The goal of this tutorial is to clarify issues that are
relevant to a discussion about
daylight and the visual perception of it. The following topics
will be discussed:
Daylight:
What it is, and some surprising findings.
The Kruithof Curve
The Human Eye
Chromatic Adaptation
References
Daylight:
What it is, and some surprising findings
The words synonymous with daylight are "natural" or
"true light".
Daylight is defined as the combination of sunlight and skylight
The daylight condition most commonly associated with a clear blue
cloudless day exists at about 6000K. Rarely, if ever, will you
hear
someone walk out into these conditions and say, "Today is
too cool"
or "too blue". However, if you take these same lighting
conditions and
view them in an indoor setting your perception will be that the
same
light you thought was "natural"outdoors now
surprisingly appears bluish.
In this paper we will explore the physiological and environmental
reasons
for this phenomena and ways to overcome it.
Background
Our first encounter with this phenomena occurred when SoLux was
used to illuminate
Vermeer paintings at the National Gallery of Art in Washington.
An
experiment was set up so the proper amount of footcandles (20-30)
illuminated
each painting. The observers were able to adjust the lighting to
the most
preferred color temperature using SoLux at 4700K and standard
MR-16's at
3000K while maintaining a stable footcandle level. The color
temperature that
was most frequently chosen was 3500K. This was a little bit
surprising
because 3500 appears yellow-orange to the eye under normal
viewing
conditions, and the spectral power distribution supports this.
These findings could not blamed on untrained eyes because the
people at the
Vermeer test were professionals with many years experience in the
lighting of
art. (A curator from the NAG and the NGA's chief lighting
designer were
amongst the group) However, it was noted that museum light levels
are much
lower than standard light indoor light levels. Does light levels
at extreme
highs and lows and at points in between impact the eye's
perception of color?
Historical and empirical evidence suggests the answer to this
question
is yes.
The Kruithof Curve
Early work that touched upon the relationship between light
levels and color
temperature was conducted by Kruithof . He developed a chart
which
defined a region of high and low levels of illumination for a
range of color
temperatures which were considered "pleasing" to a
number of observers.
Kruithof provided the lighting designer with a breakthrough
concept that has
withstood the test of time because he used only the sun and
incandescent
sources for his study which yielded the purest possible spectral
power
distributions for his study. Based on the Kruithof Curve, SoLux
low voltage
daylight lamps should be used under the following illuminance
guidelines:
Color Temperature | Footcandles | Lux |
3500K | 18-200 | 194-2,153 |
4100K | 22-1500 | 240-16,147 |
4700K | 27-5000+ | 290-50,000+ |
5000K | 40-5000+ | 430-50,000+ |
Further refinements of the Kruithof Curve is currently being
made by
Weintraub et al. using SoLux and the lighting system used at the
National
Gallery of Art.
The Human Eye
So what causes this changing perception of color? Part of the
answer lies
with the level of lighting conditions outdoors, indoors and the
way the human
eye functions under these very different lighting conditions.
The number of footcandles measured during a typical cloudy day in
Rochester,
NY is about 3,200 footcandles with a color temperature of 6550K.
A sunny day
measures out at 13,600 footcandles and 5000K. The number of
footcandles
required for reading and displaying items in a retail store is
between 75 and
150 footcandles and the number of footcandles required in a
museum
due to conservation issues is roughly 20 footcandles.
Lighting levels produced by outdoor, indoor, and museum lighting
differ by a
factor of about ten or more in each case. In order to respond to
these
changes the iris, the entrance into the human eye, is designed to
contract
and dilate rapidly. By expanding and contracting the iris
controls the amount
of light incident upon the retina. The retina contains the rods
and cones
responsible for vision. The light incident upon the retina is
proportional
the square of the pupil diameter. If the pupil doubles in size
the amount of
light entering the eye increases by a factor of four. The iris
can expand to
8mm in dim light and contract to 2mm in bright light. This factor
of four
change in the diameter of the iris corresponds to a 16 times
change of
brightness on the retina, however the light level change from the
museum to
the sunny outdoors in Rochester is 680 times. The additional
light level
factor of 42.5 that the iris can't correct leads to a dynamic
interplay
between the two light receptors, rods and cones.
There are approximately six million cones and one hundred and
nineteen
million rods intermingled non-uniformly over the retina. The
cones which are
primarily located in the center of the retina in a region called
the Fovea
and have a responsivity as shown in figure 4 peak at 555nm (green
region).
Until recently, the cones have been primarily credited with color
vision.
The rods, whose responsivity peak at 508 nm (blue region), have
traditionally
been credited only with night vision
.
In a 1996 paper entitled, "The
Reengineering of Lighting Photometry," Dr. Sam
Berman sets forth a new theory on the workings of the human eye
where the
function of the rods and cones are not mutually exclusive as
previously
believed.
To prove his theory that rod receptors were at work all the time,
Dr. Berman
measured pupil diameters which were exposed to light sources of
equal total
output but with different spectral power distributions. The
sources which
emitted more energy closer to 508nm and away from the peak cone
sensitivity
of 550nm resulted in smaller pupil sizes proving his theory that
not only
were rod receptors at work at all times, but also that the rods
controlled
pupil size and not the cones as previously thought.
What this all means to the observer looking at a painting
outdoors, indoors
or in a museum is this: Outdoors, the light is very bright
causing the pupil
to contract, however, the change in the pupil diameter is not
great enough to
offset the large increase in illumination. With the higher light
levels the
reponsitivity of the cones is dominant yet the rods to a much
smaller degree
are still contributing to the overall response. With this large
amount of
light entering the eye, 6000K appears white.
Moving indoors, the pupil size grows allowing a larger proportion
of light to
pass. Once again, the iris is not capable of maintaining a
constant level of
illumination on the retina. Under these reduced lighting
conditions, rods
with blue sensitivity come more into play and hence the 6000K
light that
looked white outdoors now appears bluish and 4700K appears white.
By
traveling to the museum, we decrease our light level another
factor of 10
times from about 200 to 20 footcandles. The rods are utilized
even more and
4700K light which appeared white under normal indoor conditions
now appears
bluish or "cold" to the museum curator and 3500 appears
white.
It is important to point out that as the iris contracts and less
light is
allowed to enter the eye, the size of the image on the retina
does not
change. Thus for a given field of view, the same amount of rods
and cones are
always exposed, it is the amount of light which triggers a larger
visual
influence of the cones for higher illumination and rods for lower
illumination.
The interaction between the iris, the rods and the cones gives a
plausible
explanation to our observation that people will label direct
sunlight at
6000K "white" yet say that 4700K at low light levels
"looks a little blue".
Chromatic Adaptation
While the Kruithof curve characterizes a physiological condition
that
influences the perception of color, chromatic adaptation is a
psychological
condition that also plays an important role. Color adaptation is
a
rebalancing of the color response of the eye as the spectral
composition of
the scene changes. The brain is constantly working to process the
information fed to it by our eyes. Sometimes the brain needs to
"massage"
the data; without this ability, most of the light sources we work
and play
under could quite possibly make most of us ill. Take for example
fluorescent
lights; to the eye, a room illuminated by cool white fluorescent
lights
appears white, however, a photograph taken in the same room
reveals a green
glow.
Combining the effects of chromatic adaptation with large
illuminance
fluctuations creates some interesting effects. For example, at
night, the
headlights on a car appear bright and white however, during the
day, the same
headlights appear dim and yellow. The inverse of this example is
to
introduce a small beam of daylight into a room illuminated by
incandescent
light. While the room appears white, the beam of daylight appears
blue.
Unless the same beam of daylight is viewed in an environment of
daylight, it
would be hard to convince the observer that incandescent light is
yellow and
daylight is "white" and not "blue"
Returning to the original question regarding daylight-is it
"blue", or is it
"white"? Obviously, there is not a single or simple
answer. It depends on
the level of intensity of the source and the surrounding
environment. At
high levels of intensity (i.e. out doors) daylight with color
temperatures
ranging from 4500-6000k all appear white to brain. At
intermediate levels
indoors, 6000k will tend to appear blue especially when directly
compared to
incandescent at 3000k and 4100K-4700K appears white. At lower
intensities
(i.e. museums) daylight anywhere in the 4500-6000k range will
seem a little
blue and 3500K-4100K will appear white.
For designers lighting an area or an object with daylight, they
must take
into consideration many aspects including- how brightly lit the
room needs to
be, and what adaptation have the viewer's eye's undergone and of
course, what
it is they are trying to illuminate. With the advent of the SoLux
lamp at
4700K and the soon to be announced 3500K and 4100K lamps, the
lighting
designer will have the ability to create the perfect lighting
condition for
whatever the application. Kevin P. McGuire
References
S.M Berman, "The Reengineering of Lighting Photometry,"
Publications of the
Lighting Research Group, Lawrence Berkeley Laboratory,
California, 1995
Roy S. Berns and Frank Grum, "Exhibiting Artwork: Consider
the Illuminating
Source, "Color Research and Application, vol. 12, no. 2,
April 1987
Robert G. Davis and Dolores N. Ginther, "Correlated Color
Temperature,
Illuminance Level, and the Kruithof Curve," Journal of the
Illuminating
Engineering Society, Winter 1990.
IES Lighting Handbook, Lighting Handbook Eighth Edition, IESNA,
New York,
1993
Optics, Hecht & Zajac, Addison-Wesley Publishing Company
Inc., 1974