Technology Corner: Randy Hoffner
What is Color?
Television has been an integral part of our day-to-day
lives since the 1950s, and it was fervently pursued as a concept
far earlier than that. Although early television was monochrome,
the pioneers of the technology were interested in transmitting color
pictures across the airwaves from the beginning.
Humans, and many other species, have highly developed
color vision. Color vision, in addition to having functional value,
enriches our lives aesthetically. The world would be drab indeed
without it, and it is not surprising that color television transmission
was so avidly pursued after monochrome transmission became a reality.
What is color, really? Reduced to its most fundamental
properties, the color of an object or a light is dependent on the
frequencies of the light waves that reach our eyes. The first scientific
description of color was made by Isaac Newton in 1666, based on
an experiment he performed in which a small hole in the wall permitted
a beam of white sunlight to enter a dark room.
Newton passed this beam of light through a glass prism,
causing the light to spread out fan-wise into a strip he called
a spectrum, a continuum of colors from red – through orange,
yellow, green, blue, indigo – to violet.
We now know that this spreading is caused by refraction,
a property of glass or other transparent material that causes light
rays to be slowed down – and thereby, bent – as they pass
through it. The amount of slowing – and thus the amount of
bending – depends on the frequency of the ray, with the lower
frequencies being bent less than the higher frequencies.
COLORS OF THE SPECTRUM
Newton performed a second experiment to determine
whether the colors of his spectrum could be further broken down.
He passed a narrow piece of his spectrum – containing a single
color – through a small slit in a light barrier and then through
a second prism. From this, Newton determined that the colors displayed
by his spectrum could not be further reduced into other components.
The visible spectrum that Newton discovered is a part
of the whole electromagnetic spectrum that is bordered on the low-frequency
side by the infra red, which we know as radiant heat and –
on the high-frequency side – by the ultra violet, which causes
sunburn. It occupies a place in the electromagnetic spectrum above
radio frequency waves and below X-rays.
While radio-frequency energy is typically referred
to in terms of frequency, light is typically referred to in terms
of wavelength. The visible light spectrum is usually displayed as
ascending wavelengths (expressed in nanometers, or meters-9) from
left to right – conversely to the way the radio-frequency spectrum
is usually displayed, as ascending frequencies from left to right.
The visible spectrum extends from approximately 800
- 400 nanometers, or approximately 400 - 800 million megahertz –
a range of only about a single octave. This is in marked contrast
to the audible spectrum, which has a range of about ten octaves.
The visible spectrum is a continuum and its identifiable
colors are somewhat arbitrary, but the spectral colors are typically
identified as red, orange, yellow, green, blue and violet. It is
unknown just why Newton added a seventh distinct color – indigo
– between blue and violet, but it could be that he just saw
things differently from others.
COLOR PERCEPTION
There are three components to color perception: the
object being viewed, the composition of the light by which it is
being viewed, and the eye of the beholder. A complete study of color
involves not just physics, but physiology and even psychology. A
reliable estimate of the number of colors the typical human can
discern is about ten million, giving one an idea of just how sophisticated
human color vision is.
If we had to somehow represent each color discretely
– for example, with a separate phosphor on a CRT – it
is readily apparent that the gamut of colors that could be reasonably
represented on television would be seriously limited.
Fortunately, this is not necessary. It has long been
known that, with certain limitations, any color may be represented
by combining the proper proportions of just three primary colors.
The primaries that may be used are somewhat arbitrary
– but there are some conventions – with different primaries
being used depending on whether we are considering a mixture of
pigments (subtractive or reflective colorings), or a mixture of
lights (additive or transmissive colorings). The artist works with
pigments and knows the primary colors as red, yellow and blue –
from which any desired color may be mixed.
We all discovered a long time ago that when you mixed
red and yellow, you got orange; and when you mixed blue and yellow,
you got green.
The pigments used by painters are absorptive colors
– that is, when light strikes them, some colors are absorbed
and some colors are reflected – following the rules of subtraction.
Thus, when all colors are absorbed or subtracted, the pigment looks
black, and when all colors are reflected and none are absorbed,
the pigment looks white. When some colors are absorbed and others
reflected, the result depends on which colors are absorbed and which
are reflected.
When two complementary colors, or combinations that
contain all three primaries – red, yellow and blue – are
mixed in the proper proportions, the resultant mixture is gray,
with black being the completely saturated form of gray. The complement
of red is green; the complement of orange is blue; and the complement
of yellow is violet. If we analyze these pairs of complements, it
is found that all three primaries are always present. For example,
red/green = red/(yellow+blue) and yellow/violet = yellow/(red+blue).
In printing, a different set of subtractive primaries
is used: cyan, magenta and yellow. These are used because they can
create the largest color mixing space using transparent inks on
white paper. Because the ink colors are not really pure, their mixture
generally does not produce a completely saturated black. In order
to print a saturated black economically (using the least ink), a
true black ink is added – resulting in the printing primaries
called CMYK – for cyan, magenta, yellow and black (black being
called "K" because "B" indicates blue).
A CRT television picture is created by bombarding
fluorescent phosphors with a beam of electrons. In this case, the
colors generated are not reflective but transmissive – being
not reflected, but emitted – from the observed object. They
follow the rules of color addition.
The conventionally used additive primaries are red,
blue and green – and we know that these are the colors of the
phosphor dots on a television screen. The additive case is the inverse
of the subtractive case. The mixture of all three primary lights
– or the mixture of complementary colors in proper proportion
– produces gray light, while white may be considered the most
brilliant form of gray. This raises the peripheral issue of color
temperature, but that is a subject for another column.
Although within certain limits any color may be represented
on a television screen by the proper mixture of red, green and blue
lights – the transmission of full bandwidth red, green and
blue signals is impractical and in fact, impossible if full compatibility
with the monochrome system is to be maintained.
RED, GREEN AND BLUE
Fortunately, full bandwidth red, green and blue signals
are not required to adequately represent color to the human visual
system. If we capture our color pictures using a television camera
with transducers that are sensitive to the red, green and blue portions
of the visible spectrum, we may pass these RGB signals through a
matrix that produces three signals: Y, R-Y and B-Y:
Y=R+G+B
R-Y=R-(R+G+B)=G+B-Y
B-Y=B-(R+G+B)=R+G-Y
The Y signal is also called the luminance signal and
it represents the monochrome, gray-scale or brightness component
of the picture, while the R-Y and B-Y signals contain the elements
required to reconstruct the color values of hue (the spectral color)
and saturation (the intensity of the color relative to its brightness).
While we do not know exactly how the human eye works,
the most informed scientific information at present is consistent
with the eye operating on a color-difference principle similar to
the above. The eye’s receptors consist of rods, which are essentially
monochrome receptors and sensitive to low light levels, and cones,
which are color receptors and require higher light levels.
There are three types of cones: rho, gamma and beta,
which are sensitive to the red, green and blue portions of the spectrum,
respectively. The scientific evidence is consistent with information
from the cones being transmitted to the brain in a form similar
to the Y, R-Y and B-Y signals just described.
The human vision system is far more sensitive to the
monochrome or gray-scale aspects of vision than to the color aspects,
so it is possible to create a satisfying color picture by storing
and transmitting the Y or luminance portion of a picture in wide-band
form, while limiting the color information to a far lower bandwidth.
The ITU-R 601 specification as typically implemented,
for example, restricts R-Y and B-Y to about half the bandwidth of
Y (2.75 MHz versus 5.75 MHz), while in NTSC the luminance portion
of the picture has a bandwidth of 4.2 MHz and the color information
– whose axes have been rotated from R-Y and B-Y to I and Q
– have bandwidths of about 1.5 MHz and 0.5 MHz, respectively.
While NTSC (or PAL or SECAM) color cannot possess
nearly the color dynamic range of Y, R-Y and B-Y, it performs quite
acceptably in most of the living rooms of America day in and day
out.
It is more than a little amazing that the enormously
complex phenomenon of color vision may be represented by red, green
and blue dots on a television screen.
Randy Hoffner is manager of technology and strategic
planning at ABC. The views expressed in his column are his own,
and not necessarily those of ABC. Write to him c/o TV Technology.
|
