ’Plumbicon’ compared withother television camera tubes, Philips Tech. Rev.,27, 1 (1966).

Beijer, L. B., Light sources for sportsground lightingwhich are also suitable for colour TV broadcasts,Electrotechniek 49, 665 (1971).

Westermann, H.O., Lighting technique for stadia,Electrotechnick 49, 748 (1971).

CIE Publication No. 13, Method of measuring andspecifying colour rendering properties of lightsources (1965).

Breimer, H., Holm, W., Tan, S.I., A colour televisioncamera with ’Plumbicon’ camera tubes, Philips Tech.Rev., 28, 346 (1967).

Tienkamp, E., About colour TV cameras, some re-strictions and possibilities, Elektrotechniek, 16, (1971).

Sauter, D., Halogen-metalldampflampen im farb-fernschen, Fernsehund Kino-Technik, 24, 255 (1970).

Judd, D. B., MacAdam, D. L., and Wyszecki, G., Spec-tral distribution of typical daylight as a function ofcorrelated colour temperature, J. opt. Soc. Am., 54,1031 (1964).

Schmidt, K., Emission characteristics of high pres-sure sodium discharges, Bull. Am. Phys. Soc., 8, 58(1963).de Vrijer, B. Hoge-druk natriumlamp, Elektrotech-niek, 43, 512 (1965).

Reiling, G. H., Characteristics of metal vapourmetallic iodide arc lamps, J. opt. Soc. Am., 54, 532(1964).

Beijer, L. B., Jacobs, C. A. J., and Tol. T., The iodidedischarge lamps, Philips Tech. Rev., 29, 361 (1968).

Rokosz, F., Decker, W.J., and Fraser, H.D., Factorsaffecting metal halide lamp parameters and operat-ing characteristics, Illum. Engng, 62, 626 (1967).

Zollweg, R. J., and Frost, L. S., Molecular radiationfrom high-pressure discharge, Proc. Phenomena inIonized Gases, 224, Vienna (1967).Leo Mori, Tadatoshi Higashi, Seifu Nomura, IsamuOsawa, A newly developed lamp with maximum colourrendering property, Toshiba Rev., (Nr. 32), 44 (1967).

Discussion

ffe R. Ackerman (BBC): This paper is the first tocome to my attention which covers sports lightingfrom both an illuminating engineering and a televi-sion engineering standpoint. Television lighting forsport has reached the stage that interior lighting hadreached in the early 1950s. An understanding hasdeveloped of ways of achieving adequate illuminancesand a concern is being shown for disability glare;we have not, as yet, developed an understanding ofhow to light for maximum quality both of the tele-vised image and of the scene as viewed by the spec-tator. The paper rightly calls for constancy of thespectral quality of the light and for strictly controlledcontrast ratios, for we have not yet developed theskills to dispense witch these crutches. Furthermore,the modelling achieved is largely fortuitous. Sidelighting gives reasonably even lighting, with towersthere are four strong shadows. I look forward towhen we can design for modelling, perhaps by theuse of vector-scalar ratios, and when we can usedifferences in colour of light creatively, perhaps

simulating the fading sun with a slight increase inreddish light in one direction and bluish light inanother.

The paper has made it clear that illuminance designand control of glare for a football ground are complexproblems only to be solved with the aid of computerswhich have also been used to optimise the mix inmercury halide lamps. The computer was fed withinformation on colour camera optical spectral re-sponses and with the spectral compositions of sometwenty-six materials typical of surfaces encounteredin the visual scene; they included grasses, pillar boxes,and Eurasian facial complexions. The programmeprovides a printout of the colour errors to be expec-ted for each of the chosen materials. The results areusually expressed in just noticeable differences (jnd’s)when assessing the quality of results to be expectedfrom a source.

Mercury halide lamps have proved almost ideal sourcesfor the lighting of stadia, and broadcasting authoritieswelcome the increasing number of installations at thegrounds of leading clubs. This trend is being en-couraged by the European Football Association whichis making lighting suitable for colour television acondition for allowing a ground to mount any majorEuropean match. But the BBC covers many othersports where the organisers have much less moneyavailable and we still have considerable problemsin erecting equipment to provide the illuminancessuggested in this paper. We would welcome the deve-lopment of transportable kits of lightweight equipmentwhich could be wheeled into, for example, Rugby Foot-ball grounds and give acceptable lighting with no morethan twenty-four hours’ preparation.I notice that the authors refer to colour television

being balanced for Illuminant C. This was true atone time, but the modern equivalent is IlluminantD65, which they recognise in Table 3.

With reference to camera lag (i.e. the disturbingeffect of smearing of the image when the camera ispanned and which gets worse when the illuminanceis inadequate), I would add that in addition to therequirement that tube current should exceed 100 nAthe tube currents should also be reasonably equal.If they are not, the picture suffers from colouredlag, which is more disturbing than neutral lag. Un-fortunately, this design parameter has to be a com-promise because if currents are equalised by re-ducing the signal to the higher current tubes, a lossof sensitivity results.

I am puzzled by the reference in Section 4. 2 that thetelevision system amplifies the sensation of lumi-nance difference by a factor of 2. The designed over-all system gamma is I, although in practice it maybe 1 Q 1 or ~. 2. In effect this means contrast ratiosare rendered reasonably realistically. Do the authorsmean a subjective impression of a factor of 2? Thecamera system a limited contrast handling caaa-city so if the contrast range is greater, detail is losteither in some of the black areas or some of thewhite areas, but the effective contrast should be simi-lar. There appears to be some misapprehension inthe diagram, where it shows the eye having the charac-teristics covering I-1(i ~3f~0. 1 do not think that the

eye does achieve this at any one instant of time.

In referring to camera optics, the authors say that’the aperture ~i.e. the ~f’ number of the zoom lens)is increased automatically with focal lengththus resulting in a constant light value on the camera



198

tubes’. Should this not have read ’the aperture iskept constant with increasing focal iength... , . , .. ’ ?However, a third type of zoom lens used frequently atfootball grounds has a l. fi : 1 or even 18 :1 1 zoom rangeand, to keep its bulk reasonable, its maximum aper-ture reduces at the narrow angle end, thus reducingthe light on the tubes. This means that under low lightconditions these narrow angles cannot be used.

Finally, I have had some difficulty in correlating MrAldworth’s proposition in his recent article in Lightand Lighting that 800 000 cd intensity is a reasonablelimit per floodlight with the authors’ suggestion of asource brightness limit of 1.t10Qt~ cd/m2. Have wereached the stage where a design method can bedescribed which will ensure a reasonable and pre-dictable glare limit?

R. C,Aldwartla (Thorn Lighting Ltd): This paperimplies that lighting for colour television can only beachieved by a side-lighting system. This is not true.Also, no useful purpose is served by arguing for onesystem rather than another. In practice the choice ofsystem will be influenced by factors far removedfrom lighting design. Both side-lighting and towersystems are needed which will give the best resultswithin the limitations set by each installation.

On the subject of glare, I feel there is some confusionbetween ’adaptation interference’ which occurs whenlooking at a corner tower and being temporarilydazzled, and the build-up of’visual fatigue’ experiencedwith side-lighting systems when the glare sourcesare constantly within the normal field of view. Myexperience of two ’near-continuous line’ and multiple-group’ arrangements conflicts with the authors’ con-~iusic~ns, In both cases the installations use similarnumbers and types of floodlights at the same mount-ing height, and my experience is that the near-con-tinuous layout is more comfortable than the multiple-group arrangement. With a near-continuous line offittings the source brightness represents a constantfactor in the adaptation to the luminance in the fieldof view, which does not vary greatly as the spectatorfollows the game. On the other hand the multiplegroups introduce ’adaptation interference’ whichleads to a build-up in ’visual fatigue’. With four towersthe risk of p adaptation interference’ is reduced andvisual fatigue therefore is less.

Except in citing BBC recommendations, the authorsdid not mention lens flare, which is the camera’suffering’ from glare. Lens flare is seldom a prob-lem with a corner-tower system but with side-light-ing systems it is common, particularly with lowmounted cameras.

Fitting luminance deos not vary with angle. The re-flector luminance in the spill light zone of a flood-light is the as that in the main axis. It is the

intensity which varies.

In an earlier paperl I~1~° Ackerman pointed out thatthe ideal lighting s~~i~-r~~ to have a ’key’, ’fill’ and7backlight’. With a four-tower system the ’key light’is provided by the tower nearest to the subject andthe camera position and the ’fill light’ by the furthesttower on the same side of the ground. The °ba~i~3.~.~ht?is provided by the two towers on the opposite side tothe camera position. Although the direction of the’key light’ will change as the player moves from oneend of the pitch to the other, and the ’key’ and fill’will be equal over the central area, the lighting con-ditions are much closer to the stated requirementsthan with a side-lighting installation.

The BBC state in their published requirements thatthey do not consider them to be a precise specificationand I would therefore take issue with Fig. 15 whichshows the small area of pitch which received light fromonly one tower. Apart from a query on the geometry,which appears to take no account of the fact that aplane normal to the camera would not usually be verti-cal, the diagram shows an area no more than 6&dquo;10 ofthe total pitch area. It is the corners and the fartouch line where television engineers are in trouble.There is no problem on the near touch line as thediagram implies. Supplementary light in this area isnot required for a corner-tower system, even whenthe towers are in line with the touch line, as with theWest Ham installation.

The important thing is that we are measuring andcalculating on a flat plane, but the camera is lookingat three-dimensional objects. For Fig. 15 to be validin practise, switching out Tower 2 would result in thecamera seeing a player in area A in sharp silhouette.This is plainly not the case. If the BBC admit thattheir specification is not a precise document, let usnot treat it as if it is.

The authors suggest scalar-vector ratios could pro-vide a guide to modelling and picture quality. I would

suggest that a vertical-cylindrical ratio would bebetter. Limited work carried out using this ratioseems to agree with the results achieved by Lemonsand Neenans.2 2

Much as I value the theoretical work on glare andother factors, we now have a number of football in-stallations to which we could apply the technique ofappraisal in the same way as we have done for in-terior lighting. These would be more difficult toorganize and to be really effective would requireMr Ackerman to attend with a colour scanner vanso that spectator and player requirements could beappraised in parallel with the colour television re-quirements.

1 Ackerman, K. R., Ltg Res. & Tech., 1 (3), 123 (1969).2 Lemons, T. B., and Neenans, C. J., Illum. Engng,

65, 97 (1970)

W.A. Price (Osram-GEC Ltd): I think we aremaking the mistake of measuring certain illuminancesin a satisfactory installation and then using thesevalues as a specification for other installations. Theonly illuminances acknowledged in the BBC colourtelevision requirements are those on planes facingthe main camera position. These levels can beachieved most economically from a single bank offloodlights behind the camera. Also, this is probablythe only practical way in which the very tight diver,-sity requirements can be met and floodlights in thisposition will not cause camera flare. Obviously ’Ll-islighting from a single point would not be satisfactoryfor players and spectators and probably not for tele-vision transmissions either, yet we are biasinginstallations way.

By specifying the illuminance only on planes normalto the camera we are ignoring that we are lightingsolid forms and that light towards the camera isvery effective in lighting grass and mud. A footballlit from a direction at right angles to the directionof view is still visible. Grass will be at least as

bright from the same light coming towards the cameraas from behind it. As the grass (or mud) forms themajor content of most television pictures, its lumi-nance must be important, which suggests that the



199

lighting levels might be reduced to one-half thosespecified if we have a symmetrical installation.

The authors have ignored many advantages of thefour-tower systems, probably because of their ob-session with illuminances on planes facing the cameraposition. However, seldom does the lighting engineerhave a free choice of system, because stands andother site considerations restrict floodlight mountingpositions, but an open mind on the advantages anddisadvantages of the various techniques will helphim design an installation within the imposed limits.

R.L. C. Tate (Thorn Lighting Ltd): Perhaps themajor objection to tower lighting is overcome whena really concentrated light source is used to givea very narrow beam. It is then possible to makefloodlighting equipment which, mounted on towers,will give a very tight beam and consequently verylittle glare. Glare will only be experienced whenlooking directly into the beam-which can only bedone from the field and the spectator is not on thefield.

~. ~.1’ra~ke~ (Osram-GEC Ltd): The 10 kW lampwas shown with a ’constant wattage’ ballast, whereasthe other lamps were shown operating on conventional’choke’ ballasts. The authors stressed the need for

constancy of colour in operation. As the supply vol-tage may change by the allowable ~6°l° during a matchthere will be change in lamp wattage. Is this thereason for using the more expensive circuit for the10 kW lamp? Are any figures available for change ofcolour temperature against supply volts for theunstabilized circuits?

S. K. Guha (Building Research Estal~lisla~aent}: iHave the authors tried the concept of scalar or cylin-drical illuminance ? I should have thought that thiswould give a better appraisal. Of the two systems offloodlighting mentioned, which gives performance,(e.g. light distribution characteristics) closest tothat of daylight? Problems of adaptation for the spec-tators might be easily eliminated by designing forrelatively high illuminances at the entrances: Hasthis principle been applied?

F. B. P. Hampton (Thorn Lighting Ltd): This paperoffers some rather dogmatic guidelines to designersof stadium lighting schemes. We are advised thatlamps should have a minimum output of 15 000 1m,and that a maximum of 100 lanterns should be used.Such figures are not relevant to the time require-ments of equipment selection which requires thatall factors affecting the total efficiency of the schemeshould be considered in terms of light provided bythe system against total cost. It is easily conceivablethat a system using larger numbers of smaller lampscould be a better solution under some circumstances.

Any useful guidance which would reduce the numberof decisions which must be taken when designingindividual ~y~t~~ =~o~:~d be welcome but theauthors, suggestions, if adopted, would impose un-warranted limitations on the designer

W. Bownes: With spectators all round a groundand up to 40 ft above the playing area, it is impossibleto design lighting which meets satisfactorily therequirements of players, spectators and the televi-sion camera. However with a six-to.wer installationthere will always be light from three directions forthe benefit of the camera, and if the lights are at a

height of 120-130 ft there can be complete freedomfrom glare.

M. G.A. Jackson (in reply): As Mr Ackerman pointedout, designing for modelling is important and we agreethat the effect achieved is largelyffortuitous. A side-mounted arrangement (multiple group or broken linesystem) is more likely to produce acceptable modell-ing than a four-corner tower scheme where shadowswill be comparatively harsh. Mr Ackerman suggests,as we do, the use of vector/scalar ratios as an indexof modelling, whilst Mr Aldworth favours the vertical-cylindrical ratio. We need to keep an open mind onthis until more experience has been gained.

There are a number of problems concerned withproviding portable lightweight equipment to supple-ment the existing lighting at some stadia as a tempo-rary measure for CTV transmission. Not the leastis the location of the equipment to provide the re-quired high vertical illuminance with reasonable uni-formity and minimum glare. If the existing installa-tion provides at least 400 lux (vertical) additionaltemporary lighting may be feasible on the standwhere the main camera is positioned, given a suitablemounting height. A side-mounted arrangement wouldusually give better results from the point of view oflight distribution with minimum glare, and installationproblems would be reduced compared with temporarylighting from corner towers. Conditions vary verymuch between different stadia so that lighting installa-tions, temporary or otherwise, need to be purpose-designed for success.

The comments on the subject of glare made by severalspeakers show that we have yet to produce a practicalcode of glare control for football stadia. Mr Aldworthhas suggested an intensity limit of 800 000 cd perfloodlight to the eye of the nearest spectator. Thisfigure is based simply on a limit of 100 000 cd sug-gested by Dorey, Peirce and Price in 1964, and in-creased in direct proportion to the illuminance ofinstallation. Whilst the original figure of 100 000 cdseemed satisfactory, it is perhaps too early to acceptMr Aldworth’s extrapolation. We think that our sug-gestion of a luminance design method has distinctpossibilities. Much work has yet to be done and wehave used the results of interior lighting studies asa guide. We have therefore expressed possible glarelimits in terms of the average luminance of a groupof luminaires or isolated individual luminaires asseen by a spectator compared with the backgroundluminance. The maximum luminance in a part ofany reflector viewed in the spill light region maybe the same as that in the main axis, but the eyewill normally take into account the less bright areasand assess the average luminance. Mathematicallythis is the sum of the intensities in the direction ofthe viewer divided by the projected area taken up bythe groups of luminaires. Much depends, of course,on the light distribution pattern of the floodlight.With reference to Mr ’~at~~~ comments, the spillregion be in a projector atightly controlled beam, but such a fitting is not nec-essarily suitable for large area floodlighting purposes.This may answer Mr Ackerman’s final question andalso the point made by Mr Aldworth with referenceto our relating glare to fitting luminance.

Mr Aldworth’s comment on glare in relation to themerits or otherwise of multiple-group and near-con-tinuous line systems seems confused. A multiple-group system, as he envisages it, is more likely to



200

lead to glare problems than a near-continuous linelayout, but I am sure he is thinking of comparativelylarge arrays of floodlights mounted in three or fourgroups along each side, whereas we have illustratedand demonstrated what is really a broken-line arange-ment which, perhaps, should not be referred to as amultiple-group system. Such a broken-line systemmay in practice prove superior to a continuous-linearrangement both from the point of view of adaptationinterference and glare.

Returning to Mr Aldwarth’s opening comment, thepaper does not say that CTV floodlighting can onlybe achieved by a side-lighting system. We have triedto show that given the proper mounting height, side-lighting systems give the best opportunities to meetthe current BBC specification on illuminance. Ob-viously four-corner systems must be catered for-and we illustrated ane-but we made the point thatsome supplementary side lighting is then most likelyto be required. We have illustrated this in Fig. 15which Mr Aldworth criticizes. Whilst we take his

point about ’key’, ’fill’ and ’backlight’ and three di-mensional objects, the fact remains that if the BBCspecification is to be met, the area about the neartouchline to the main camera position is importantand supplementary side lighting is necessary wherefour-corner towers provide the main lighting. Ifone wishes to argue that the BBC requirements placeundue emphasis on illuminance in the direction ofthe main camera, this could be the subject of anotherpaper. We accept Mr Aldworth’s point that the BBCdo not claim that their requirements constitute aprecise document, and we appreciate Mr Price’s com-ments on this point. It should not be forgotten, how-ever, that a four-corner system will require approxi-mately twice the total wattage of a side-mountedsystem for a given vertical illuminance. We agreethat the choice of floodlighting systems depends onmany factors and site restrictions, but we would notagree that no useful purpose is served in discussingdifferent systems. _

Lens flare is a problem and must be born in mindparticularly when designing side lighting. I findMr Aldworth’s proposal for appraisal techniquesinteresting; this is a sound idea for designers whoshould appraise their own installations carefully andobjectively as we still have much to learn.

In reply to Mr Guha we have not tried the conceptof either scalar or cylindrical illuminance. We cal-culate ’vertical’ illuminance in four directions, notonly in the direction of the main camera postion, andhorizontal illuminance. Investigations into modellingenable the merits of these two concepts to be assessedas possible criteria in the design of floodlighting forsport. Neither side-mounted nor four-corner schemescome very close to daylight which, at one extreme,may be very diffuse and resemble an overall luminous

ceiling, and at the other may have a strong directionalcomponent from the sun. A side-mounted schemewhich is biased on the camera side per-haps give lighting closest to daylight.

°

The provision of a relatively high illuminance at theentrance for spectators is sound and we would al-ways recommend such practice. It should also be

applied to the players’ exit on to the pitch.

Mr Bownes makes out a case for six towers. This isa possible solution and an improvement on a four-corner system as the centre tower on each side willconsiderably help to meet the requirements of vertical

illuminance. In this way it resembles a side-mountedscheme.

We would not agree that it is impossible to light satis-factorily for the camera, spectator and players. Thereare obvious difficulties, and the answers will not al-ways be perfect. Each ground will have its own prob-lems and the final result depends upon the skill ofthe lighting engineer, who has to make judgements onsite which in practice may show departures from thetheoretical aiming plan in order to overcome a glareproblem. Thus the illuminance may fall or increaseslightly in one area or another when compared withthe calculated values. We are to some extent in a newworld in stadium floodlighting for colour television,but there are already practical installations whichare giving very fair results.

B. C. F~ogers (in reply): In reply to Mr Parker the10 kW metal halide lamp control circuit illustratedemploys a constant wattage ballast. It is not essentialto employ such a circuit to operate this type of lamp,but there are advantages with constant wattage opera-tion. These are (a) a change of supply voltage pro-duces little change in light output, (b) the chances ofoverloading the lamps because of excessive supplyvoltages are reduced, and hence a longer life isachieved, and (c) the cost of such control gear is littlemore than for other types since the lamp arc voltageof about 550 V is well in excess of the supply voltageand a means of stepping-up the supply voltage isrequired.The colour of the light from some types of metalhalide lamps does vary slightly with lamp wattage.For example, in a typical series inductance lampcircuit a 1% increase in supply voltage will produce a1. 8’70 increase of lamp watts and about 20 K decreasein colour temperature. This colour change has notbeen a problem in practice and does not warrant theuse of constant wattage control gear on any lamptypes known to us.

Mr Hampton criticized the idea of specifying parti-cular values of light output of lamps and maximumnumbers of luminaires employed. It must be remem-bered that the lighting installations are large, andelectrical and mechanical installation costs can beof the same order or greater than the lighting equip-ment cost. It is our experience that the most econo-mic solution lies in minimizing installation as wellas equipment costs; generally a large number of low-wattage luminaires cost more to install and maintainthan a lesser number of high-wattage luminaires forthe same field illuminance. Each installation has tobe considered individually and use made, where pos-sible, of existing towers, stand roofs, cables, etc.,

:~< ~’. Davies (in repl~r): ~r Ackerman ¡ s comment thatthe three camera tube currents should be reasonablyequal, as well as exceeding 100 nA, is correct; it isdesirable to achieve good correlation between thespectral distribution of the light source and the in-dividual camera channel sensitivities. Colour lagwill be minimal with a lamp spectral distributionwhich results in equal currents in the three camerachannels. As Mr Ackerman points out, however, ~.compromise has to be accepted for sensitivity rea-sons.

Referring to Section 4.2 of the paper and the sensationof luminance difference, Bodmann found that the dis-cernible range of luminances between the border lineof blackness and discomfort glare with a background



201

luminance of 10 cd/M2 was from 0. 7 to 800 cd/m2;with a background luminance of 100 ed/M2 the rangewas 0. 5 to 4000 cd/M2. It could be that there is abasic difference between static and moving scenes,but it has to be recognised, as Mr Ackermann points

out, that a gamma of 1. 2 rather than ~.. fl is oftenpreferred by television viewers and the further shiftof luminosity sensation may perhaps be explained bysubjective impression brought about by the limitedluminance range.



Lighting techniques and associated - CiteSeerX

Documents