ILLUMINATION QUALITY METER

 

TABLE OF CONTENTS

 

I.                     General Description …………………………………………………………………….         2

II.                   To Operate ………………………………………………………………………………         2

III.                  To Get the Results ………………………………………………………………………         2

IV.               To Understand the Results  ……………………………………………………………     3

V.                 Details of Operation……………………………………………………………………..         3

VI.               Care and Feeding ..…………………………………………………………………….          4

VII.              The Computed Characteristics ………………………………………………………..          4

VIII.            Detection System  ………………………………………………………………………         7

IX.               Technical Details ………………………………………………………………………..    10

X.                   Applications ………………………………………………………………………………   11

 

 

 

 

 

 

 

 

I. GENERAL DESCRIPTION

The ILLUMINATION QUALITY (TM) meter:

Tomorrow's light-meter...eight "eyes"... instant analysis of illumination... twenty computed characteristics...a complete microcomputer... read-only memory... random-access memory... keypad control…Iiquid-crystal display and printer output... interface to large computer... rechargeable batteries... three days' normal use between charges... battery condition readout... one to one-thousand dynamic range...two-year conditional warrantee on parts and labor.

The Illumination Quality meter (the IQ meter) is a sophisticated, eight-eyed light meter embodying a full-fledged microcomputer and a large software memory to perform its computations.

The IQ meter sees, by looking directly at the lamplight, what you as an expert see by looking into the room illuminated by that lamplight.

With its eight detectors, the IQ meter can do what the most experienced and highly trained human observer cannot do: sense all the characteristics of a certain lamplight by merely looking at the lamplight.

The human observer needs the usual complex colored scene illuminated by that lamplight, to assess the many effects of the lamplight on the appearance of the illuminated scene. We call "the many effects of the lamplight on the appearance of the illuminated scene”, the characteristics of that lamplight.

The human observer can, for example, pronounce the color- attractiveness of the illuminated scene as "good" or "fair" or "poor." He can, by careful comparison of similar adjacent scenes illuminated differently, decide which one of these excels the other in color- attractiveness, if only by a bit. But two difficulties remain: (1) how to put a numerical value - a number - on the color-attractiveness of a certain scene, and (2) how to be sure that the "average person" will assess the color-attractiveness in about the same way.

Much of the value of the new IQ meter resides in the provision of several new scales of the kind alluded to in the preceding paragraph. These new scales (1) are numerical, (2) are representative of assessments of lamplight quality by "average persons”, and (3) are completely visually-based.

The IQ meter measures - puts a NUMBER on - twenty important characteristics of the lamplight.

You can HOLD IN YOUR HAND the instrument which gives all these answers.

It is simply a light-meter. All you do is push a button. But it measures 20 characteristics of the lighting, all the traditional ones, plus the important ones.

 

The important ones are:                (These are strictly visual                                                                                               characteristics of  lamplight)

·          Brightness

·          Color-Attractiveness

·          Visibility

·          Color-Scheme-Stability

·          Brightness per Footcandle

 

Additional important ones are: .

·          Color-Preference Index

·          Color-Gamut

·          Visibility per Footcandle

·          Visible watts per square meter

·          Violet watts per square meter

·          Ultraviolet watts per square meter

·          Visible microwatts per lumen

·          Violet microwatts per lumen

·          Ultraviolet microwatts per lumen

·          Brightness per UV microwatt per sq. meter

·          Visibility per UV microwatt per sq. meter

 

The traditional ones, also measured, are:

·          Footcandles

·          Chromaticity x,y

·          Color-Temperature

·          Color-Rendering Index

 

Description: The Illumination Quality meter, or IQ meter for short, is a portable device to sample the illumination in a particular area, and to compute and display a large number of qualities of that illumination. The IQ meter's ultra-sensitive Detector-Head consists of eight photodiodes each covered by a different optical filter, so that each is responsive to a different portion of the spectrum. Thus the signal from each diode represents the intensity of that portion of the spectrum to which it is responsive. The photodiode-filter combinations are arranged in an array and exposed to the illumination, which it is desired to analyze.

Each photodiode-filter combination generates a signal in response to its portion of the incident light. Each signal is amplified by an operational amplifier, and charges a capacitor. The analog voltage on the capacitor is converted to a digit by an analog/digital converter. The eight digits, representing the eight portions of the incoming light, are stored in the microcomputer memory. Mathematical functions are stored in the read-only memory; upon inserting the eight incoming digits into each of these functions, each function computes a value of a particular characteristic of the incident illumination.

The revolutionary ILLUMINATION QUALITY meter measures the old-fashioned qualities of illumination: footcandles, color-temperature, color, and color-rendering index. More importantly, the IQ meter also measures the visually-based qualities: color-attractiveness, color-scheme-stability, perceived brightness and visibility. These scales are derived from judgments by experienced observers assessing each quality under different kinds and intensities of lamplight.

Put the IQ meter in whatever illumination you wish to know more about, touch the 'read' button, and give it a second to sample the light through its eight eyes. By that time it has memorized the eight inputs, and already computed the many qualities of the illumination. Touch more buttons and call out the illumination qualities one by one to the liquid crystal display, or print them on an included printer.

The IQ meter puts numbers on both the hard-nosed and the esthetic qualities of your lamplight or lighting installation, or one by your competition. What makes it pleasing? What makes it successful? It measures about all there is to know about illumination. It displays the secrets of lamplight. The IQ meter substitutes for a group of friendly and talented experts, which you don't always have at your elbow.

 

II. TO OPERATE

 

1.          Turn the switch on. The IQ meter display will say "ready”.

2.          Set the IQ meter in the lamplight you wish to measure, at the location you wish to measure it. The Detector-Head, at upper right of the top surface of the meter, is the light-sensitive area. The eight "eyes" are under the white, scattering layer. Avoid casting any shadow on the IQ meter.

3.          Touch the "read" button (R). IQ meter display says "testing”. ("Touch" means touch lightly, but hard enough to activate the switch beneath the key.)

 

 

III. TO GET THE RESULTS

 

·          In a few seconds the IQ meter will give an audible signal that it is finished looking at the lamplight.

·          The display will say "done testing," or will show one of the computed characteristics of the lamplight.

·          All twenty characteristics have now been computed.

·          To call them out of memory:

 

Touch "C," and then "1," to get C1 = Color Attractiveness, for example. C1 is the call-code of Color Attractiveness. There are twenty call-codes in all.  The computed value of C1 will appear on the display.

 

V. TO UNDERSTAND THE RESULTS

 

·          Allow yourself a comfortable learning period to become familiar with the useful scales of units.

·          Each one of the twenty computed characteristics of lamplight has a useful scale of units.

·          By all means take the IQ meter out and use it. But realize that the enormous amount of information you now have at your fingertips needs a great deal of study - by you - to get out of that information all of the usefulness packed into it.

·          A computer disgorges more information than one can at first handle, or assimilate. Uncomfortable feeling. But one can learn gradually to use the power of the computer. At first you may feel you have "a bull by the tail"; later, you will feel in complete command. The computer in the IQ meter can do a hundred thousand calculations in one second. One can learn to use this power. Pressing the button and letting the IQ meter observe the illumination for a second or two, essentially 'takes a snapshot' of the lighting conditions. Then, with all the necessary information stored in the computer's memory, one can (at his leisure) ask for, one by one, the numbers describing each of the separate qualities of the illumination which was present at the time the button was pressed. At any one time, and in any given illuminated place, some of these qualities are more important than others. So do not feel pressed to learn immediately the role each of the twenty characteristics plays in modern lamplight; on the other hand, beware of falling into the sheepish habit of leaning on the old, time-worn and unreliable characteristics just because they are somewhat familiar.

·          It is almost impossible to harm the computer during normal operation. Touching wrong buttons will do no harm. If the display shows something incomprehensible (probably a procedure used during manufacture or calibration of the IQ meter) simply turn the switch off, and then back on. This clears the IQ meter's memory and returns it to the "ready" state. If one gets mixed up in asking for all the available information from the IQ meter, it is only necessary to start analysis over again; not to take the 'snapshot' again. Or, one can take the snapshot over again if he wishes.

·          It is very informative to choose two rooms side by side, preferably lighted by very different kinds of lamps (for example, standard fluorescent lamps and prime-color fluorescent lamps); then sample the illumination in one space, leisurely determine its many qualities, whether good or bad. Now sample the illumination in the other space and compare each quality with the corresponding one in the first space. Make careful comparisons between what you see, and what the IQ meter says you should be seeing.

·          Study the meanings of each of the twenty computed characteristics (Section VII).

·          You may find that, at first, you may wish to concentrate mostly on a few particular characteristics, becoming thoroughly familiar with their units and scales.

·          A wise choice, to begin, is "Brightness" and "Color-Attractive- ness."

·          Note carefully how the value of a certain computed characteristic corresponds to what you yourself see.

·          Begin immediately to build up a personal sense of, and confidence in, what the computed characteristics mean, in terms of quality and quantity of lighting.

 

V. DETAILS OF OPERATION

 

1.        To choose a particular characteristic of the lamplight, to be read out immediately and automatically after every reading: (a) Touch "*98,” (b) then the call-code of the characteristic desired. This sets what will be the first display.

2.        To print on the printer whatever is displayed: Touch "#".

3.        To print out all twenty of the computed characteristics to the printer: Touch "*96".

4.        The IQ meter "looks at" the illumination falling upon its Detector-Head for only a few seconds after the read-button R is touched. That is, the eight eyes are operational only during the read-period. In that short time, it has gathered and stored all it needs to know. At any other time, whatever illumination may be falling on the Detector-Head is ignored.

5.        Therefore, carrying the IQ meter around will not change the characteristics last-computed.

6.        When you want to recheck a result, to analyze a new or different illumination, or a different brightness of the same illumination, or the illumination in a different location, simply put the IQ meter in the proper place and re-read, (R). This changes the remembered values of the characteristics, so be sure you have written down, or have a print-out of, or don't care about, the last-computed characteristics. On the other hand, it is usually simple to re-read a certain illumination, if the information has been lost.

7.        The characteristics last computed are stored in memory, and remain there until R is touched again, or the meter is turned off. A new reading wipes out the old one.

 

VI. CARE AND FEEDING

 

·          A little care and a little feeding of your IQ meter will give you years of useful service from it.

·          Carefully protect its Detector-Head. Its eight eyes, although they are rugged, need to be treated like the sensitive instruments they are - like any eyes. Let nothing but LIGHT impact the Detector-Head.

·          As for the rest of the IQ meter, treat it like any valuable computer. Don't knock it. Don't subject it to extremes of temperature or humidity. Don't leave it in the sun. Don't leave it in the car. Protect it from dust and dirt. Turn it off when not in use. Put it in the case when not in use. Bear in mind that disassembly voids the warranty.

Feeding:

1.        To display battery voltage: Touch "*91."  It should be between 8 and 9 volts. If it is lower, charge for an hour or more. Check it daily.

2.        To connect charger: Plug charger into IQ meter, THEN into wall outlet.

3.        Do not charge overnight. Do not over-charge.

4.        To display "low-battery threshold" setting: Touch "*92." This is the battery voltage below which, during operation, the IQ meter will automatically complain of low-voltage. It should be set normally to about "7.5." When the display shows "battery low," charge the IQ meter when convenient.

5.        If "battery low" is displayed, and you wish to take a few more readings, reset the threshold below the present "battery volts" reading. You may then make a few more readings before recharging.

6.        To enter "low-battery threshold" setting: Touch "*90." Enter "7.5," for example.

 

VII. THE COMPUTED CHARACTERISTICS

 

A. Introduction

The usual light-meter outputs a single number. Sometimes, but not always, that single number gives the user an accurate measure of how bright a space appears to the average person. Or how well one can see in the space. Or what camera exposure to use. The meaning of its scale of units must always be carefully and patiently learned. If the meter is a "footcandle meter," experience teaches the user this sort of thing: (1) a reading of 1 footcandle warns that some difficulty probably will be encountered in reading a newspaper; (2) a reading of 50 footcandles is likely to be comfortable; (3) a reading of 500 footcandles will sometimes seem too bright. Not much more than that can be inferred from reading a footcandle meter. The footcandle meter is a very inadequate stand-in for human vision, or for the human observer. Even so, its scale of units takes time to understand and use.

The footcandle-meter has a single eye, as does the light meter on a camera. With a single eye, a light meter can distinguish lightness and darkness - nothing more. What a single eye senses can be described with a single dimension, like inches on a yardstick.

The IQ meter, on the other hand, has eight eyes.  Eight independent inputs. The eight eyes are sensitive to different colors of light; that is, to different parts of the spectrum (Figure 1). (This feature of the IQ meter is described in detail in Section VIII.) The eight eyes respond to: Ultraviolet, violet, blue-violet, blue-green, green, yellow, orange-red, and deep red, respectively.

This means that the IQ meter can measure eight dimensions of lamplight. Most of us cannot imagine what it means to measure, and then visualize, more than three dimensions - length, width and height of an object, or hue, saturation, and brightness of a light. So the remaining five dimensions sensed by the IQ meter remain outside our ability to visualize.

However, what we can do is relate those eight dimensions of lamplight to some complex characteristic of the lamplight, for example BRIGHTNESS, which we can understand.

 

This is the second step - the computation step - performed by the IQ meter.

 

B. Example: Color-Attractiveness

 

1. Computation.

It is important to grasp the fact that most types of lamplight are mixtures of colored lights; in fact, most lamplights are mixtures of lights of all colors, even though the lamplight usually appears white to the eye. It is the composition of the lamplight, the varying amounts of these colored lights that make up the lamplight, which is responsible for the great differences we note from one lamplight to another.

Immerse the IQ meter in a certain lamplight we will call "lamplight A.” Touch the read-button. It "looks at" lamplight A for a few seconds. During this time, each eye accumulates an electric charge, proportional to the intensity of that segment of the lamplight to which that particular eye is sensitive. For example, the blue-violet- sensitive eye accumulates a charge proportional to the amount of blue-violet light present in lamplight A, and similarly for the other seven eyes. The electronics in the IQ meter convert the eight electric charges to eight proportional numbers. Suppose the resulting eight numbers, accumulated by the eight eyes, and characteristic of a certain intensity of lamplight A, are:

 

   UV     V     BV     BG      G      Y     OR     DR

  ___   ___   ___    ___    ___   ___   ___    ___    (Lamplight A)

 

   31            73   122    158    188    245   221   132

 

The IQ meter automatically commences to make a long, complex calculation of each of the twenty characteristics of that certain intensity of lamplight A.  The calculation is guided by the eight numbers 31, 73, 122, 158, 188, 245, 221, 132, which it accumulated while looking at lamplight A. In perhaps a second, the twenty calculations are finished, and stored in its memory.

Now touch C1 (Color-Attractiveness). Out comes the number 613 to the display. As you will learn, a Color-Attractiveness of 613 is reasonably good.

Now repeat the entire process with the IQ meter immersed in a certain intensity of lamplight B, resulting in:

 

   UV     V     BV     BG      G      Y     OR     DR

  ___   ___   ___    ___    ___   ___   ___    ___    (Lamplight B)

 

   20            51   102    141    194    169   251   158

 

This time, each eye has accumulated a little different number. The blue-violet-content (and each of the seven other "contents") of lamplight B is at least a little different from that of lamplight A. All the calculations are redone. Upon touching C1, out comes the number 842, a value of Color-Attractiveness representing a lamp-light which is very good indeed.

The IQ meter has transformed the eight measured dimensions of the lamplight into a point on a simple scale (yardstick) of Color- Attractiveness which we can understand (Figure 2).

It does this by comparing its eight readings to a data-bank on many other lamplights of widely-varying characteristics, all stored in its memory. It has done this for all twenty lamplight characteristics, only one of which we have as yet called for.

 

2. The "Color-Attractiveness" Scale.

Let us take for concreteness the scale of color-attractiveness, and examine the means by which it was arrived at.

Sixty or seventy real lamps were assembled, including most of the commercial lamp-types on the market today (incandescent, fluorescent, and high-intensity-discharge lamps) and in addition many experimental lamps with properties both considerably superior to, and considerably worse than, those of actual marketed lamps.

Five identical enclosures were provided, with identical arrays of familiar colored objects such as fruit, vegetables, meat, bread, butter, well-known examples of commercial packaging, metallic objects. The geometry of the lighting was carefully controlled to be constant despite changes in type of lamp.

Each of the five enclosures was illuminated by a lamp with a

different spectral power distribution, i.e. by a different lamp-type. Brightness in each enclosure was adjusted to be approximately the same for all enclosures. All five enclosures were viewed simultaneously by the experienced human observer. The observer was required to rank the color-attractiveness of the arrays of objects each in its particular lamplight, and to assign each enclosure (Iamp type) a number from one to ten, the poorest being assigned the number one, and the most attractive coloration the number ten. One set of five lamp-types might be assigned the numbers: 3, 9, 10, 1, 2. Another set might be assigned the numbers: 8, 1, 3, 10, 5. Once the individual observers had made their assignments, two or more observers were asked to rank the five lamplights by consensus.

Next, three of the lamp-types were changed, always leaving two lamps from the previous session, but randomized in position. The observations and ranking were carried out as before, and the process repeated until all lamp-types had been compared in groups of five.

The assessments of the single observers were remarkably alike. Consensus was in most cases easy to agree upon. It was therefore apparent that the "average person" is well defined, as regards his criteria for assessing attractiveness of coloration of an illuminated scene.

The averaged magnitude-estimations were interleaved (the reason for having two lamp-types common to sets of five in successive sessions). The result was a scale, in which each lamp- type was assigned a certain rank, and in which the scale-distance between adjacent lamps was not constant; i.e. the scale-distance between types ranked 63 and 64 would probably be different from the distance between types ranked 42 and 43. To the worst lamplight was assigned "color-attractiveness = 1" and to the best, "color-attractiveness = 1000." Most future lamps will undoubtedly fall in the range of color-attractiveness of 1 to 1000. Hopefully a lamp of the future will exceed a value of 1000 (which will introduce no problem, either to the IQ meter or to conceptualization by the user), and also hopefully no commercial lamp will require a negative value.**

"Color-attractiveness," arrived at in the above manner, fulfills the three criteria: it is numerical, it represents the "average person," and it is completely visually based.

 

3. Other Scales.

"Color-scheme-stability" was arrived at in essentially the same manner. Instead of foods, etc, identical arrays of fabrics were used, with the obvious criterion of degree of upset of the color-scheme with change in lamp-type.

The scale of "Brightness Units" was found in a conceptually somewhat simpler manner. Only two adjacent, identically furnished enclosures, each with a different lamp-type, were used. Since neither quality nor trueness nor preferredness of coloration was the criterion, but rather the perceived brightness of the scene, the nature of the colored objects in the identical scenes was presumed to be irrelevant. The footcandle-level in both enclosures was initially set to 100. Then the brighter enclosure was dimmed until the perceived brightness in each enclosure was pronounced the same. At that point, the footcandle levels were again read and their ratio taken. For each enclosure (lamp-type) a relative perceived-brightness-per-footcandle value (B/FC) was thus obtained. For the yellow illumination of the "Gold" fluorescent lamp, perceived-brightness-per-footcandle was arbitrarily set at "1.00,” and the values for all other lamp-types related to that. As examples, the B/FC values for a few familiar commercial lamplights are given in the Table below:

 

Table I

 

Lamplight	B/FC
Warm White fluorescent	0.9
Cool White fluorescent	1.0
Daylight fluorescent	1.1
North light	1.4

 

 

 

Knowing the B/FC values for all of the lamps, one has only to multiply the measured footcandle values by the appropriate value of B/FC to obtain the correct number of "Brightness Units."

The scale of "Visibility Units" was found in the above manner. However, the brighter enclosure was dimmed not to equal bright- ness, but to equal "visibility." This criterion is a little harder to define than "brightness." It involves the ability to see the elements of a scene clearly. In many cases, for example, the observer dimmed the one enclosure much farther than he would have, had the criterion been "equal brightness." It points up the fact that "brightness," although a more useful unit than "footcandle level" with which to correlate the ability to see well, yields to the concept of "visibility."

Picture oneself in a boat on a lake on a calm morning, in a thin, sunlit fog: there may be a great deal of brightness, but one is able to see very little: the "visibility" is low.

In everyday human visual experience, it is necessary to distinguish between, and to do the best we can to measure, "bright- ness" and "visibility."

In the experiments being described, consensus was arrived at in the matter of equality of visibility, although considerably more roughly than in the case of color-attractiveness. Then, a value of visibility-units-per-footcandle was arrived at for each lamp. As before, for the yellow illumination of the "Gold" fluorescent lamp, visibility-per-footcandle was arbitrarily set at "1.00,” and the values for all other lamp-types related to that. Knowing the V/FC values for all of the lamps, one has only to multiply the measured footcandle values by the appropriate value of V/FC to obtain the correct number of "Visibility Units."

These new visual scales are the subjects of continuing research, for they must be determined with increasing precision. Nevertheless, already their validity far surpasses anything heretofore available for the quantitative assessment of lighting quality.

C. The Twenty Measured Characteristics of Lamplight

1. The FOOTCANDLE.

The footcandle is an illuminance of one lumen per square foot incident upon a surface. One manner of defining the "lumen": In one watt of pure yellow-green light of wavelength 555 nm, there are 683 lumens. If that one watt of light were to fall on one square foot of white paper, the resulting 683 footcandles would appear bright yellow-green to the normal observer (average person).

The IQ meter reads footcandles as does the footcandle meter, but with the increased accuracy deriving from its eight eyes, rather than from the footcandle meter's one eye. Number of lumens = number of watts x 683 (lumens per watt for the spectral color 555 nm) x the luminous efficiency of the light. Luminous efficiency is luminous flux/radiant flux. Radiant flux is the integral of the spectral power distribution, in watts of light incident per unit area; luminous flux is the integral of the spectral power distribution weighted by the photopic spectral luminous efficiency function V-lambda, variously called the 'luminosity function,' the 'visibility curve,' etc. The foot-candle meter 'has one eye'; its sensitivity is defined by the photopic spectral luminous efficiency function V-lambda. Footlambert: an equivalent footcandle; one lumen per square foot will be reflected from a perfectly white, perfectly diffusing surface upon which one footcandle (one lumen per square foot) is falling. A surface is said to have a luminance of one footlambert if it is emitting one lumen per square foot.

The user of this manual is advised to take great care to distinguish carefully between perceived brightness, and either the lumen, the footcandle, or the footlambert. None of these latter units can substitute for what is normally called "brightness," and untold confusion and distress have resulted over the last sixty years from  the attempt. The footcandle does not correlate well to what people see as 'bright.'

To compute footcandle level, inputs from the IQ meter's eight eyes are properly weighted according to the prescribed photopic relative luminous efficiency function V-lambda, and the footcandle level incident on the detector-head is accurately computed.

 

2. The BRIGHTNESS UNIT.

The brightness unit is contrasted to the footcandle by visual experiment. Many types of experiments require a trained observer with normal vision to adjust two observed lights, or illuminated spaces, so that their brightness is perceived to be the same.*

Once the observer pronounces the lights to be equally bright, the footcandle (or footlambert) level is then measured, and the ratio B/L of perceived brightness per unit of illuminance, or luminance, is determined. Over the years, thousands of data of this type have been amassed; it has been found that (1) observers with normal color vision agree adequately well on B/L for a given light, and (2) B/L varies a great deal among lights of different color or at least of different spectral power distribution. For our purposes, in relation to the IQ meter, 'normal' yellow light, such as incandescence through a yellow filter (yellow traffic light) is taken to have a brightness-unit- per-footcandle or brightness-unit-per-footlambert ratio of 1.00.

Many published results on a large number of illuminants of varying spectral power distribution contribute to the function, stored in the IQ meter's memory, by which brightness units are computed from the eight inputs. Chapanis and Halsey** measured B/L for several hundred colored lights; they showed that pure blue light has about ten times the brightness units per footcandle as does normal yellow light, and pure green light and pure red light have about twice the brightness units per footcandle as does yellow light. More interesting still, light of the same color may have quite different brightness-per-lumen depending upon its spectral power distribution; this is a natural consequence of the three independent, and spectrally different, inputs of the human visual system.

3. BRIGHTNESS-PER-FOOTCANDLE.

Brightness-per-footcandle was determined, as described above, for a large number of types of lamplight. It is simply the quotient of computed brightness units and computed footcandles, for any illumination sampled and computed by the IQ meter. Example: For 5000K filtered incandescence, brightness-per-footcandle will appear on the IQ meter readout as 1.3; that is, such illumination provides 30% more brightness units than it does footcandles. Alternatively stated, 100 footcandles of that illuminant provide the same perceived brightness as do 130 footcandles of normal yellow light (for which B/FC = 1.00).

 

4. The VISIBILITY UNIT.

The visibility unit is also compared to the footcandle by visual experiment in exactly the same manner as is the brightness unit. Here, rather than setting to equal perceived brightness, the observer adjusts the light level in two illuminated spaces so that he can see equally well, equally clearly. As mentioned above, the visual characteristics of illumination here called Brightness and Visibility are not at all the same. A room can seem full of light, and yet one may not be able to see particularly well. Visual experiments on visibility units are ongoing; results collected so far are incorporated in the visibility function, stored in the IQ meter's memory, by which visibility units are computed from the eight inputs.

5. VISIBILITY-PER-FOOTCANDLE.

Visibility-per-footcandle was determined, as described above, for a large number of types of lamplight. It is the quotient of computed visibility units and computed footcandles. Example: For the Cool White Deluxe fluorescent lamp, visibility-per-footcandle will appear on the IQ meter readout as 2.4; that is, such illumination provides 2.4 times as many visibility units as it does footcandles. Alternatively stated, 10 footcandles of Cool White Deluxe illumination will provide the same visibility as will 24 footcandles of yellow light such as that from the 'gold' fluorescent lamp.

 

 

6,7,8. VISIBLE MILLIWATTS PER SQUARE METER, VIOLET MILLIWATTS PER SQUARE METER, ULTRAVIOLET MILLIWATTS PER SQUARE METER.

The approximate wavelength bands which are scanned by the IQ meter for visible, violet, and ultraviolet content of the illumination are shown in Figure 1. For example, 100 footcandles of light from an incandescent lamp carries with it about 3.5 visible watts per square meter, 110 milliwatts of violet light, and about 20 milliwatts per square meter of ultraviolet, if the light-source is an unshielded quartz-tungsten incandescent lamp. Common household bulbs will emit a smaller proportion of ultraviolet radiation.

Bear in mind that lamplight may contain radiation outside the range of wavelength covered by the IQ meter. For example, lamplight from the incandescent lamp contains a great deal of infra-red radiation, at longer wavelengths beyond the deep-red. Such radiation plays no part in the visual processes of interest to us, and therefore the IQ meter has not been equipped to detect it.

9,10,11. VISIBLE MICROWATTS PER LUMEN, VIOLET MICROWATTS PER LUMEN, ULTRAVIOLET MICRO- WATTS PER LUMEN.

In a certain lamplight, once the visible watts per square meter and the footcandle level are measured, the quotient (times a constant) yields the visible microwatts per lumen; similarly for violet and ultraviolet "content" of a lumen of that lamplight. For the museum conservator, for example, the measurement of ultraviolet microwatts per lumen relates to the desired balance between the opposing requirements of (1) visibility to the observer of a precious article, and (2) its decomposition by incident light. Example: Stan-dard Cool White fluorescent light contains 60 microwatts of ultra-violet per lumen, while clear mercury HID contains 680 and one variety of 5000K filtered incandescence only 11.

12,13. BRIGHTNESS UNITS/UV MILLIWATT/ SQUARE METER, VISIBILITY UNITS/UV MILLIWATT/ SQUARE METER.

In a certain lamplight, once the brightness units and the ultraviolet milliwatts per square meter are measured, the quotient yields the number of brightness units per ultraviolet milliwatt per square meter. These are also units of intense interest to the museum curator, as are visibility units per ultraviolet microwatt per square meter of museum area, for example. Examples: clear mercury HID supplies 0.3 visibility units per ultraviolet milliwatt per square meter, Cool White fluorescent 1.2, and the 5000K filtered tungsten incan- descent light supplies 2.1 units.

5000K prime-color fluorescent lamplight supplies about 30 units, both because ultraviolet content is lower and because visibility per footcandle is higher.

14. CHROMATICITY.

Chromaticity x and yare coordinates on a "color map" (Figure 3) called the 1931 CIE x,y chromaticity diagram. Coordinates x and y are computed by the IQ meter for the incident lamplight. As examples: Lamplight from the ordinary incandescent bulb has approximately the chromaticity x = 0.448 and y = 0.408. That of the standard Cool White fluorescent lamp is about x = 0.367, Y = 0.384. That of the Gold fluorescent lamp is about x = 0.517, Y = 0.479. The chromaticity of a lamplight will not actually indicate its apparent color; recall that an incandescent lamp inside a home, but observed from outside in average daylight, appears very yellow; but after dark from inside the home the same lamp appears quite white. Two types of illumination may of course have the same chromaticity but be very different in their other characteristics. For example, average daylight (6500K) and the standard Daylight fluorescent lamp both have a chromaticity of about x = 0.315, Y = 0.335, in the 'white' region of the diagram; their color-rendering characteristics are however quite different.

 

 

 

 

 

 

 

 

 

 

* A list of thirty-three published references, documenting the general problem of visual determination of the relative brightness of different lights, will be found in "Brightness Meter,” Journal of the Illuminating Engineering Society, October 1980.

**A. Chapanis and R.M. Halsey, "Luminance of Equally Bright Colors,” Journal of the Optical Society of America, Vol. 45, No.1, January 1955, p.1.

 

 

15. COLOR-TEMPERATURE.

"Color-temperature" is a convenient way of specifying the color of certain familiar and useful lights. The white light from an overcast sky is one of these; it is one of a unique series of natural lights, characteristic of incandescent glowing bodies like the sun. At one end of this series of lights is the glow from a red-hot poker in a blacksmith's shop. "Red-hot" corresponds to a temperature of about 1000 Kelvins (1300 degrees Fahrenheit); the reddish color is characteristic of an incandescent body at 1000 Kelvins. The in- candescent tungsten filament in the household lightbulb, near 2800 Kelvins, glows with a yellowish light. Sunlight has the color of a glowing body at about 5000 Kelvins, and blue sky has the color of a glowing body at some very high temperature, say 100,000 Kelvins. Each temperature in Kelvins has associated with it a color: red, yellow, white, blue-white, as temperature increases. So we have a scale of "color temperatures"; the color temperature of the sun is about 5000K because the color of its light when it reaches us is that of an incandescent body at 5000K. The various phases of daylight are all members of the series of "natural" lights, to each of which is associated a color temperature. Light from a yellow sunset may have a color temperature of 2000K, from a warm-looking cloud in the sunlight 4200K, from an overcast sky 6500K. Note that there is nothing in the "warm-looking cloud" that is anywhere near 4200K in actual temperature; it just shares the color of an incandescent body which is glowing at that temperature. Light from the commonest of fluorescent lamps has a 4200K color temperature, but the lamp is hardly warm; light from the glowing embers in a fireplace has a 1000-2000K color temperature, and in that case the embers are really that hot.

The curve within the color diagrams of Figures 3 and 4 is the locus of chromaticities of the unique series of natural lights, which include phases of daylight, and light from incandescent bodies. On this curve lie the chromaticities (colors) with each of which is associated a color-temperature in Kelvins.

What the IQ meter actually calculates, as defined for the 1960 CIE u,v color diagram (Figure 4), is known as the "correlated color temperature" of the lamplight. Many lamplights are either a bit greener or a bit more purple than the "natural" lights of the preceding paragraph, and so do not really have a "color temperature.”  Because the unit is a useful one, although very approximate, correlated color temperature is commonly specified for lamplights. It is the color closest to the color of one of the natural lights. Because the definition of "closest to" depends upon the particular color-diagram used (and there are many in use for different purposes), the specification of "correlated color temperature" is, for many types of lamplight, rather sloppy. Choice of another 'official' color diagram (there are difficult problems with the present choice - or with any choice - of color diagram) is a possibility that is currently being considered. A change would make significant shifts in the assigned correlated color temperature of any lamplight differing from the natural series. Correlated color temperature of the sodium lamp is about 2000K; the Warm White fluorescent lamp is about 3000K; 'Cool White' about 4200K; sunlight about 5000K; overcast sky about 6500K.

16. COLOR-RENDERING INDEX.

Color-rendering index is a measure of the similarity of color- rendering of a lamplight to that of a reference phase of daylight closest to it in correlated color temperature. The CIE defines eight test-colors by their spectral reflectance curves. The eight chromaticities of these test-colors, as rendered by the lamplight, are plotted in the u,v color diagram (Figure 5). The eight chromaticities as rendered by the proper reference illuminant (which we can take as a phase of daylight closest in color to that of the illuminant) are also plotted. If the rendered chromaticities by reference and test illuminants exactly coincide, CRI = 100. If they do not, i.e. if some or all of the test-colors are shifted by the test illuminant, the average color-shift is computed, and the CRI reduced. CRI = 50 for the standard Warm White fluorescent lamp (rather poor color-rendering), and can even go negative for illuminants like low-pressure sodium. Two of the main problems with CRI are: (1) If CRI is appreciably less than 100, there is no way of guessing whether coloration is pleasant and acceptable, or unpleasant and unacceptable. As one example, CRI = 60 for both the metal-halide HID lamp and for the prime-color HID lamp; the colors of fruit, vegetables, complexions, etc., appear pale and washed-out under the former lamp, and colorful and vibrant under the latter lamp. (2) CRI of all reference illuminants is, by definition, 100; but color-rendering by blue sky is very different from that by the light from a yellow sunset, although for both CRI = 100, and both are prescribed reference illuminants.

17. COLOR-GAMUT.

Color-gamut is simply the area on the u,v color diagram enclosed by the eight-sided figure (Figure 5) formed by the chromaticities of the eight test-objects as rendered by the test illuminant. Normalized to 100 for average 6500K daylight. The larger the gamut, the more saturated all colors will tend to appear. CG = 60 for the standard Cool White fluorescent lamp, 100 (as large as that for daylight) for the Cool White prime-color fluorescent lamp. See "Color-Discrimination Index," Journal of the Optical Society of America, February, 1972.

 

18. COLOR-PREFERENCE INDEX.

Computed somewhat similarly to the CRI, except that the target, or ideal, chromaticities of the test-objects are not those as rendered by some reference illuminant, but rather those which the normal observer wishes to see. As prescribed by D. B. Judd of NBS in 1967, based on a collection of results of visual experiments on choice of colors for familiar objects. The computed CPI of an illuminant agrees well with the ranking of that illuminant as to acceptability of the coloration it gives to a scene. For average daylight, CPI = 100. Pure prime-Color illumination reaches a CPI of 120 or so. For deluxe fluorescent lamps CPI ranges 80-100. For standard fluorescent lamps, CPI is 50-70; for clear mercury HID, about zero; lower for the sodium lamps. See Deane B. Judd, "A Flattery Index for Artificial IIluminants," Illuminating Engineering, Volume 62, page 593, October, 1967; and "Customer Acceptance of the Color-Rendering of Illumination," Lighting Design and Application, March, 1982.

19. COLOR-ATTRACTIVENESS.

A visual scale arrived at by side-by-side comparison of coloration given to identical arrays of familiar objects by sixty widely varying illuminants. The illuminants were compared four or five at a time, and the results interleaved to form a continuous scale running from one to about 1000 arbitrary units. Examples: For the incandescent lamp, CA = 823; Cool White Deluxe fluorescent 680; standard Cool White 430; clear mercury HID 60.

20. COLOR-SCHEME STABILITY.

The interior designer chooses his color-scheme in excellent quality illumination. He then hopes his color scheme survives a change of lamplight. A visual scale was arrived at in the same way as for color-attractiveness, using color-schemes of fabrics as test- arrays. Visual experiments on color-scheme stability are ongoing; results collected so far are incorporated in the color scheme-stability function, stored in the IQ meter's memory, by which CSS units are computed from the eight inputs. CSS = 700 for the incandescent lamp and 450 for the standard Cool White fluorescent lamp. See "The Design of Safety Colors," Journal of the Illuminating Engineering Society of North America, January, 1977, for related discussion.

 

VIII. DETECTION SYSTEM

Of the illumination falling on the IQ meter, the wavelength region from 250 nm in the ultraviolet  to 680 nm in the deep-red is analyzed by the IQ meter (Figure 1). Of the eight eyes, or detectors, one is sensitive to the ultraviolet region, one is sensitive to the violet region, and the remaining six detectors cover six precisely defined regions of the easily-visible region of the spectrum as follows: blue-violet, blue-green, green, yellow, orange-red, and deep-red (Figure 1).

Note that the familiar primaries - blue, green, and red - are not appropriate divisions of the spectrum where human vision is concerned. Nor where the IQ meter, whose objective is to substitute for the human visual system, is concerned.

 

 

Figure 1. The spectral responses of the IQ meter. Top curve: The spectral power distribution of sunlight. Watts per nanometer wavelength interval (the power of sunlight in a narrow vertical slice) versus wavelength in nanometers. Middle curves: The spectral responses of each of the eight "eyes" of the IQ meter. No.1 is the special ultraviolet-sensitive detector; nos. 2 -8 are sensitive in precisely-defined bands in the visible region of the spectrum. Bottom: The designation of the eight bands by name.

 

Figure 2. The "yardstick" representing the Color-Attractiveness scale. The computed values for lamplights "A" and "B". The IQ meter transforms the hard-to-visualize eight measured dimensions of the lamplight, to a single dimension (for example Color-Attractiveness) which is easy to understand.

 

 

Figure 3. Definition of "color-diagram," "chromaticity," and "color-temperature." The 1931 CIE x,y color-diagram: Periphery is the locus of chromaticities x and y of the spectral colors (pure, brilliantly-colored lights corresponding to narrow vertical slices of Figure 1); a few wavelengths are labeled in nanometers. Inner curve: Locus of chromaticities of natural lights of different color-temperature; i.e., of color corresponding to that of an incandescent glowing body at some Kelvin temperature. The perceived color of a body at 1OOOK (one thousand Kelvin) is red, 2000K is orange, 3000K is yellowish, and at higher temperatures the perceived color becomes white and then blue-white; the observer is presumed to be adapted to average daylight.

 

Figure 4. The 1960 CIE u,v color-diagram; a simple transformation of the diagram of Figure 3.

 

 

 

 

 

 

 

 

 

 

Figure 5. The 1960 CIE u,v color-diagram; a small area in the neighborhood of the chromaticity of average 6500K daylight D65 (spot at the center of the pattern). Also shown are the eight chromaticities of the CIE test-colors as they are rendered by the 6500K daylight. A species of lamplight of the same color, and of the same chromaticity (spot at the center of the pattern), would probably render the same eight test-colors rather differently. Often, but fortunately not always, artificial lighting renders colors paler than does daylight, so their rendered chromaticities would generally fall inside the daylight-pattern. Color-Rendering Index would be less, and Color-Gamut would be smaller (see text).

Sunlight, with the spectral power distribution of Figure 1, is composed of power distributed quite evenly at all wavelengths through the region to which the IQ meter is sensitive. The exception is the "short-ultraviolet" region, where the atmosphere subtracts much of the short-ultraviolet radiation from the sunlight. Artificial illumination of some kinds does, however, contain power at short- ultraviolet wavelengths. There are a number of important reasons why short-ultraviolet radiation should be monitored, and in many applications guarded against. For example: (1) It can cause painful, although temporary, damage to the eyes. (2) It will fade and decompose many materials such as paint, dye, ink, cloth, paper.

Number-one detector responds to the band of wavelengths called the ultraviolet region, including both "short-ultraviolet" and "long-ultraviolet" sectors. Because of this sensitivity, the IQ meter responds to tanning wavelengths, in both short-ultraviolet and long- ultraviolet regions of the spectrum. It is thus capable of monitoring these important radiations. The problem is that it is still unknown which wavelengths, and which optimum combination of wave- lengths, cause tanning, or tanning with "sunburning.” It is suggested that the user regard as very tentative any prescription of ultraviolet radiation for any purpose whatever. The IQ meter is best suited to monitor ultraviolet radiation from the point of view of "the less the better:” Some exceedingly useful measures of ultraviolet-content of illumination are output by the IQ meter. One example is "microwatts of ultraviolet per visible lumen." These measures are to be used primarily by those who, for a host of reasons, wish to control, and usually minimize, the ultraviolet content of illumination.

Number-two detector is sensitive to the violet-content of illumination. Power at "violet" wavelengths is more or less midway between the often-harmful fading and decomposing wavelengths of ultra-violet radiation, and the important visible wavelengths upon which we depend to see. Violet light is very much more visible than traditional handed-down thinking acknowledges, and than the internationally adopted standards specify. The violet-content of light quite powerfully influences the perceived color of the light. Violet light is also quite powerful as a fader and decomposer of materials, although not as virulent as ultraviolet wavelengths. Materials differ greatly, of course, in their susceptibility to damage by either ultraviolet or violet light. The user of the IQ meter should, if that possibility is of interest or concern to him, carefully document the importance to his activities of both ultraviolet and violet light, separately. Alternatively, in the course of time he can relate IQ meter readings in those regions of the spectrum to his own observations on the effect of the illumination containing the violet and ultraviolet radiation.

Violet-content of light is a factor in IQ meter readings of "brightness" and "visibility," for example, as well as in human perception of "brightness" and "visibility." The footcandle-meter is blind to violet light; human vision is definitely not. It follows that, although the IQ meter reads "footcandles," because footcandles are a traditional measure of amount of light, the user of the IQ meter should immediately begin to re-educate himself, and cease to equate "footcandles" with "brightness" or "visibility." The straight- forward way to do this is simply by getting in the habit of comparing the user's own visual assessment of the illuminated scene to the readings of the IQ meter; most readings will correlate well, because they are properly visually based, but "footcandle" readings can be very misleading indeed.

Detectors 3, 5, and 7 respond to the "prime-colors" of human vision: blue-violet, green, and orange-red.

Detectors 4, 6, and 8 respond to the "anti-prime-colors," blue-green, yellow, and deep-red. The above is a carefully based and unique division of the visual spectrum. It is due to this particular segmentation of the spectrum (Figure 1) that the IQ meter yields a powerful analysis - in close accord with visual assessment - with so few detectors. Shift the responses one way or another and you lose the prime capability of the IQ meter: its ability to substitute for the human visual system.

Each of the eight photodiodes (without its associated filter) is sensitive to radiation in a broad range of wavelengths. Photodiodes vary appreciably in their spectral sensitivity, from one diode to another, making it unwise to assume that any two diodes are identical in their spectral response. Stable filters are used, a different filter with each diode, to restrict incoming light to the band desired for that diode. For example, the proper filter admits only blue-green light to photodiode number four, a different filter admits only deep-red light to photodiode number eight, etc. Yet it is also unwise to depend on absolute reproducibility of, say, blue-green- filter transmission from batch to batch of filter material, any more than to assume that the photodiode under the blue-green filter of IQ meter #9701 is identical in its response to the photodiode under the blue-green filter of IQ meter #9702. So none of these assumptions is made.

Each IQ meter, together with its assigned detector-head, is calibrated against a large number of samples of real lamplight of widely disparate characteristics. The characteristics of each lamp light sample are determined precisely by the Prime-Color, Inc. spectroradiometer. From the ensuing huge amount of data, the memory of each IQ meter is filled with its own unique set of factors. These factors tell it how it should respond, under any lamplight whatever, when any of the twenty computed characteristics of that lamplight is called for.

Since most artificial lamplight flickers twice each voltage-cycle (120 times per second), and the light output waveform of the lamp in adjacent half-cycles is often not identical in shape, the incoming lamplight must be sampled thoroughly throughout one or a small number of complete cycles, or throughout a large number of cycles. For example, the sampling should be carried out over a period of time rather exactly equal to N times 16,667 microseconds, if N is a small number. On the other hand, smoothing with RC filters, or by other means, and sampling for a second or so is acceptable. What will not suffice is to sample for very short periods of time, milli- seconds or microseconds, even if the sampling takes place at the same point of the voltage cycle.

IX. TECHNICAL DETAILS

 

Type: Light, compact, hand-held eight-eyed illumination com- puter-analyzer (illumination quality meter) with digital readout of twenty computed characteristics of lamplight, by liquid-crystal display and microprocessor; multiple uses and interfacing.

Receptors: High-sensitivity photodiodes (one GaAsP photodiode and seven silicon photodiodes) filtered to detect eight closely-specified bands of radiation covering the ultraviolet and visible regions of the spectrum, under integrating flat opal diffuser.

Spectral response: Individual photodiode-filter combinations engineered to respond to ultraviolet out to 250 nm, violet, the three prime-colors (blue-violet, green, and orange-red), and three anti- prime-colors (blue-green, yellow, and deep-red).

Calibration: Each individual IQ meter is calibrated against a number of actual lamps of widely disparate spectral power distribution, and for each of which the values of the twenty output functions are known.

Measured characteristics: (See Table below)

Minimum illumination required: Twenty milliwatts of white light per square meter, or five milliwatts per square meter of colored light.

Allowable temperature range of operation: 10 to 120 degrees Fahrenheit (-10 to 50 degrees Centigrade). Nickel-cadmium batteries should be kept above freezing. Some drift of analog values and spectral responses at temperature extremes.

Controls: Sixteen-key pad, main switch.

 

Measured Lamplight Characteristics:

 

 

*Whichever is larger.

Display: LCD type, 16-digit, 5 x 7 matrix, decimal and units.

Audible signal: When finished reading.

Operating functions:

1. Print display on printer.

2. Print all twenty computed characteristics of lamplight.

3. Enter new baud rate.

4. Enter test-end instruction.

5. Dump eight raw values to serial port.

6. Display battery volts.

7. Enter low-battery threshold.

8. Display low-battery threshold.

 

Accessories: Printer, case.

Dimensions: 305 x 260 x 95 mm (12 x 10.25 x 3.75 inches)

Weight: About four pounds (1800 grams).

(Specifications subject to change without notice.)

X. APPLICATIONS

 

A. Introduction

Illumination Quality and Quantity is best judged by consensus, using a group of expert, unbiased, experienced human observers with normal color vision. Given such a group of observers, one can expect to get a reliable evaluation of any aspect of illumination. After all, "illumination" has little meaning except as it relates to human activities, and as judged by human observers.

Is the office lit brightly enough for painstaking paperwork? Does the restaurant illumination make food look appetizing? Does the concert hall enable a patron to read his program and yet highlight the orchestra on the stage? Does the lighting in the hospital room encourage the patient by benefitting his appearance, to his own satisfaction and that of his visitors? Does the hotel lobby lighting yield attractive coloration of guests and decor? Will the draftsman in the proposed engineering complex be able to work long hours without visual fatigue?

A group of expert observers, with lots of time and patience, can answer these questions. The human visual system can make such judgements, and of course helped formulate the questions in the first place. But such an investment in man-hours is seldom possible.

Is there a substitute for the group of human observers, a sort of "secondary standard"? Partly. Consider the complete spectral power distribution of the illumination, i.e. the composition of the lamplight, in one of the situations above. A first-class spectroradiometer can determine the spectral content of the lamplight. Once that is

 

accomplished, a first-class computer can evaluate the content of the lamplight, and come up with ratings which partially answer many questions like those above.

But you now own, and can hold in your hand, an instrument which can do better even than that.

What was needed: (1) An optical device simpler than the spectro-radiometer, (2) more sophistication in relating the results to what we actually see, and (3) the same first-class computer.

The Illumination Quality meter looks at the lamplight, but sees much of what you see when you look at the illuminated scene. Among other things, the IQ meter sees brightness, footcandles, attractiveness of coloration, color, color-temperature, color-render- ing index, gamut of coloration, fading risks, color-scheme stability, visibility. It can tell you how many brightness units you are getting for every footcandle, and for every watt of light; how many visible watts fall on a square meter of work surface. It will give you an idea of how stable your color-schemes will be in a proposed illumination.

No longer is the user of commercial lamps more or less at the mercy of the lamp manufacturer. The user can now know exactly what his customers will enjoy about the lighting - and what they will find distasteful.

Among the important qualities of illumination, particularly arti- ficial illumination as supplied by commercial lamplight, are:

1.        Brightness, as perceived by the user.

2.        Coloration of the scene

a.        as judged by the trueness of the observed colors,

b.        as judged by the pleasantness and satisfaction given

  by the colors of the surroundings

                c.      as judged by how clearly colors are seen.

None of these three qualities of illumination has been adequately measured by meters available to the user. The familiar footcandle meter can differ by 100% or more from expert visual judgement of the brightness of a scene, while no good measure at all of visibility was possible. The Color-Rendering Index (CRI), which can be calculated with the use of a full-scale computer, is an index of similarity of coloration to that afforded by real phases of daylight; it fails to agree with the observer's assessment of coloration of a scene in all three respects, trueness, pleasantness, and clarity.

 

B. Some things the IQ meter can help you do:

Assess particularly pleasing and successful lighting installations, to determine which qualities are important ones in those cases.

Assess a problem installation, to determine what qualities are lacking.

Assess potential customers' present lighting, to establish a basis on which to make recommendations for improvement.