SPECIFYING LIGHT & COLOR

SEEING COLOR.

How we see color depends on the color of the light source, the color of the object and, most importantly, the combination of the eye and brain ("eye") of the viewer. The human "eye" responds to that very small portion of the electromagnetic spectrum known as the visual spectrum-the red, orange, yellow, green, blue, indigo, and violet of the rainbow. However, it does not respond uniformly, as shown in (Fig. 1). Given the same output of power at each wavelength or color, the "eye" will sense the yellow-green region as the brightest-and the red and blue region as the darkest. That is why a light source which has most of its power in the yellow-yellow/green area will have the highest visual efficiency: i.e. the highest lumens per watt.

Figure 1

However, without a reasonable proportion of red and blue in its output, a light source will not be able to render colors satisfactorily. With most sources, the better the color rendition the lower the efficiency.

How the "eye" really sees color is still only a theory. What is most difficult to understand is that the "eye's" conception of the color of an object it is seeing is a constantly changing, highly dynamic process. It depends on what colors surround the object, how long the "eye" has been exposed to the scene, what it was looking at before, what it expects to see, and perhaps even what it would like to see. Each person sees the scene differently. Added to that is the fact that about 5% of the male population and about 0.4% of the female population has a color deficiency or is "color blind" to some degree.

Trying to describe exactly what we see is, of course, extremely difficult since it depends on language to describe what is really in our mind. It is difficult to visualize the color effect in a space-and then to specify the necessary object and light source colors to make it happen.

SPECIFYING LIGHT & COLOR.

Specifying the surface colors of the visual environment-paint, wall and floor coverings, furniture, etc.-is relatively easy from material samples. Knowing which colors to specify is obviously the real task. In the same way, specifying lamp colors by the manufacturer's ordering code is straightforward. Knowing what the impact of the lamp color will be on the surface colors and on the visual environment is often considerably more difficult and creates many problems in practice.

BLACK BODY LOCUS ON C.l.E. CHROMATICITY DIAGRAM

Figure 2

CIE PASTEL TEST COLORS FOR CALCULATING CRI

Figure 3

SPECIFYING LAMP COLOR.

Lamp color can be specified by several methods, none of which is completely satisfactory, but all of which can be useful. These include x & y coordinates on an internationally agreed on chromaticity diagram (Fig. 2), chromaticity measured in kelvins (K), color rendering index (CRI) and spectral power distribution curves.

C.l.E CHROMATICITY DIAGRAM.

The C.l.E. (International Commission on Illumination) diagram is based on the idea that any color of light can be created by mixing varying proportions of hypothetical primaries of red, green and blue. This can be mathematically represented on a graph by a "triangle", in which the perimeter encompasses spectrally pure colors (seen in nature only in rainbows & prisms) ranging from red to blue. Moving toward the center "dilutes" the color until it ultimately becomes a "white". Specifying the x & y coordinates locates a color on the color triangle.

CHROMATICITY.

As a piece of metal (or a theoretical black body) is heated up, it changes color from red to yellow to white to blue-white. The color at any point can be described in terms of the absolute temperature of the metal measured in kelvins (K). This progression in color is plotted on the color diagram (Fig. 2) and is called the black body locus. The color can now be specified in either x, y coordinates or more simply, in kelvins on the black body locus.

Lamps, other than incandescent, which typically do not have a smooth, continuous spectrum can also be specified by their color temperature in kelvins represented by their x, y location. This is called correlated color temperature (CCT). Two lamps, one plotted above the black body locus and one below, could have the same CCT. However, the one above will appear green, the one below pink. Color temperature, therefore, is not really a very precise measure of color, except when specifying the chromaticity of incandescent lamps.

COLOR RENDERING.

The chromaticity of a light source defines its "whiteness", its yellowness or blueness, its warmth or coolness. It does not define how natural or unnatural colors of objects will look when lighted by the source. Two colors of lamps can have the same chromaticity, but render colors very differently. For example, "warm white" fluorescent has about the same chromaticity as high wattage incandescent, but it has far less deep red in its spectrum. Therefore, red colors will not appear as bright under WW as under incandescent.

COLOR RENDERING INDEX.

To provide guidance to colorists, a system was devised some years ago that mathematically compares how a light source shifts the location on a version of the C.l.E. color triangle of eight specified pastel colors (See Fig. 3, CIE Pastel Test Colors for Calculating CRI) as compared to a reference source of the same chromaticity.

If there is no change in appearance, the source in question is given a CRI of 100 by definition. From 2000K to 5000K, the reference source is the black body radiator and above 5000K, it is an agreed-upon form of daylight.

An incandescent lamp, virtually by definition, has a Color Rendering Index (CRI) close to 100. This does not mean that an incandescent lamp is a perfect color rendering light source. It is not. It is very weak in blue, as anybody knows who has tried to sort out navy blues, royal blues and black under low levels of incandescent lighting. On the other hand, outdoor north sky daylight at 7500K is weak in red, so it really isn't a "perfect" color rendering source either. Yet, it also has a CRI of 100 by definition.

CRI is useful in specifying color only if its limitations are understood. It was designed originally to compare continuous spectrum sources whose CRl's were above 90. Below 90 it is possible to have two sources with the same CRI, but which render color very differently. At the same time, the colors lighted by sources whose CRl's differ by 5 points or more may look the same. Colors viewed under sources with line spectra such as mercury, General Electric Multi-Vapor~ metal halide or Lucalox high pressure sodium lamps, may actually look better than their CRI would indicate. However, some exotic fluorescent lamp colors may have very high CRl's, while distorting some object color substantially.

CRl's can only be compared with sources of the same chromaticities. This is dramatically pointed out when one considers that Warm White Deluxe (WWX) fluorescent at 3000K and daylight fluorescent at 6250K have CRl's of 77 and 75 respectively. Colors under WWX obviously look significantly better than under "daylight." Yet the CRl's are the same. The CRl's of various fluorescent lamp colors are shown in (Fig. 4), plotted against chromaticity. Note that the cooler (bluer) the source, the higher its CRI.

CHROMATICITY-KELVINS

Figure 4

Why use CRI if it has so many drawbacks? It's the only internationally agreed upon color rendering system that does provide some guidance. It will be used until the scientific community can develop a better system to describe what we really see. It is an indicator of the relative color rendering ability of a light source and should only be used as such.

COLOR MATCHING.

Matching colors, whether they be textile, paint, plastic, or ink requires a different technique than choosing colors for their appearance or harmony. This is particularly true where different materials, pigments or dye lots are being matched. To avoid metamerism (colors that match under one color source but not under another), samples to be matched should be compared under two widely different sources, i.e., daylight fluorescent (lacks red) and incandescent (lacks blue). If the samples match each other under both of these sources alternately, they will likely match under any source. A simple, white painted box with two F20T12/D fluorescent lamps and two 40-watt silvered bowl lamps, separately switched, is all that is needed to solve most color matching problems (Fig. 5).

Figure 5

COLOR SELECTION.

The choice of lamp color and space color(s) is the prerogative of the lighting designer, the interior designer and the owner. The decisions should be made jointly since both affect the success of the installation.

There are a number of common sense rules of thumb that, if followed, can help make color selection easier:

All space colors (wall & floor covering, furniture, drapes, accents, etc.) should be chosen under the lamp color specified for the installation. Custom suggests that warm sources should be used at low lighting levels, cool sources at high levels. However, the choice may be influenced by space colors and by degree of luminaire brightness control.
Warm color schemes may appear overpoweringly warm if lighted with a warm source to relatively high levels-use a cooler source.
Cool color schemes may need warm sources, particularly at low lighting levels.
Well-shielded lighting systems such as those using wedge louvers, or low brightness lenses will "cool-off" a room. Solution: use a slightly warmer source such as the General Electric SP35 lamp instead of standard cool white to counteract the effect.
Where color rendition is highly critical, use high CRI continuous spectrum sources such as Deluxe Cool White (CWX), Chroma 50 (C50) or Chroma 75 (C75).
Where both color and lighting level are important, use either the SP30, SP35, or SP41 General Electric Specification Series Color lamps, with 3-peak rare earth phosphors, for good color rendering and high efficiency; or the GE SPX30, SPX35, or SPX41 lamps for very good color and high efficiency.
The Specification Series Colors tend to make spaces look more colorful because the 3peak phosphors compress all colors into the blue, green and red-orange bands. This increases the contrast between colors. Three-peak lamps do not, however, increase the contrast of black & white tasks. Claims that less light is needed for typical office and industrial tasks when they are used are scientifically unfounded.
The color of a light source does not affect productivity or the ability to do black-on-white seeing tasks. Studies at Ohio State University show that people, especially older people, perform better as lighting levels increase, but that, for example, switching from cool white fluorescent to metal halide to high pressure sodium makes little or no difference at all.

SELECTING THE BEST COLOR

To the best of our knowledge, the color of light does not affect the health and well-being of people. Most such claims are based on anecdotal observations, improperly controlled studies or on non-scientific conclusions. The light from an electric light source is no different than the light from the sun. It differs only in the relative amounts of power at each wavelength. People may be happier in a room lighted with a warm color than with a cool color, or vice versa. This is primarily a matter of personal preference, just as it is with wall, floor and furniture colors. Experience suggests that complaints about the color of a lighting system often turn out to be glare problems associated with the basic lighting layout, not color at all.

There is no "best" color lamp nor is there any such thing as "true" color. Each spectral distribution "distorts" object colors compared to another, whether it be from a natural source such as sunshine, north skylight, sunset, or electric sources such as incandescent, fluorescent, and HID. The "right" color source for a given application depends on personal preference, custom and, to a very large extent, an evaluation of the tradeoffs in efficiency, cost, and color rendition. The table below suggests some appropriate choices in order of preference.

FLUORESCENT COLORS FOR SPECIFIC APPLICATIONS:

GENERAL OFFICES	SP35, SP41, SP30, LW, CW, WW
Color-important areas	SPX35, SPX41, SPX30, SP35, SP30
Color-critical areas	C50, CWX, WWX, SPX35, SPX41, SPX30
RETAIL STORES
Food, Drug, Variety, Hardware	SPX35, SPX30, SPX41, SP35, SP30, SP41
Meat display	SPX30, SPX35, SPX41, CWX, WWX, SP30, SP35, N, plus Incandescent
Jewelry	CWX, C50, SPX41, SP41, plus Incandescent
Florist	CWX, C50, SPX41, SPX35, SP35, SP41
Shoe	SPX35, SPX30, SP35, SP41, SP30, CWX, WWX
Women's Wear	SPX30, SPX35, SPX41, SP30, SP35, WWX, CWX, SP41
Men's Wear	SPX35, SPX41, SP35, SP41, CWX
INDUSTRIAL
General	LW, SP35, SP41, SP3O, CW, WW
Printing	C50, C75, (ANSI Std. PH2.32-1972)
Textile (color checking)	C50
Paint (color checking)	C50, CWX
Meat (Inspection)	CWX, C50
PUBLIC BUILDINGS
Museums	CWX, C50, SPX35, SPX41, SPX30, WWX
Hotels/Motels	SPX30, SPX35, SPX41, SP30, SP35, SP41
Hospitals	CWX, SPX35, SPX30, SPX41, WWX, SP35, SP30, SP41
-Nurseries	C50
-Labs (color critical)	C50, CWX
-Treatment, intensive care, etc.	CWX, C50

SUGGESTED COLOR APPLICATIONS FOR HID LAMPS:

Clear Mercury	Landscape lighting, specialized floodlighting such as green copper roofs.
DX Mercury	Stores, public spaces-Multi-Vapor lamps; however, are preferred.
WDX Mercury	Stores�Multi-Vapor. Il lamps, are preferred.
MV	Stores, public spaces, industrial, gymnasiums, floodlighting signs & buildings, parking areas, sports.
MV/C	Same as MV�warmer color�diffuse coating reduces brightness.
MY/II	Stores, public spaces�warmer color. Similar in chromaticity to incandescent.
LU	Street lighting, parking areas, industrial, floodlighting, security, CCTV.

SPECTRAL POWER DISTRIBUTION CURVES.

SPD curves provide the user with a visual profile of the color characteristics of a light source. They show the radiant power emitted by the source at each wavelength or band of wavelengths over the visible region (400 to 700nm). Incandescent lamps and natural daylight produce smooth, continuous spectra. High intensity discharge (HID) lamps produce light in discrete lines or bands (used in spectral analysis to identify or fingerprint the material producing the light). Fluorescent lamps are generally a combination of a continuous or broad band spectra from their phosphor, plus the line spectra of the mercury discharge. In general, continuous spectra or very full-line spectra produce less distortion of object colors than a few discrete lines.

Basic Flourescent Colors

The following Spectrographs are enlargements from the above Chart. They appear in order, left to right, top to bottom

* Used in Mod-U-Line' Twin Tube fluorescent lamps to simulate the color of low wattage incandescent lighting.

ADAPTATION.

The first impression of the color of a room should not be taken too seriously-it will change with time. Just as a hand adapts to the temperature of hot water in which it is placed, so also will the eye adapt to color.

Sources used for general lighting will gradually appear to become "white" to the viewer, whether they be yellow/white like incandescent, or Lucalox high pressure sodium lamps-or, blue/white like daylight. Within reason, the human color vision process tends to compensate or fill in for those colors lacking in the spectrum: red in the case of daylight, blue for incandescent, etc.

The "eye's" previous state of adaptation is also a factor. A warm environment will look even warmer to the occupant if he enters it from a cold, bluish space. It will look cooler if he's been in a yellowish or pinkish one. Then the "eye" slowly adapts until the space appears to be lighted with "white" light-no matter what the "eye" was adapted to previously.

This suggests that, while side-by-side color comparisons are an excellent way to show the differences between light sources, since the eye never becomes adapted to either source but to a combination of both, a final color evaluation would be better made using a relatively large space, with only one color lamp lighted at a time. In the final analysis, the ultimate test is to live with the color or colors for an extended period of time. In that way, first reactions and adaptation effects are accounted for.

LAMP COLOR SPECIFICATIONS.

Lamp (Fluor.)	Nominal LPW*	x, y (approx.)	CRI (approx.)	CCT {K) (approx.)	Whiteness	Colors Enhanced	Colors Greyed	Notes

WW	81.3	x = .440 y = .403	52	3000	Yellowish	Orange, Yellow	Red, Blue, Green
W	80.0	x = .409 y = .394	60	3450	Pale Yellowish	Orange, Yellow	Red, Blue, Green
CW	78.8	x = .372 y = .375	62	4150	White White	Yellow, Orange, Blue	Red
LW	83.8	x = .379 y = .395	49	4200	Pale Greenish	Yellow, Blue	Red, Orange
D	65.0	x = .313 y = .337	75	6250	Bluish	Green, Blue	Red, Orange
SP30	81.3	x = .440 y = .403	70	3000	Yellowish	Red, Orange	Deep Red, Blue
SP35	82.0	x = .413 y = .393	73	3500	Pale Yellowish	Red, Orange Green	Deep Red	Rare-earth Phosphors
SP41	82.0	x = .376 y = .387	70	4100	Pale Greenish	Red, Orange	Deep Red Green, Blue	Rare-earth Phosphors
SPX27	�	x = .463 y = .415	81	2700	Warm Yellow	Red, Orange	Blue	Rare-earth Phosphors Simulates Incandescent
SPX30	81.5	x = .437 y = .402	82	3000	White (Pinkish)	Red, Orange, Yellow	Deep Red	Rare-earth Phosphors
SPX35	82.5	x = .413 y = .393	82	3500	White	Red, Orange Yellow, Green	Deep Red	Rare-earth Phosphors
SPX41	84.3	x = .376 y = .387	82	4100	White White	All	Deep Red	Rare-earth Phosphors
WWX	55.0	x = .437 y = .404	77	3025	Yellowish	Red, Orange, Yellow Green	Blue	Simulates Incandescent
CWX	56.3	x = .371 y = .368	89	4175	White (Pinkish)	All	None	Simulates outdoor daylight (cloudy day)
C50	55.3	x = .346 y = .359	90	5000	(Bluish)		None	Simulates sunlight sun-sky-clouds
C75	50.0	x = .300 y = .312	92	7500	Bluish	All	None	Simulates North skylight clear
N	52.5	x = .388 y = .366	90	3700	Pinkish	Red, Orange	Blue	Flatters complexions, meat displays seml-"cosmetic"
SGN	60.0	x = .334 y = .349	82	5200	(Bluish)	All	None	Deluxe color for plastic signs
PL	21.3	x = .320 y = .238	-2	6750	Purplish	Blue, Deep Red	Green, Yellow	Plant/Flower enhancement & growth
PL/AQ	46.6	x = .408 y = .354	90	3050	Pale Purplish	Blue, Deep Red	Green, Yellow	Plant/Flower enhancement & growth
CG	71.3	x = .307 y = .359	68	6450	Greenish White	Yellow, Green	Red, Blue
INC. (100 W)	17.5	x = .445 y = .405	99+	2900	Yellowish	Deep Red, Red Orange Yellow	Blue Green
H175 A39	45.4	x = .326 y = .390	15	5710	Blue, Green	Blue, Green	Red, Orange	Poor overall color rendering
H175 DX39	49.1	x = .388 y = .384	50	3900	Pale Purplish	Blue, Red	Green	Shifts to greenish over life
H175 WDX39	40.0	x = .413 y = .396	50	3300	Pinkish	Blue, Red	Green	Shifts to greenish over life
MVR 175	80.0	x = .384 y = .389	65	4100	White	Blue, Green, Yellow	Red	Shifts to pinkish over life
MVR 175/C	80.0	x = .398 y = .397	70	3900	White	Blue, Green, Yellow	Red	Shifts to pinkish over life
MXR 175	91.4	x = .427 y = .392	65	3100	Yellowish	Red, Orange, Yellow, Green	Red	Shifts cooler over life
LU250	110.0	x = .512 y = .420	21	2100	Yellowish	Yellow	Red, Blue
LU 250/DX	90.0	x = .505 y = .410	65	2200	Yellowish White	Red, Green, Yellow, Blue	Deep Red Deep Blue	CRI decreases slightly over life

* For Fluorescent lamps, Lumens Per Watt (LPW) applies to the standard F40 type (lamp only).