LIBRARY OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 AT URBANA-CHAMPAIGN 
 
 510.84 
 
 no. 582-587 
 
 cop. Z 
 
Digitized by the Internet Archive 
 in 2013 
 
 http://archive.org/details/penetronlandcolo583pani 
 
Ji(jt^ UIUCDCS-R-T3-583 
 JU) S~% 3 
 
 7j 
 
 /JUL^M 
 
 C00-1U69-0235 
 
 7 
 
 PENETRON LAND COLOR DISPLAY SYSTEM (PENTECOST) 
 AND SOME OBSERVATIONS CONCERNING COLOR PERCEPTION 
 
 by 
 
 G. Panigrahi 
 
 October 1973 
 
 ■ 
 
 DEPARTMENT OF COMPUTER SCIENCE 
 
 UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN 
 
 URBANA, ILLINOIS 
 
UIUCDCS-R-73-583 
 
 PENETRON LAND COLOR DISPLAY SYSTEM (PENTECOST)* 
 AND SOME OBSERVATIONS CONCERNING COLOR PERCEPTION 
 
 by 
 G. Panigrahi 
 
 October 1973 
 
 DEPARTMENT OF COMPUTER SCIENCE 
 
 UNIVERSITY OF ILLINOIS AT URBAN A- CHAMPAIGN 
 
 URBANA, ILLINOIS 6l801 
 
 Supported in part by the Atomic Energy Commission under grant 
 US AEC AT(ll-l)lU69 and submitted in partial fulfillment of the 
 requirements of the Graduate College for the degree of Doctor 
 of Philosophy in Electrical Engineering. 
 
PENETRON LAND COLOR DISPLAY SYSTEM 
 (PENTECOST) AND SOME OBSERVATIONS 
 CONCERNING COLOR PERCEPTION 
 
 Godavarish Panigrahi, Ph.D. 
 
 Department of Electrical Engineering 
 
 University of Illinois at Urbana-Champaign, 1973 
 
 This work examines the human color vision mechanism in light of 
 Dr. E. H. Land's two-color experiments on color vision. The problem of 
 color perception in machines is looked into. Some of the two-color ex- 
 periments are described. But the main emphasis has been on the building 
 of a two-color television display system based on the two-color projection 
 experiments. 
 
 The Pene tron Electronic Color System (PENTECOST) is a two-primary 
 color television system intended to examine the scope and the limitations 
 of using the Land two-primary scheme for high resolution color information 
 displays. It employs a penetration type cathode ray tube (Penetron) having 
 red and white layers of phosphors that are sequentially excited every al- 
 ternate field to display the red and green record of a scene taken syn- 
 chronously through the red-green color filter-wheel. The Penetron tube 
 as a color display tube is evaluated and the different switching and regis- 
 tration circuits are described. The camera system with the lead-oxide 
 vidicon (Plumbicon) pick-up tube, the color filter-wheel, and the associ- 
 ated synchronizing circuits are also described. The use of an electron- 
 ically controlled solid state filter like Gadolinium Molybdate instead of 
 the color wheel is considered. 
 
Ill 
 
 ACKNOWLEDGEMENTS 
 
 The author wishes to express his gratitude to Professor W. J. 
 Poppelbaum for suggesting this thesis topic and for his continued guidance 
 and support. He is also grateful to Professor W. J. Kubitz for his assis- 
 tance and encouragement during all phases of this project. 
 
 The author thanks all the members of the Hardware Systems and 
 Circuits Research Group for their friendship. He also thanks all the men 
 in the fabrication group and machine shop for the work they contributed 
 towards building of PENTECOST. Thanks are also due to Evelyn Huxhold for 
 typing the final draft of the thesis. 
 
 The author is indebted to his wife Manju for her encouragement 
 and for typing the rough drafts of the thesis many times. 
 
 This thesis is dedicated to the memory of Gautam and Bhagawan 
 who are no more . 
 
IV 
 
 TABLE OF CONTENTS 
 
 Page 
 
 1. INTRODUCTION 1 
 
 2. COLOR PERCEPTION IN HUMANS 3 
 
 2.1 Human Visual System 3 
 
 2.2 Theories of Color Vision 9 
 
 2.3 Receptors of Human Color Vision ik 
 
 2.k Representation of Color l6 
 
 2.5 Two-Color Perception 19 
 
 2.6 Retinex Theory 22 
 
 2.7 Visual Adaptation • 2k 
 
 2.8 Memory Organization and the Problem of Memory Color . . . 2k 
 
 2.9 Visual Induction 25 
 
 2.10 Brain and Visual System Organization 
 
 in Relation to Color Vision 25 
 
 2.11 A Synthetized Theory of Color Vision 28 
 
 3. COLOR PERCEPTION IN MACHINES 30 
 
 3.1 Visual Information Processing Systems 30 
 
 3.2 Color as an Attribute in Vision Systems 32 
 
 3.3 The Retinal Computer 3k 
 
 3.1+ Machine Representation of Color Information 37 
 
 3.5 Processing of Color Information . 38 
 
 3.6 Wo rld-Mo deling and Interpretation «*0 
 
 k. EXPERIMENTS IN COLOR VISION k2 
 
 k.l Land Projection Experiments k2 
 
 k .2 Sequential Projection Experiments ^3 
 
 U.3 McCulloch Effect Experiments U5 
 
 5. PENTECOST SYSTEM ORGANIZATION *+8 
 
 5.1 System Goals ^° 
 
 5.2 Organization *° 
 
 6. PENTECOST RECEIVER .51 
 
 6.1 Penetron Cathode Ray Tube 51 
 
 6.2 High Voltage Switch 5 1 * 
 
 6.3 Focus Voltage Switch 56 
 
 6.k Horizontal Deflection Control 58 
 
 6.5 Vertical Size Control 60 
 
 6.6 Video Analog Switch 62 
 
 6.7 Brightness Control and Blanking Circuits 6U 
 
 6.8 Vertical Sync. Separator and Multivibrator 66 
 
 6.9 Receiver Control Circuits 66 
 
V 
 
 Page 
 
 7. PATTERN DISPLAY SYSTEM 69 
 
 7.1 Organization of the Pattern Display System 69 
 
 7.2 Pseudo-random Binary Sequence Generator 69 
 
 7.3 Pattern Memory and Driving Logic 75 
 
 7.U Read-Write Logic 75 
 
 7.5 Pattern Timing Logic 80 
 
 7.6 D/A Convertor and Video Mixer 80 
 
 8. PENTECOST CAMERA 83 
 
 8.1 Camera Tube 83 
 
 8.2 Filter Wheel 86 
 
 8.3 Photo-detector and Amplifier 86 
 
 Q.k Camera Synchronization Circuits • 88 
 
 8.5 Color Field Indicator 88 
 
 9. ELECTRONICALLY -CONTROLLED COLOR FILTERS 93 
 
 9.1 Solid State Color Filters 93 
 
 9.2 Ferroelectric Ceramic Filter 9*+ 
 
 9.3 Theory of Ferroelectric-Ferroelastic Filter 98 
 
 9.h The Gadolinium Molybdate (MOG) Color Modulator 101 
 
 10. SYSTEM PERFORMANCE AND RESULTS 103 
 
 10.1 Pentecost Receiver 103 
 
 10.2 Pentecost Camera 108 
 
 10.3 Color Modulator Testing Ill 
 
 10. k Land Color Experiments 11*+ 
 
 11. CONCLUSION 117 
 
 APPENDICES 
 
 A. Cathode Ray Tube Selection 119 
 
 B. Camera Tube Selection 12U 
 
 C. Land Slides and Photographs 130 
 
 LIST OF REFERENCES .136 
 
 VITA lUl 
 
VI 
 
 LIST OF FIGURES 
 
 Figure Page 
 
 2.1 Information Flow in Human Brain k 
 
 2.2 Horizontal Cross-Section of the Eye 5 
 
 2.3 Plan of the Retinal Neurons 6 
 
 2.U Visual Pathways to the Lateral Geniculate Bodies 8 
 
 2.5 Co-ordinates of Color Perception.- 11 
 
 2.6 Biernson's Model ■ 12 
 
 2.7 Difference Spectra of Rod and Cone Pigments 15 
 
 2.8 Standard Chromaticity Diagram 18 
 
 2.9 Land Color Versus Long and Short Wave Sources . 20 
 
 2.10 Diagram of Components of the Visual System in Primates. . 26 
 
 3.1 Visual Information Processing System 31 
 
 3.2 Pattern Recognition System 33 
 
 3.3 Retinal Computer: Machine and Human Models 36 
 
 k.l Checkerboard Pattern kh 
 
 k.2 Sequential Projection Experiment 46 
 
 5.1 Pentecost System ^9 
 
 6.1 Pentecost Receiver 52 
 
 6.2 Screen Voltage Switch 55 
 
 6.3 Focus Voltage Switch 57 
 
 6.4 Horizontal Size Control 59 
 
 6.5 Vertical Size Control Circuit 6l 
 
 6.6 Video Analog Switch 63 
 
 6.7 Brightness Control 65 
 
 6.8 Vertical Sync. Separator 67 
 
Vll 
 
 Figure Page 
 
 6.9 Receiver Control Circuits 68 
 
 7.1 Pattern Display System 70 
 
 7.2 Shift Register with Linear Logic 72 
 
 7.3 Pseudo-random Binary Noise Generator with Programmable 
 Maximum Length and Initial Conditions 7^ 
 
 f.k Semiconductor Memory 76 
 
 7.5 Addressing Logic 77 
 
 7.6 Write Logic and Buffer 78 
 
 7-7 Read Logic 79 
 
 7.8 Pattern Timing Logic 8l 
 
 7.9 D/A Converter and Video Mixer 82 
 
 8.1 Pentecost Camera System with Filter Wheel 8U 
 
 8.2 Photo-detector Amplifier 87 
 
 8.3 Sync. Generator Board 89 
 
 Q.k Camera Synchronization Circuit 90 
 
 8.5 Color Field Indicator 91 
 
 9.1 Ferroelectric Ceramic Filter 96 
 
 9.2 Hysteresis Loop of Ferroelectric Ceramic 97 
 
 9.3 Hysteresis Loop for Gd (m 0, ) 10 ° 
 
 9.k Cross -section of MOG Color Modulator 1°2 
 
 10.1 Pentecost Receiver 10^+ 
 
 10.2 Receiver Waveforms 106 
 
 10.3 Pentecost Camera 109 
 
 10. k Drive Signals for the Color Modulators 112 
 
 10.5 Color Modulator Driving Circuits 113 
 
Vlll 
 
 Figure Page 
 
 C.l Land Black and White Transparency (Longwave Record). . . 131 
 
 C.2 Land Black and White Transparency (Shortwave Record) . . 132 
 
 C.3 Land Color Projection of Longwave and Shortwave Records. 133 
 
 C.k Land Color PENTECOST Display 13l+ 
 
 C5 Land Color PENTECOST Display (Color Reversal) 135 
 
1. INTRODUCTION 
 
 The human visual system has certain inherent peculiarities that 
 can be taken advantage of to reduce the amount of data necessary to 
 
 transmit a color picture. Dr. E. H. Land's experiments with two-primary 
 
 1-k 
 color projections demonstrate the extraordinary ability of the human 
 
 visual system to perceive object colors of all hues, even though the 
 visual system receives color information from only two primaries. These 
 experiments indicate how important a role brain mechanisms play in color 
 perception. The Circuit and Hardware Systems Research Group of the De- 
 partment of Computer Sciences at the University of Illinois has been 
 involved in uncommon methods of Information Processing. We became inter- 
 ested in Dr. Land's experiments for two reasons; one was the prospect of 
 reducing the transmission bandwith of picture signals and the other was 
 to understand the visual information processing in the human brain. 
 Dr. W. J. Poppelbaum proposed to construct a two-primary closed-circuit 
 television system called 'PENTECOST' for Penetron Electronic Color System, 
 based upon Dr. Land's experiments. It was noted at the outset that the 
 Land system reproduces color pictures of low saturation (such as would be 
 encountered in transmission of pictures of human faces, etc.) very well 
 and so could be quite suited for video transmission. An additional goal 
 of the project was to investigate some aspects of the color vision pro- 
 cessing capabilities of the human retina-cortical system, coined "Retinex' 
 by Dr. Land. Understanding the basis of the human visual system, we could 
 probably draw some analogy for the efficient organization of an intelligent 
 machine using color information for its visual perception. 
 
 Several technological advances in display devices were taken ad- 
 vantage of in building PENTECOST. One was the Penetration Cathode Ray 
 
Tube (Penetron) which became commercially available at this time. The 
 other was the prospect of having an electrically controlled color filter 
 which became possible through recent advances with ferroelectric ceramics 
 and ferroelectric-ferroelastic materials. 
 
 The organization of the thesis can be conveniently divided into 
 two parts. The first part consists of Sections 2.0 - k.O and considers 
 the problem of color vision and investigates the human visual system and 
 then speculates on congnitive machines that could have color vision. The 
 second part of the thesis, Sections 5.Q - 10.0, relates to the organization 
 and actual construction of the PENTECOST system. 
 
2. COLOR PERCEPTION IN HUMANS 
 
 2.1 The Human Visual System 
 
 The human visual system represents the utmost sophistication in 
 the organization of the nervous system. The nervous system, as such, can 
 "be divided into the peripheral system, the central nervous system (CNS), 
 and the autonomic system. The peripheral system includes the sensory in- 
 puts and the motor outputs controlling muscular activities. The CNS 
 processes all of the input data and takes action depending upon present 
 and past data. The autonomic system regulates all of .the chemical and 
 physiological processes that go on constantly to sustain life. 
 
 There are five major sensory data input channels to the CNS. The 
 data gathered by the skin receptors and that concerning the muscle and 
 joint tension and pressures are transmitted via the spinal cord. The 
 other three — visual, aural and gustatory information — are directly trans- 
 mitted to the lower portions of the brain. Figure 2.1 shows a schematic 
 representation of the human brain and the different information paths to 
 the brain. 
 
 Here, we are primarily concerned with the visual information sys- 
 tem. At the peripheral end of the visual system, we have the eye which 
 receives the light energy formed as an image of the world it looks upon. 
 Figure 2.2 shows the horizontal cross-section of the eye. Incoming 
 light passes successively through the cornea, aqueous humor, lens and 
 the vitreous humor and finally falls on the retina. The retina contains 
 the principal receptors of the light energy. A cross-section of the 
 retina, shown in Figure 2.3, explains its organization. It basically 
 contains three layers of neuron cells: the receptors consisting of rods 
 and cones, the bipolar cells, and the ganglion cells. The receptors per- 
 form photodetection. The bipolar cells interconnect the receptors and 
 
FRONiM- 
 REGION 
 
 CORTEX 
 
 OCCIPITAL 
 
 REGION 
 (BACK OF 
 
 HEAD) 
 
 INPUT f- 
 
 SOMESTHETIC £ 
 
 DATA 
 
 OUTPUT 
 J MECHANICAL 
 »\ CONTROL 
 J SIGNALS 
 
 Figure 2.1. Information Flow in Human Brain 
 
Rectus tendon 
 
 Ciliary muscle 
 
 Canal of Schlemm 
 Conjunctiva 
 
 Iris 
 Cornet 
 
 Anterior 
 chamber 
 
 Posterior 
 chamber 
 
 Limbal zone 
 
 Macula 
 
 lutea 
 
 Lamina cribrosa 
 
 Chorkxd 
 Sclera 
 
 Figure 2.2. Horizontal Cross-Section of the Eye 
 
^- y", u^ . j, \V— , / 
 
 Figure 2.3. Plan of the Retinal Neurons' 
 
the ganglion cells in a complex fashion. The ganglion cells form the 
 optic nerve which carries visual information to the occipital cortex by- 
 way of laternal geniculate bodies. 
 
 There are approximately 150 million rods and 7 million cones in 
 the retina. The rod receptor is of cylindrical shape, about 1 urn in dia- 
 meter and 50 urn long. It is divided into two sections of approximately 
 equal length, called the inner and outer segments. The outer segment of 
 a cone has the same shape as that of a rod* but the inner segment is 
 broad at the end, and contains a conical section of gradual taper from 
 the outer segment. The cones are responsible for color vision of high 
 acuity; the rods only for achromatic vision of low acuity. There is a 
 great amount of data compression in the retina as is evidenced by the 
 fact that there are only about 1 million optic nerve fibers. The foveal 
 region of the retina contains only cones and there is an optic nerve 
 channel for each cone. As we go from the foveal region to the periphery, 
 the density of the cones decreases but that of the rods increases. At 
 the peripheral portions, the output from hundreds of receptors is combined 
 in some complex manner into a single channel in the optic nerve. 
 
 The optic bundle from each eye starts from the blind spot in the 
 back of the eye and travels towards the brain. They intersect at the optic 
 chiasma. The optic bundles are split so that the optic nerve going to the 
 right hemisphere of the brain carries information from left half of each 
 retina and the one going to the left hemisphere carries information from 
 right half of each retina. 
 
 The optic nerves terminate in two-dimensional layers in the thal- 
 amus called lateral geniculate bodies. This is shown in Figure 2.U. A 
 small amount of processing is performed on the data in the thalamus and 
 
LAT. 
 GENIC ULATELS 
 
 Figure 2.k. Visual Pathways, to the Lateral Geniculate Bodies 
 
information is then transmitted to the primary visual area of the cortex 
 (also called 'area 17' or 'area striate') through bundled output axons. 
 The human train lying above the thalamus consists of so-called 
 
 'gray matter' and 'white matter'. The 'gray matter' or cortex is a sheet 
 
 2 
 of interconnected neurons about 2.5 mm thick but with an area of 2200 cm . 
 
 The 'white matter' consists of the bundles of transmission axons that 
 interconnect the thalamus and the cortex, and the various regions of the 
 cortex. 
 
 The visual information reaching the cortex is .still in conformal 
 form and takes up about 26 sq.cm. of area in the occipital part of the 
 brain. But this conformality is scrambled after the visual data are pro- 
 cessed beyond the input region of the cortex. It is experimentally proved 
 that information flows between the primary visual cortex and the secondary 
 visual cortex (area 18 which surrounds area IT). Visual perception in 
 the brain is the result of some co-relation and cross-corelation between the 
 
 Q 
 
 information in these two cortices. 
 
 2.2 Theories of Color Vision 
 
 Among the various theories of color vision the Young-Helmholtz 
 theory and the Hering theory are the most widely known. The first one 
 emphasizes the physical aspects of color and the later one is very much 
 response-oriented. In addition, a number of modifications, variations 
 and combinations of these two theories have been proposed over the years. 
 
 Young-Helmholtz Theory: Thomas Young proposed a set of three 
 
 q 
 receptors sensitive to 'red', 'green*, and 'blue'. Maxwell incorporated 
 
 these into some experiments and equations. German physicist Helmholtz 
 
 slightly modified Young's hypothesis and his contribution to a sound 
 
10 
 
 scientific explanation was so immense that the theory became known as the 
 Young- He Imholtz theory of color vision. In essence, the theory postulates 
 that there are three kinds of receptors which produce 'red', 'green', and 
 '"blue' responses and that these responses are transmitted to the brain 
 where it is processed to give color perception. 
 
 Hering Theory: German physiologist Eward Hering suggested that 
 the eye contains three chemically different substances (Empfangstof fen) 
 which absorb light and interact with a receptor mechanism (seshubstanz ) to 
 
 yield three kinds of opponent responses: white-black, red-green, and blue- 
 
 12 
 yellow. The co-ordinates of color perception according to Hering are 
 
 shown below in Figure 2.5. 
 
 Hering' s co-ordinate system has one drawback. Instead of having 
 
 zero sensation at the origin we experience gray. Wallach modified Hering 1 s 
 
 theory to include another co-ordinate of luminous-gray sensation that is 
 
 distinct from white-black sensation. The luminous gray sensation gives an 
 
 indication of the general lighting in the scene and provides sensations as 
 
 surface gloss and metallic luster. 
 
 Feedback Model of Vision: Biernson proposed a feedback model of 
 achromatic vision which is then generalized to include color vision. His 
 model is based on the Hering Theory but attempts to find a functional basis 
 for it. He developed a functional achromatic model relating to known physi- 
 ological findings on the structure of the retina and the presence of modu- 
 lation signals in the receptor output. ' This model, in the form of a 
 feedback-controlled bridge network incorporating time-average feedback, 
 spatial-average feedback and gain control action, is shown in Figure 2.6. 
 The photopigment molecules alternating between bleaching and regenerating 
 states provide the time-average feedback. The time-constant of photopigment 
 
11 
 
 BLUE 
 
 RED 
 
 BLACK 
 
 GREEN 
 
 YELLOW 
 
 Figure 2.5. Co-ordinates of Color Percepti 
 
 on 
 
12 
 
 H 
 0) 
 
 o 
 
 S 
 
 o 
 
 en 
 
 U 
 <U 
 
 •H 
 
 pq 
 
 CM 
 
 u 
 
 •H 
 P-4 
 
13 
 
 regeneration is two minutes for cone receptors and ten minutes for rod 
 
 receptors. Speaking of spatial-average feedback he says, 
 
 "The output voltage, Vo, from the receptor is coupled to the 
 conduction layer of the retina through a high resistance, which 
 feeds a spatial feedback current to the receptors. The currents 
 flowing through the conduction layer establish at each point on 
 the retina a bias voltage, Vb, which represents a weighted spa- 
 tial average of all the output voltages of all the receptors." 
 
 The variable resistance in the feedback path provides for the 
 automatic gain control function. There are two output signals from the 
 network: The white-black signal Vo, and the luminous gray signal Vo. 
 
 Biernson extends this achromatic model to provide color vision in 
 
 the following way. 
 
 "Since the outer segment of a retinal cone is only about two 
 wavelengths of visible light in diameter, it acts as a dielectric 
 waveguide, and light can propagate down it only as a summation of 
 a few particular waveguide modes. The modes cause the light energy 
 in the cone to vary with radial distance from the axis, and the 
 radial mode pattern changes in a complex manner with wavelength. 
 These patterns contain considerable spectral information and so are 
 theoretically capable of providing color discrimination. A simple 
 way in which the cone might detect its mode pattern is to employ a 
 scanning process which scans within each disk of the outer segment, 
 starting at the circumference and progressing inward to the axis in 
 the form of a contracting ring. The scan would cause the output 
 voltage from the cone to vary with time, modulating it with a wave- 
 form having the shape of the spatial mode pattern. A demodulation 
 process synchronized to the scan could derive the red-green and 
 yellow-blue chromatic signals from the modulation on the cone out- 
 put signal." 
 
 "It is assumed that the scanning processes in all the cones 
 are synchronized so that all the cone photopigment molecules in 
 the retina at a given value of outersegment radius are sampled at 
 the same instant of time. If the spatial feedback is much faster 
 than the scan, the adaptation processes would operate independently 
 at each value of radius. Hence the cones would adapt to the time- 
 average and spatial-average mode patterns. This would compensate 
 accurately for the spectral characteristics of illumination, so 
 that the mode pattern information carried as a modulation on a 
 cone output voltage would describe the reflectance spectra of the 
 object being sensed. Demodulating this waveform would provide 
 signals representing the color of the object, which are essentially 
 independent of illurainant spectrum." 
 
11+ 
 
 Biernson's electronic model gains support from the fact that 
 though three kinds of cone spectral responses have been detected it has 
 not yet been possible to detect any three kinds of cones or connections 
 between them. There is also some similarity between the scanning process 
 and the nerve impulse, except the latter have higher amplitude and propa- 
 gation rates. 
 
 2. 3 Receptors of Color Vision 
 
 Light passes through eight of the retinal layers to the receptor 
 layers where it is absorbed. The first of the light-sensitive pigments 
 to be discovered was rhodopsin. The rhodopsin molecule is made up of 
 opsin and 11-cis retinene. When light is absorbed only retinene is iso- 
 merized to produce intermediate products. Rhodopsin is the visual pigment 
 of rod receptors ' and has a difference spectrum that is similar to the 
 507 nm peak of the scotopic luminosity curve. Rods are not important for 
 color vision and occur only in the extra-foveal regions. 
 
 The first two of the pigments responsible for color vision were 
 
 20 
 isolated by Rushton. They are the green-sensitive pigment called ery- 
 
 throlable and the red-sensitive pigment called chlorolabe. The blue pigment 
 
 21 
 cyanolabe was later isolated by Wald. Erythrolabe, chlorolabe, and 
 
 cyanolabe have peaks at 55 nm, 525 nm, and U50 nm respectively. These 
 
 pigments are associated with the cones. It is likely that they have the 
 
 19 
 same retinenes but different opsins from that of the rods. The differ- 
 ence spectrum of rhodopsin as well as that of cyanolabe, chlorolabe and 
 erythrolabe are shown in Figure 2.7. 
 
 The spectral absorption characteristics overlap to a large extent 
 and it is apparent that there must be some form of encoding to extract the 
 
15 
 
 HUMAN ROD 
 
 HUMAN CONES 
 
 o 
 
 c 
 o 
 
 £ ~ 
 
 o 
 
 I 
 
 O 
 
 0) 
 
 c 
 o 
 
 a 
 
 400 
 
 400 
 
 Wavelength (nm) 
 
 500 600 
 
 B 
 
 Figure 2.7. Difference Spectra of Rod and Cone Pigments 
 
16 
 
 red, green and blue information. Some indication about the encoding and 
 processing that take place at the receptor level can be obtained from 
 Figure 2.3. The footnote in the figure explains the complex interconnec- 
 tion patterns which signify a complex data processing operation. The 
 centripetal bipolar cells are amplifier-mixers that gather signals from a 
 number of receptors and pass them on to a higher level. These are the 
 horizontal bipolars which interconnect different receptors. The centrigual 
 bipolar cells provide feedback directly from higher brain centers carried 
 through afferent fibers. 
 
 2. h Representation of Color 
 
 We will discuss in this section how color is represented in the 
 trichromatic theory of color vision. According to trichromatic theory, 
 colors are fully characterized by three parameters. Hence, it is usual 
 to represent them in three-dimensional space. In one method, where they 
 are represented by brightness, hue, and saturation, the z-axis represents 
 brightness, the angle represents hue and the radius r represents the satura- 
 tion. The notion of brightness, hue, and saturation is psychological in 
 nature. In psychophysical studies color is usually represented by red, 
 green and blue components. The psychophysical color vector Q can be repre- 
 sented in a three-dimensional space as 
 
 Q = RR + GG + BB 
 where R, G, B are the amounts of fixed primaries R, G, B that are required 
 to match color Q. R, G, B are called the tristimulus values. The chromat- 
 icity co-ordinates corresponding to these tristimulus values are defined 
 by the equations 
 
17 
 
 r = R/(R+G+B), 
 
 g = G/(R+G+B), 
 
 b = B/(R+G+B). 
 In colorimetry an ideal observer's color-matching functions are 
 defined by specifying three independent functions of wavelength. The 1931 
 CIE defined two linearly related standard observer's color-matching func- 
 tions. The RGB system employs reference monochromatic primaries of wave- 
 lengths 700, 5^6.1, and 1+35.8 nm with the white lying in the centre of 
 the diagram. One of the drawbacks of the RGB system is that one of the 
 chromaticity co-ordinates is always negative for spectral colors and 
 this complicates the calculations. The 1931 CIE adopted another trans- 
 formed trichromatic system, the XYZ system with primaries X, Y, Z that are 
 based on the primaries R. G. B. The X, Y, Z primaries are chosen to have 
 the following properties. 
 
 1. The Tri stimulus value Y is equal to the luminous flux. X and 
 Z points represent zero luminous flux. 
 
 2. The triangle XYZ is formed by the sides XY and YZ that are tan- 
 gents to the spectrum locus. 
 
 3. The white source, W, is at the centre of the diagram as in the 
 RGB system. 
 
 The XYZ triangle completely encloses the spectrum locus and the 
 purple line, and hence the chromaticity co-ordinates corresponding to the 
 tristimulus values X, Y, Z of any real color are always positive. The XYZ 
 chromaticity diagram is shown in Figure 2.8. It shows how the color solid 
 is constructed and labels names for different regions in the chromaticity 
 diagram. The chromaticity co-ordinates of the CIE standard sources A, B, C 
 are shown along with the equi-energy white. 
 
18 
 
 u 
 
 w 
 cd 
 
 •H 
 
 Q 
 -P 
 
 •H 
 U 
 •H 
 
 13 
 
 a 
 o 
 
 o 
 
 -p 
 
 CO 
 
 CO 
 CM 
 
 •H 
 
 Pn 
 
19 
 
 2.5 Two-Color Perception 
 
 Dr. Edwin H. Land's two-color experiments show some striking char- 
 
 l-k 
 
 acteristics of the color perception mechanism. Land produced color- 
 separation positives by taking two photographs, one through a filter trans- 
 mitting the longwave third of the visible spectrum (585 - TOO my with Kodak 
 Wratten No. 2k) t and the other through a filter transmitting the middle 
 third ( i+90 - 600 my with Kodak Wratten No. 58) • These two positives are 
 respectively called the long and short wave records. The long-wave record 
 is usually projected with red light and the short-wave record with incan- 
 descent lamplight. Other projecting lights have also been used. With the 
 two images properly registered on the screen, Land observed the following 
 results. 
 
 If the projection wavelengths are sufficiently different, a variety 
 of colors appear. The colors that do not appear for the various combinations 
 are shown in Figure 2.9 drawn from Land's empirical data. There is an 
 achromatic region near the diagonal and it corresponds to projection wave- 
 lengths of insufficient separation. The color reversal effect, achieved 
 when the shorter projection wavelength is used with long-wave record and 
 vice versa, is shown below the diagonal. In this instance complimentary 
 colors appear. Land also found that when red and white light were used to 
 project the long- and short-wave records, the projected picture could be 
 photographed with color film yielding a quite realistic colored picture. 
 The relative intensities of the two projecting beams was varied between 
 wide limits and did not produce any serious distrubance of the colors per- 
 ceived. The chromatic effects were more prominent with random type scenes. 
 
 In explaining his results, Land is of the view that "the color, at 
 least in images derived from primaries, depends neither on the wavelengths 
 of these primaries nor on the relative energy of these primaries at a given 
 
20 
 
 400 500 600 
 
 WAVELENGTH OF STIMULUS FOR SHORT RECORD MILLIMICRONS) 
 
 Figure 2.9. Land Color Versus Long and" Short Wave Sources 
 
21 
 
 point in the image." Instead, "color at a point in an image depends on 
 the ratio of ratios; namely, as numerator, the amount of long-wave stimulus 
 at a point as compared with the amount that might be there; and, as denomi- 
 nator, the amount of a shorter wave stimulus at that point as compared with 
 
 the amount that might be there." 
 
 22-2H 
 This view is refuted by Judd, Wool f son and several others who 
 
 are of the opinion that all of Land's results could be explained within the 
 framework of the classical color theory. Empirical formulas for the pre- 
 diction in terms of hue, lightness and saturation of the color perceived of 
 
 25 
 any object viewed in any kind of light have been developed by Judd. The 
 
 illuminant color is discounted by defining, on the Maxwell triangle, a point 
 that corresponds to the perception of gray in the object mode. This achro- 
 matic point is close to the chromaticity point of the illuminant color. By 
 
 26 
 
 Helson's principle, this point is also a function of the luminance of the 
 
 object relative to the average luminance of the scene. Judd combined these 
 two principles in the Helson-Judd formulation and successfully applied this 
 to predict colors in Land's two-primary projection. 
 
 Judd also refers to some possible effect due to simultaneous and 
 successive contrast and memory color. But he summerizes, "The hypothesis 
 that middle-wave and long-wave information provides all the information 
 necessary to determine the object-color perception is obviously doomed to 
 
 failure. Two-primary color processes must fail to yield faithful 
 
 color rendition to an extent essentially greater than the all too large 
 departures from reality afflicting current three-primary color processes." 
 
 Woolfson applied the Young-He lmboltz theory of color vision to 
 
 Land's two-primary projection and came up with similar results. Recently, 
 
 27-29 
 Pearson, et al. have applied the Helson-Judd formulation to quantita- 
 tively predict Land's results in two-primary projection and television 
 
22 
 
 situations. The success of the Helson-Judd formulation in predicting the 
 Land color points to the fact that the so-called phenomenon of color- 
 constancy is responsible to a great extent for the appearance of Land 
 color. 
 
 2.6 Retinex Theory 
 
 Land later proposed his retinex theory to explain his results. He 
 accepted the hypothesis that the retina contains receptors having peak sen- 
 sitivities in the regions of approximately 600, 550 and U70 my and that the 
 response curves for these receptors widely overlapped each other. But Land 
 proposes "that all of the receptors with maximum sensitivity to the long 
 waves in the spectrum, for example, operate as a unit to form a complete 
 record of long wavelength stimuli from objects being observed." He calls 
 this suggested retinal-cerebral system a 'retinex'. Thus, there are three 
 retinexes. The retina-cortical system draws three independent conclusions 
 in three retinexes about the lightness of the object. The result of the 
 comparison of these three conclusions gives the sense of color. Land ex- 
 plains the red-white projection experiments in terms of the retinex theory 
 in the following stanzas. 
 
 "in red and white projection, the long-wave retinex system will 
 form a lightness scale from the optical image of the black and white slide 
 projected in red light. This perceived lightness scale will be invariant 
 to changes in the brightness of the projector, to variations in the uniform- 
 ity of illumination of that projector, and to reducing the time of projection 
 to a microsecond. The other color separation slide, the one usually photo- 
 graphed through a green filter, is projected with white light. All three 
 retinexes, the long, the middle and the short-wave, Will form lightness 
 
23 
 
 scales of this optical image. We can assume that these retinex images in 
 terms of lightness will be like each other and that the objects within them 
 will have the same rank order. The lightness scale on the long-wave retinex 
 will be determined by the composite of the long and the middle wave optical 
 image. The lightness scale on the short-wave retinex is identical to the 
 one on the middle wave retinex in this experiment. The middle-wave record 
 is the one that is on both the middle and short retinexes, so that the 
 small difference in behaviour of the middle retinex and the short retinex 
 manifests itself by the appearance in binary projection of both red objects 
 on the one hand and blue or green objects on the other, depending on subtle 
 differences in the absorption curve of these latter objects. Both the short 
 and middle retinexes, because of their similarity in behaviour, contribute 
 together and without conflict to the sensations of grays, browns, oranges 
 and so on, when they are simultaneously correlated with the lightness scale 
 of the long retinexes." 
 
 There is yet no physiological basis against the existence of the 
 retinex. Land's retinex theory postulates an information processing view- 
 point presuming certain physiology for the color vision mechanism as opposed 
 
 to the usual psychological or psychophysical explanations of the color 
 
 30 31 
 perception process. More experiments have been recently reported 'to 
 
 corroborate Land's retinex theory. But we are still far from the point 
 when the visual system can be examined from an information-processing point 
 of view so as to relate its function to the structure of the retina- 
 cortical system. In the next few sections we examine some of the psycho- 
 logical effects working in favor of Land effect and look for the morpho- 
 logical basis for all these phenomena. 
 
2k 
 
 2.7 Visual Adaptation 
 
 Visual adaptation is the ability of the visual system to respond 
 over wide ranges of luminance and chrominance. This is accomplished by 
 adapting the chemical processes at the receptor level. Visual adaptation 
 also refers to 'cerebral adaptation' which automatically biases the cere- 
 bral mechanism to perceive familiar objects much the same way under dif- 
 ferent conditions of luminance and chrominance. One such example: A 
 white paper looks the same under daylight as under tungsten light though 
 the color temperature and luminance levels are widely different. This is 
 also sometimes termed color constancy. Another illustration of the influ- 
 ence of expectation on perception is the problem of memory color which is 
 discussed in the next section. 
 
 2. 8 Memory Organization and the Problem of Memory Color 
 Numerous researchers have proposed a large number of different 
 
 models of the memory in the human brain and its functional relation with 
 the perceptual processes like vision and speech. For example, auditory 
 recognition is modeled in terms of short-term memory, long-term memory 
 and different associative mechanisms. There is a labyrinth of association 
 fibers that connect various sensory receptor regions with various other 
 regions in the brain. What is perceived is the result of the sensory 
 input and the different associative activities that it excites. 
 
 What we remember about an object influences the way it appears. 
 People tend to remember the dominant attributes of familiar objects and 
 color is such an attribute. Such remembered color attributes are called 
 memory colors. Sometimes, the memory color of an object is different from 
 the actual color and the visual system would prefer to have a perception 
 of the memory color. This effect is known in color television where 
 
25 
 
 peoples' memory color of the human face is more tanned than it really is 
 and they prefer to see such a tanned color rather than the real color. 
 
 2.9 Visual Induction 
 
 'Induction' also called 'contrast' refers to the fact that the re- 
 sponse at one place influences the response at another place in the retina. 
 A white dot in a black surrounding looks whiter than it really is. The 
 problem of induction over small areas has been investigated to a great 
 extent. Here, we are more concerned with the large-area induction effects 
 which are thought to play a major role in producing Land color. It should 
 be mentioned that all of Land's experiments are large-field experiments. 
 The juxtaposition of various widely differing stimulus areas enhances the 
 chromatic and luminance difference. This involves interaction between the 
 various inducing elements in the experimental field. 
 
 2.10 Brain and Visual System Organization in Relation to Color Vision 
 
 An illuminating discussion on the neuronal organization of the 
 
 32 
 color vision mechanism is given by Ekaterina Skol'nik-Yarros. Some of 
 
 the salient points pertaining to color vision are given here. 
 
 The morphology of the color vision mechanism at the receptor level 
 has been investigated in great detail and with some success, but there are 
 extremely contradictory views on the morphological basis of color vision at 
 the cortical and sub-cortical level. A diagram of the components of the 
 visual system is given in Figure 2.10. It shows the different centripetal 
 and centrifugal interconnection paths between the cortical and the sub- 
 cortical structures. It also gives one a reasonable qualitative feel about 
 the neuronal structure of the cortical areas IT, 18, and 19. The optic 
 
26 
 
 Figure 2.10. Diagram of Components of the Visual System in Primates 
 
 32 
 
27 
 
 nerve is a projection bundle of centripetal and centrifugal fibres from the 
 cortex and the sub-cortical structures. The afferent fibres from the lateral 
 geniculate bodies give off terminal branches in Area 17, layer IV, sublayers 
 b and c. The areas surrounding Area 17 have a high concentration of pyrami- 
 dal cells that have the capability of forming thousands of connections with 
 surrounding and distant neurons. This is said to be the morphological basis 
 for association of visual images with others. 
 
 It is now generally accepted that three types of cones transmit the 
 color stimuli. How further processing takes place is a controversial ques- 
 tion. Le Gros Clark hypothesized that the six layers of the lateral genic- 
 ulate bodies link to the three cone systems. Some have ascribed color 
 vision capability to the subcortical structures and still others mainly to 
 the cortex. To ascertain the morphology responsible for color vision 
 Shkol 1 nik-Yarros compared the neuronal structure of man and primates having 
 color vision with those which do not have color vision. She noticed an 
 important cytoarchitectonic difference in the lateral geniculate body. 
 This is the presence of midget neurons which are numerous in the upper 
 layers of the lateral geniculate body in primates with well-developed 
 color vision. These midget cells resemble the small round neurons having 
 few dendrites found in large numbers in sublayer IVc of Area 17 of monkeys 
 and man. Polyak had observed in primate retina the existence of bipolar 
 cells termed 'midget cells' connecting individually with each cone. This 
 midget bipolar cell is connected with a midget ganglion cell. Polyak des- 
 cribed this cone-midget-bipolar-midget ganglion cell system as the cone 
 system. From these observations Skol' nik-Yarros concludes that, 
 
 "The proven presence of midget (in the sense of the extent 
 of their axo-dendritic connections, not of their size) cells in 
 the central part of the visual system and the fact that color 
 
28 
 
 vision may "be lost in patients with "brain lesions leads to the 
 conclusion that a special cone-midget system exists in the visual 
 analyzer of the primates. The macular part of this system is 
 most extensive, and it is evidently concerned not only with the 
 perception of color stimuli, but also with visual acuity, with 
 the fineness and precision of visual discrimination i.e., with 
 the properties characteristic of macular vision. The peripheral 
 part of this system (the rare midget cells in layers 1-2 of the 
 lateral geniculate body corresponding to the midget bipolar and 
 ganglion cells at the periphery of the retina) is much less ex- 
 tensive." 
 
 This morphological specialization of the retina-cortical system 
 
 supports Dr. Land's retinex theory which assumes independent sets of 
 
 retinexes. 
 
 2.11 A Synthetized Theory of Color Vision 
 
 In previous sections we discussed various theories of color vision. 
 The trichromatic theory emphasizes the physical aspects of color vision at 
 the receptor level. The Hering theory is based on psychological opponent 
 responses. There are numerous other variations of these theories that model 
 the color vision mechanism at the receptor level. Land's retinex theory is 
 an attempt to examine the retina-cortical system as a whole instead of 
 ignoring the cortical processes. Considerable advances in neurology and 
 cybernetics are called for before we can have a viable theory that accounts 
 for the morphological and information-processing basis of the color vision 
 
 mechanism. 
 
 33-35 
 
 The investigations of DeValois, et al. have at least shown 
 
 evidence that color vision in man is represented by two simultaneously 
 working systems, the trichromatic system and the Hering system. DeValois 
 observed in the lateral geniculate body that layers 1 and 2 do not respond 
 to color stimuli but the neurons of layers 3, h, 5, and 6 do respond dif- 
 ferentially to spectral stimuli. Single neurons in layers 3 and k gave 
 either on or off responses. For example, an on response is evoked with 
 
29 
 
 blue light and an off response with yellow light i.e., it responded to 
 complimentary colors. Neurons of layers 5 and 6 evoked both on and off 
 responses to the stimuli of the same color. But different neurons re- 
 sponded to different light stimuli. Thus layers 3 and k are a mechanism 
 working in accordance with Herings' opponent-response theory. Layers 5 
 and 6 are a mechanism working in agreement with trichromatic theory. 
 
 It could be now argued that color vision at the receptor level is 
 a three-component mechanism but that the visual information is coded into 
 different forms in subsequent processing steps prior to transmission of 
 this information to the cortex. In conclusion to this section on human 
 color vision, we again note that we are far from understanding fully the 
 intricacy of the color vision process. Any such understanding coming from 
 the works of Neurologists and Information Scientists will not only shed 
 light on brain mechanisms but also stimulate research on Artificial Intel- 
 ligence Machines that are capable of human perceptual processes. 
 
30 
 
 3. COLOR PERCEPTION IN MACHINES 
 
 The computer- controlled world of the future needs machines that 
 will carry on intellectual pursuits besides doing their unintelligent and 
 repititive chores. We will require machines that can perceive, recognize 
 and act upon their world. Vision being such an important perceptual pro- 
 cess, machines will be called upon to 'see' the world. Though these 'see- 
 ing machines' or the visual computers need not be homomorphic, we might 
 draw upon the experience of a comparative study of the human perception 
 system to understand machine perception. So an attempt is made here to 
 examine the visual systems in the light of machine vision, specifically 
 color vision. 
 
 3.1 Visual Information Processing Systems 
 
 A visual information processing system is shown in Figure 3.1. 
 It has a set of receptors that receive the visual stimuli, a processor 
 that filters out the data relevant to the system and a decision unit 
 that interpretes the current information on the basis of associated past 
 experience. 
 
 Since the early years of Artificial Intelligence, statistical pro- 
 cedures have been employed for vision systems, especially for pattern 
 recognition purposes such as optical character recognition, etc. In 
 Statistical Pattern Recognition Theory, a visual information processing 
 system is described in terms of cascaded stages of picture preprocessors, 
 classifiers and interpreters. The input stimuli is preprocessed to extract 
 the information relevant to the particular pattern recognizing activity. 
 Features are extracted from this information and the feature values are 
 
31 
 
 Output 
 
 Input 
 Stimuli 
 
 Receptors 
 
 Picture 
 Processor 
 
 Decision 
 Unit 
 
 Short term 
 Working 
 Memory 
 
 Long term 
 Memory 
 
 Figure 3.1. Visual Information Processing Syste 
 
 m 
 
32 
 
 used to classify the picture into various categories. A subsequent stage 
 interpretes the result of the classification process. Pattern classifica- 
 tion by statistical decision theory becomes unwieldly when the feature- 
 space becomes large. In a more efficient and realistic pattern recogni- 
 tion system, there should be feedback between the preprocessing and classi- 
 fication stages and the interpretation stages. Such feedback control is 
 incorporated into the human visual system where higher cortical centres 
 adjust the filtering and the processing steps, depending on the mode of 
 activity and the results of the interpretation phase. A feedback con- 
 trolled artificial pattern recognition system is shown in Figure 3.2. The 
 specific nature of feedback between the various stages is less settled at 
 this time but it should be recognized that the mathematical formalism of 
 statistical classification theory, though adequate for optical character 
 recognition, is unwise in the analysis of real scenes. In scene analysis 
 the problems of semantic and syntactic ambiguity arising from imperfect 
 edgedata, shadows, occlusion of objects, and the slow variation of tex- 
 ture, brightness complicates the problem of machine vision. In such cases 
 the interpretation is guided by a priori knowldge of the world. The 
 world-model is the visual memory that stores this knowledge of the world 
 from past learning experience and updates it as it reacts with the environ- 
 ment . 
 
 3.2 Color as an Attribute in Vision Systems 
 
 There are various attributes such as color, shape, size, and tex- 
 ture that characterize a pattern. A pattern recognition process consists 
 of searching for these attributes and then comparing them with the model 
 of the attributes that has been built by past experience. Color is a 
 
33 
 
 Input 
 Stimuli 
 
 Receptor* 
 
 Preprocets 
 or 
 
 Feature 
 Extractor 
 
 Claeeifier 
 1 
 
 Recognition 
 
 Unit 
 
 Output 
 
 Stored 
 Knowledge 
 
 Figure 3.2. Pattern Recognition System 
 
3U 
 
 very important attribute of any object. The power of any color information, 
 however crude it may be for pattern recognition purposes, is exemplified by 
 recent vision systems. ' Use of color information greatly simplifies 
 pre-processing, segmentation, and semantic procedures on the pictorial 
 data. In the past, acquisition of color information for pattern recog- 
 nition systems has not been incorporated because of the lack of proper 
 hardware and a desire to reduce the amount of data. The fact that we al- 
 most get all the necessary information from, a black and white television 
 picture does not prove the insignificance of the color information. Here 
 we rely on the highly complex human data processing system and the associ- 
 ated world-model, the internalization of the external world. 
 
 It should be emphasized at this point that color is a psycho- 
 physical attribute that has direct correspondence to the human world. One 
 could always design a visual machine that would spectrally analyze the 
 input picture in 2, 3, h or any number of dimensions. But, if there is no 
 direct transformation to the human color world it would be difficult to 
 supply any human experience so as to build the world-model. Different 
 members of the animal world have different morphology for their vision 
 systems and their perception of the world is different from ours. So 
 also, could visual computers be designed with special hardware and the 
 corresponding perceptive rules. To communicate with humans, they must 
 transpose their perception to the human perceptive system. This may not 
 be an easy task. So it is reasonable for us to design machine color per- 
 ception systems modeled after the human color perception system. 
 
 3.3 The Retinal Computer 
 
 Gregory said, "Seeing machines should have visual illusions." 
 Humans have visual illusions that are characteristic of the organization 
 
35 
 
 of the visual system. To have human-like illusions the visual computers 
 have to he functionally homomorphic. Another advantage of having func- 
 tionally homomorphic organization is the ease of communication between 
 humans and machines and convenience in relating of their experiences to 
 each other. Such a computer, called the Retinal Computer, having hardware- 
 software functional parallels with the human visual system, is shown in 
 
 Figure 3-3. The analogous functional elements of the human visual system 
 
 39 
 are also indicated in dashed lines. Corresponding to the red, green 
 
 and "blue cone system in the eye there are three television cameras equipped 
 respectively with red, green and blue filters. Binocular cameras are 
 used for depth perception. The contrast networks correspond to the inter- 
 connection network in the retina. The range finding system corresponds 
 to the lateral geniculate bodies. Some information about the visual 
 mechanism is also transmitted to the superior colliculus that determines 
 and controls the direction of gaze. The information from the lateral 
 geniculate bodies are sent to the occipital cortex Area IT and some to 
 Area 18 and 19. We do not know the exact preprocessing steps in these cor- 
 tices. But we identify these activities in visual machines as segmenta- 
 tion, region growing interpretation, etc. The world-model computer repre- 
 sents the knowledge of the world internalized in the machine. The central 
 computer co-ordiantes the activities of all the subsystems and the feed- 
 back control mechanism is included in this central computer. 
 
 Properties like chromatic adaptation and color constancy which are 
 characteristic of the human visual system could be mimicked by the visual 
 machine by proper adjusting of the camera filters. In this connection one 
 could mention some of the hardware sensing elements for the receptors. 
 Photocells and phototransistor arrays have been used in the past in addi- 
 tion to television cameras. Another possibility is the use of ferroelectric 
 
36 
 
 eye a 
 
 CONE 
 RECEPTORS 
 
 BIPOLAR 8 
 
 GANGLION 
 
 CELL, 
 
 INTERCONNECTION 
 
 NETWORK 
 
 LATERAL 
 
 GENICULATE 
 
 BODIES, 
 
 SUPERIOR 
 COLLICULUS 
 
 AREA 17 OF 
 
 OCCIPITAL 
 
 CORTEX 
 
 AREA 18 OF 
 
 OCCIPITAL 
 
 CORTEX 
 
 t_J t. 
 
 RETICULAR 
 
 CORE 
 
 COLLICULUS 
 
 CEREBELLUM 
 
 I I 
 
 RED 
 
 |f2 BINOCULAR 
 
 I 
 
 GREEN 
 
 t 
 
 t 
 
 COLOR 
 CONTRAST, 
 BRIGHTNESS 
 CONTRAST, 
 CONTRAST 
 ENHANCEMENT 
 
 LOCATOR OF 
 EDGE 8 SHADE, 
 RANGE FINDER 
 AND DETECTOR 
 OF DIRECTION 
 OF GAZE 
 
 COLOR 
 
 INFORMATION 
 
 PROCESSOR. 
 
 (SEGMENTATION, 
 
 REGION 
 
 GROWING AND 
 
 INTERPRETATION) 
 
 CONTROL OF 
 
 LENS AND 
 
 FILTER 
 
 t t t 
 
 WORLD- MODEL 
 
 COMPUTER 
 
 (SYNTACTIC 8 
 
 SEMANTIC 
 
 INFORMATION 
 
 OF THE 
 
 VISUAL 
 
 WORLD ) 
 
 CENTRAL 
 COMPUTER 
 
 Figure 3.3. Retinal Computer: Machine and Human Models 
 
37 
 
 ceramics or ferroelectric-ferroelastic elements. These elements could be 
 arranged in a grid pattern and the filtering properties of each region could 
 "be independently controlled in a feedback control environment. Such a 
 receptor scheme would he the hardware equivalent of the adaptive physio- 
 chemical elements in the retina. 
 
 3>h Machine Representation of Color Information 
 
 As described in the last section, the XYZ-system is generally used 
 for the representation and measurement of color. It has the added advan- 
 tage over the RGB system in that the tristimulus value Y carries all the 
 luminance information. The XYZ system is similar to representing color 
 in terms of intensity, hue and satruation. The difference is that the 
 former employs a rectangular co-ordinate system whereas the latter employs 
 a polar co-ordinate system. Broadcast Color Television employs a color 
 representation based on the XYZ system. 
 
 The sensor hardware extracts the separate red, green and blue 
 information from the scene. The chromatic coefficients at a point (k,l) 
 in a two-dimensional pattern can be calculated from the red, green and 
 blue components (R^, G^, B^) of light. If, T^ = R kl + G R1 + B kl 
 
 R kl °kl A v, B kl 
 
 then r kl = f" > «ki " tT • and \i = tT ' 
 
 kl kl kl 
 
 The xyz chromaticity co-ordiantes could be calculated from the rgb set of 
 
 coefficients by the following transformations: 
 
 x = 
 
 . O.U9000r + 0.31000g + 0.20000b 
 0.66697r + 1.132U0g + 1. 20063b ' 
 
 m 0.17697r + 0.8l2U0g + 0.01063b 
 y " 0.66697r + l.l32Uog + l.20063b ' 
 
 O.OOOOOr + O.OlOOOg + 0.99000b 
 0.66697r + 1.132U0g + 1.20063b 
 
38 
 
 Instead of calculating the x, y, z chromaticity co-ordinates of 
 each point it would he more realistic (from the consideration of the 
 amount of calculation involved at each point) to represent the points "by 
 r, g, h co-ordinates. All the pre-processing like edge-following, seg- 
 mentation, and the development of the color map should be carried out with 
 this internal representation. Once the color map has been developed with 
 the list structure corresponding to the boundary lines and a (r, g, b) 
 triad denoting the color of the region, we could transform the (r, g, b) 
 triad to a color name. This would be done by calculating x and y 
 chromaticity co-ordinates and then plotting it on the chromaticity diagram. 
 The color name could be identified from there. The regions would be 
 identified by a color name or by the wavelength sensation they excite 
 and the purity. Any high-level syntatic interpretation based on world- 
 modeling could use this color naming for interpretation of the visual scene, 
 
 Alternately, the chrominance information could be represented by 
 a single number by properly quantizing the area in the color triangle. For 
 effective use of the information, one would quantize such that there are 
 
 more points in the regions where our sensitivity is high such as the green- 
 
 Ul k2 
 red region in contrast to the blue region. ' 
 
 3-5 Processing of Color Information 
 
 The picture is first segmented into component subsets and then the 
 subsets are classified with the help of Statistical Classification Theory 
 and World modeling. The picture can be segmented to different regions on 
 the basis of the color at each point. Two points corresponding to color 
 co-ordinates (r, g, b) and (r\ g', b') or (x, y, z) and (x' , y' , z*) 
 would be equivalent if the distance D between the two vectors is within a 
 
39 
 
 certain limiting value. The distance D is calculated by choosing one 
 of the following criteria: 
 
 1. /(r-r*) 2 + 
 
 2. {|(r-r')|+| 
 
 3 . max { ( r-r ' 
 
 k. /(x-x') 2 + 
 
 5. {|(x-x')|+| 
 
 6. max { I (x-x' 
 
 g-g') 2 + (b-b') 2 
 g-g')|+|(b-b')|} 
 \, i(g-g')|, |(b-b')|} 
 
 ,y-y') 2 + (z-z') 2 
 ^y-y' )|+| (z-z' )|} 
 , |(y-y')|, |(z-z')|} 
 
 The metric D could be chosen to have different criteria and different 
 values for analyzing different regions. 
 
 In another method the (r, g, b) or (x, y, z) sets of co-ordinates 
 would be plotted respectively on the RGB or XYZ diagram. By joining this 
 point with the white point W and finding the intersection of this line 
 with the boundary of the chromaticity diagram, we determine the radiation 
 wavelength equivalent to the color sensation. This corresponds to hue. 
 The distance between the color point and the white point is the satura- 
 tion. The color mapping could be carried out using either the hue alone 
 or hue and saturation together. The distance metric should be nonlinearly 
 controlled to match the sensitivity in different regions of the chromati- 
 city diagram. Some perception phenomena like color constancy (as in the 
 Land effect) could be taken care of by suitably shifting the white point 
 W with the help of the Helson-Judd formulation. 
 
 The methods used for segmentation of the picture mainly fall into 
 
 two categories; microtechniques that operates on local neighborhoods and 
 
 U3 
 macrotechniques that operate globally. Contour following is one of the 
 
 microtechniques that is used to derive the cartoon of the picture after 
 
 which all the bounded regions are labelled. When a color boundary point 
 
ko 
 
 is determined, the boundary is followed through. Contour following is 
 very susceptible to noise and some smoothing is necessary prior to the 
 contour following operation. 
 
 For developing the color map one may also use a region-growing 
 macrotechnique that starts from a seed point and then propagates in all 
 directions. Such a technique is less susceptible to noise. Once the 
 scene is segmented into different color regions we can consider grouping 
 some of the regions based on contiguity, color and context. 
 
 3.6 World-Modeling and Interpretation 
 
 The color regions developed in the processing would be analyzed to 
 form color super regions. Such grouping of color regions must be guided 
 by other considerations like contiguity, texture, shape, size and the 
 semantic structure of the picture. Any such regrouping and subsequent 
 interpretation phases are supported by an internalized world-model. The 
 machine representation of the world is a difficult philosophical question 
 and once the representation question for a problem is solved the answers 
 may follow directly. In Pattern Recognition Systems, the external world 
 is internalized by storing different pictures and their cues and then 
 comparing and contrasting this with the input stimuli. In scene analysis 
 such a scheme is very inadequate. What is needed is the representation 
 of the world in terms of a set of relations. The relational structure is 
 then stored in the computer. Such picture languages, describing the visual 
 patterns, have recently been described in the literature. ' Interpre- 
 tation is a complex process of associating the generated description of 
 the input stimuli with the world-model. It is very difficult at this 
 moment to build up a world-model that would be general enough to cope with 
 
hi 
 
 all kinds of objects and scenes. Such a model would require a vast amount 
 of memory and logic and would be extremely hard to control. Visual com- 
 puters have been developed that limit their world to, say, the world of 
 
 1+6 
 three-dimensional polyhedra or landscapes. In this way, they can be 
 
 more efficient in their limited domain. This is not a limitation of 
 
 visual computers per se, but rather has direct analogy with humans who 
 
 specialize their functions for better efficiency also. 
 
\2 
 
 k. EXPERIMENTS IN COLOR VISION 
 
 k.l Land Projection Experiments 
 
 Some of the experiments performed by Dr. Land were repeated by 
 taking the picture of a scene through a red filter (Wratten No. 2h) and a 
 green filter (Wratten No. 58). The picture taken through the red filter 
 was projected on the screen with the same red filter filtering the light 
 from the projector. The picture taken through the green filter was pro- 
 jected with white light. Neutral density filters were used to control the 
 relative intensity of the light from the two projectors. The projectors 
 were kept side by side and the projected picture was registered on the 
 screen. It was difficult to register the picture over the whole field as 
 the lenses in the projectors were not properly matched and had non-uniform 
 magnification over the field. Despite this slight misregistration one 
 could observe a colored picture with blues, greens and reds in it. The 
 experiments were performed with the following different kinds of scenes. 
 
 1. Scenes of mostly red objects: Such scenes could be produced 
 quite well with two-primary red-white projections. If the 
 
 red and white fields are not properly registered, the misregis- 
 tered area of the white field looks blue-green as is expected 
 from color constancy. 
 
 2. Scenes of mostly green objects: It is difficult to produce 
 saturated green. The green obtained is mostly blue-green. 
 
 3. Scenes of mostly blue objects: It is again difficult to pro- 
 duce saturated blues. The blues are dark and sometimes little 
 blue-green. 
 
43 
 
 k. Checker board pattern: The two-primary red-white projection 
 experiments were repeated with the checkerboard pattern shown 
 in Figure 4.1. One would anticipate a pattern with a varying 
 amount of red saturation corresponding to mixture of different 
 amounts of red and white light. But, we also observed other 
 colors. 
 5. Scenes of all different colored objects: Reds, greens and 
 
 blues were observed in such projections. The blues were dark 
 blues. It was observed that the degree of the Land color ef- 
 fect was most pronounced when the scene had a juxtaposed set 
 of various differently colored objects. 
 Some of these experiments were repeated by putting a Mondrian 
 density pattern in front of the projector to spatially vary the intensity 
 of the projected light. No appreciable change was observed in the colors 
 perceived. 
 
 The projected superimposed picture of the scene with all different 
 colored objects was photographed using ordinary Kodachrome II slide film. 
 The developed slide, when projected with white light, showed the same Land 
 effect. These films were exposed only to red light but still we see greens 
 and blues upon projection. When a region of the film representing a green 
 object was examined through a microscope over a small area, it looked gray- 
 ish not greenish. This confirms that the Land effect is a large field 
 effect. 
 
 U.2 Sequential Projection Experiments 
 
 To examine the feasibility of a sequential color system based upon 
 the Land experiments a preliminary test was made on a set-up where the red 
 
kk 
 
 R 
 
 G 
 
 B 
 
 Y 
 
 B 
 
 Y 
 
 R 
 
 G 
 
 R 
 
 G 
 
 B 
 
 Y 
 
 B 
 
 Y 
 
 R 
 
 G 
 
 Figure k.l. Checkerboard Pattern 
 
h5 
 
 and green scenes were sequentially projected on the screen. As shown in 
 Figure U.2, a synchronous motor (l800 rpm, 1/100 H.P. ) was mounted on the 
 front plate of a box on which the projectors were placed. The synchronous 
 motor drove a filter wheel made of epoxy-glass, half of it painted black 
 and the other half left transparent. The filter wheel was large enough to 
 cover both the projectors. When driven by the motor it sequentially 
 allowed the projected light from the two projectors to be registered on 
 the screen. A red filter was kept in front of the projector that had the 
 longwave record slide in it. Thus, red and white fields were successively 
 projected on the screen. 
 
 The Land color effect was still present with this sequential pre- 
 sentation. As anticipated, our eye integrated the red and white fields 
 and one still observed the Land color effect. There was some flicker 
 present as the color frame frequency was only 30 frames per second cor- 
 responding to the speed of 30 revolutions per second for the filter wheel. 
 
 U.3 McCulloch Effect Experiments 
 
 hi 
 McCulloch projected a grating of vertical black stripes on an 
 
 orange ground for a few seconds alternating it with an identical grating 
 of horizontal stripes on a blue ground. He reported that an observer 
 watching this for a few minutes would see weak negative colors when watch- 
 ing black horizontal and vertical stripes. The vertical stripes had a 
 
 blue-green negative color and the horizontal stripes an orange negative 
 
 1+8 
 color. Stromeyer has recently shown a McCulloch effect analog of two- 
 color projections. He employed black-and-red striped horizontal gratings 
 and black-and-green striped vertical gratings for adaptation. A neutral 
 test matrix of alternating vertical and horizontal gratings of various 
 
1*6 
 
 PROJECTOR 1 
 
 rr 
 
 PROJECTOR 2 
 
 rc 
 
 SCREEN 
 
 FILTER WHEEL 
 
 RED FILTER 
 WRATTEN NO. 24 
 
 Figure k.2. Sequential Projection Experiment 
 
hi 
 
 lightness and contrast was viewed thereafter and colors of various hues 
 and lightnesses were observed. 
 
 We repeated the McCulloch effect experiments and employed the red 
 and green after-effect to observe the test pattern described by Stromeyer. 
 We observed different weak colors on this pattern. 
 
 However, our main interest has been the following: Having observed 
 these two experiments could we say anything about the Land effect. McCulloch 
 explained the color after-effect in terms, of 'color adaptation of orienta- 
 tion-specific edge-detectors.' 'They indicate that edge-dectector mechanisms 
 in the visual system are subject to color adaptation, responding with de- 
 creased sensitivity to these wavelengths with which they have recently been 
 most strongly stimulated. ' We also noted from the work of Festinger, et aL 
 that patterns of temporal intensity-changes with a constant background 
 could produce flicker colors. This is attributed to 'artificially creating, 
 somewhere in the visual system, a temporal pattern of neural firing, a se- 
 quence of intensity changes of the retina that produce the proper modulation 
 of firing rates.' 
 
 We conclude that some of these color adaptation phenomena are also 
 contributing somewhat to the Land effect. Assuming a temporal modulation 
 theory of color coding, some of the colors can be created by movements 
 that essentially impose a temporal pattern on the same receptors. 
 
U8 
 
 5. PENTECOST SYSTEM 
 
 5.1 System Goals 
 
 The Pene tron Electronic Color System (PENTECOST) is a two-primary 
 color television system intended to examine the scope and the limitations 
 of using the Land two-primary scheme for high resolution color information 
 displays. The goals are to determine the degree of the Land effect using 
 both static and dynamic displays. In addition, the Penetron tube as a color 
 display tube is evaluated and the use of an electronically controlled solid 
 state filter for color display systems is considered. 
 
 5.2 System Organization 
 
 The PENTECOST system is a two-color closed circuit television system 
 based on Dr. Land's two-color red and white projection experiments. The 
 selection of the cathode ray tube, a special Penetron tube, is described in 
 Appendix A. This Penetron tube has layers of red emitting and white emit- 
 ting phosphors. The red and green version of a scene are sequentially pre- 
 sented on the CRT by synchronously switching the beam voltage to alter- 
 nately excite the layers of red and white emitting phosphors. A black-and- 
 white monitor was modified to drive the Penetron CRT. The use of the Pene- 
 tron tube calls for special high voltage switching circuits and gain cor- 
 rection circuits. 
 
 The camera is a black and white vidicon camera that was modified to 
 accept a Plumbicon tube. A two-color red and green filter wheel rotates in 
 front of the camera. An electronically controlled solid state filter could 
 also be used in place of the filter-wheel. The Pentecost system with the 
 Penetron and the solid state filter is shown -in Figure 5.1. Section 6 des- 
 cribes the Pentecost Receiver system. 
 
h9 
 
 Black and White 
 
 Monitor 
 
 Gain Correction 
 
 Circuits 
 (Video, Brightness, 
 Vertical, Horizontal) 
 
 z 
 
 Color 
 
 Field 
 
 Indicator 
 
 Red Phosphor 
 
 Penetron 
 CRT 
 
 3~f~ 
 
 Focus Voltage 
 
 Switch 
 I-5KV-2 5KV 
 
 T 
 
 Black and White 
 Camera 
 
 White 
 Phosphor 
 
 High Voltage 
 
 Switch 
 IIKV-I6KV 
 
 Variable Color Solid 
 State Filter 
 
 Figure 5.1. Pentecost System 
 
50 
 
 A pattern generating system for displaying a random checker board 
 pattern was designed and built and is described in Section J. The camera 
 system is described in Section 8 and the solid state filter is evaluated 
 in Section 9« Experimental observations on the PENTECOST system are noted 
 in Section 10. 
 
51 
 
 6. PENTECOST RECEIVER 
 
 A "black-and-white video monitor (Conrac CQF 17/N, 525 lines, 60 
 fields) was modified for the Pentecost receiver system. The CQF 17 is a 
 high resolution monitor having a video frequency response of 30 MHZ +_ 1 dB. 
 A block diagram of the receiver system is shown in Figure 6.1. The modified 
 portions of the monitor are shown by dashed lines. The modifications incor- 
 porated are: 
 
 1. Penetron Cathode ray tube 
 
 2. High voltage switch 
 
 3. Focus voltage switch 
 
 k. Horizontal Deflection Control 
 
 5. Vertical size control 
 
 6. Video analog switch 
 
 7. Brightness control and blanking circuits 
 
 8. Sync, separator circuits 
 9- Receiver timing circuits 
 
 These are described in the following sections. 
 
 6.1 Penetron Cathode Ray Tube 
 
 The heart of the receiver system is the Penetron tube which has a 
 
 1+9-53 
 voltage sensitive multi-phosphor screen. The screen consists of two 
 
 different layers of colored light emitting phosphors usually separated by 
 an insulating barrier. The color of the luminance is a function of the 
 screen voltage since the low energy electrons cannot penetrate through the 
 insulating barrier, whereas the high energy electrons can excite the second 
 phosphor. The transition of color with voltage is gradual and various com- 
 binations are possible. 
 
52 
 
 r 
 
 ~i 
 
 K 
 
 O 
 
 I 
 
 a 
 
 8 
 
 X 
 
 a 
 O 
 
 I 
 0- 
 
 i 
 
 LJ 
 
 a. 
 
 1- 
 
 o 
 
 X 
 
 UJ 
 
 * 
 
 ■e 
 
 < 1- UJ 
 
 O 3 u. 
 
 or 5 o. 
 
 > <i 
 
 UJ St 
 
 L. 
 
 o 
 
 -I 
 < 
 
 £?uj 
 a: < ar 
 uj o a: 
 > o 
 
 £l 
 
 "~l 
 
 «*uj 
 
 >- d 
 or 3 
 uj a. ; 
 > 
 
 1 
 
 < u u 
 
 5 uj « 
 
 in o: 5 
 
 or u 
 
 O 
 
 <n 
 
 UJ oc — 
 o g u 
 
 >2 
 
 in 
 
 <r 
 
 -J z 
 
 a: to or 
 
 o o 
 
 x o 
 
 in 
 
 UJ 
 UJ? I 
 
 ° m y 
 < i </) 
 
 o 
 
 I- t- UJ 
 
 (-1 
 
 2 .. u 
 
 UJ 
 
 o > * 
 
 u. I </> 
 
 o 
 
 i *1 
 
 I 
 
 ~1 
 
 4-^i 
 
 _) o 
 
 UJ <-> 
 
 >8 
 
 I 
 
 < or 
 
 ■e 
 
 :or 
 
 0. 3 
 
 ?0«: 
 
 *** i "ia^^-Ii 
 
 or 
 
 UJ UJ 
 
 ? ? o 
 
 -3.g 
 
 _^JJ 
 
 a 
 
 
 
 o 
 
 
 
 t- 
 
 
 cr 
 
 < 
 
 
 o 
 
 
 
 i- 
 
 
 >- 
 CO 
 
 < 
 
 2 
 
 UJ 
 7 
 
 a 
 
 
 hi 
 
 _l 
 
 
 O 
 
 uj 
 
 
 
 u. 
 
 
 
 
 u 
 
 
 a> 
 
 
 S> 
 
 
 •H 
 
 
 <L> 
 
 
 O 
 
 
 <0 
 
 
 K 
 
 
 -p 
 
 
 w 
 
 
 o 
 
 
 o 
 
 uj 
 > 
 
 <v 
 ■p 
 
 
 
 LJ 
 
 CM 
 
 o 
 
 
 UJ 
 
 
 <r 
 
 . 
 
 i- 
 
 H 
 
 (A 
 
 • 
 
 O 
 
 vo 
 
 o 
 
 
 UJ 
 
 (- 
 z 
 
 UJ 
 
 n 
 
 
 
 •H 
 
 
 Pn 
 
 10 a 
 
 S 
 
 o-« 
 
 UJ 
 
 >- _i 
 
 (/) CL 
 
 5 
 
 z in? 
 
 o < ^ 
 M X 5 
 
 or ^ cc 
 O <-> 
 X 55 
 
 Q 
 
 ^ 
 
 or c_> 
 O c/i 
 x O 
 
 5 Z uj ? 
 
 o o uj °- 
 u- or to ri 
 
53 
 
 The characteristics of the penetration type tube are achieved by- 
 using linear and nonlinear phosphors. The linear phosphor emits light 
 proportional to the accelerating potential whereas the nonlinear phosphor 
 starts emitting rapidly after a threshold accelerating voltage. One can 
 make the tube by first depositing a conventional screen of phosphor for 
 the high voltage color. Over this a thin penetrable dielectric layer fol- 
 lowed by a layer of phosphor for the low voltage color is deposited. This 
 results in good saturation of the colors and uniformity over the whole 
 screen. Alternately, in the multi-layered or onion-skin particle method, 
 the penetron phosphor is produced separately with each particle having its 
 own dielectric barrier. This is mixed with the linear phosphor and then 
 deposited on the screen. This gives good color separation and uniformity 
 between different tubes. 
 
 The Penetron tube has some of the following limitations. It is 
 difficult to have a fully saturated high voltage color since the low vol- 
 tage color is also excited. The brightness levels also differ because of 
 the differing electron energies. The difference in the brightness levels 
 and the separation between the two colors are inter-related phenomena. 
 Another drawback of the Penetron is that accelerating voltage, focus vol- 
 tage, vertical, horizontal, brightness and video signals have to be switched 
 whenever the color has to be changed. This imposes additional circuitry. 
 The Zenith two-gun Penetron obviates the switching problem by having one 
 gun each to excite the phosphor layers. But it introduces the necessity 
 of keystone correction, raster registration circuits and yoke crosstalk 
 compensation. Current-sensitive single gun color tubes simplify the 
 switching problem but the difficulty with these tubes is their limited 
 color separation and the problem of simultaneous variation of brightness 
 with color. 
 
5^ 
 
 The Penetron tube used was a RCA developmental type (similar to the 
 C2U092) having a dual color phosphor screen, a 17" rectangular dimension, 
 70° magnetic deflection and a single high-voltage electrostatic-focus elec- 
 tron gun. This kinescope repreduces pictures in either red, at a dominant 
 wavelength of 6070° A, or in a white that is approximately equivalent to 
 CLE. illuminant C. It is supplied with a filter glass face plate (trans- 
 mission factor 66%). The luminance is 21 foot lamberts for the red display 
 at an anode voltage of 11KV and focus voltage of 1U5O-I85OV and k2 foot 
 lamberts for the white display at an anode voltage of 16KV and focus voltage 
 of 2100-2700V. 
 
 6.2 High Voltage Switch 
 
 A high voltage switch capable of switching the penetron screen 
 voltage between 11KV and l6KV during the vertical blanking period (l.U ms) 
 was built. This is shown in Figure 6.2. The color frame indicator signal 
 is applied to the input of the MOSFET differential amplifier through the 
 2N36kk transistor amplifier. The output from the differential amplifier is 
 applied to the base of the U0327 transistor that switches the collector 
 output between low and high states. When the collector output is high the 
 lower 7235 tube is off and the plate voltage is 8KV. The series zener 
 diodes drop 8KV (200V x kO) across them so that the output voltage is l6KV. 
 When the collector output of the k032'J is low, the lower 7235 tube is on 
 and the output high voltage is discharged to 11KV through this tube. When 
 the collector voltage is high, the lower 7235 is off but the upper 7235 
 tube V charges the load to 16KV. This charging tube is driven by an 
 emitter follower transistor amplifier. The supply for this transistor is 
 obtained by a voltage tripler that rectifies the 6.3V AC filament voltage 
 from the 20KV isolation filament transformer. The high voltage output is 
 attenuated by a high-voltage, high-resistance voltage divider and then 
 
55 
 
 20KV 
 ISOLATION 
 
 115V- 
 
 22a 
 
 115V 
 2A 
 
 -# — 'I o- 
 
 115V 
 
 63V 
 
 TO 
 HEATER 
 
 tt 
 
 1N4004 
 
 500ft 
 vi/v — 
 
 ip •> 
 
 + 100 M 
 20V 
 (3) 
 
 '200K 
 
 IN 
 4148 
 
 1 
 ■-KH- 
 
 40327 
 
 u 
 
 IN 
 4759 
 
 18KV 
 
 * 
 
 100K 
 
 U\ 
 
 V. 7235 
 
 11/16.5KV 
 
 200 Mft 
 36 Mft 
 
 +75K 
 
 * 
 
 910K 
 
 -A. v^- 
 
 1N3070 
 
 « ♦ * 
 
 1N4148A :;2.2K 2.2K:; 
 
 10 
 
 •22K 
 
 I I 
 
 8KV , | 
 
 ZENER i i 
 
 STRING 
 
 (40) 
 1N9926 
 1 I 200V 
 
 1 » 7 
 
 Vj 7233 
 
 -» 2 
 
 10K IK 
 
 Figure 6.2. Screen Voltage Switch 
 
56 
 
 applied to the negative input of the differential amplifier. The MOSFET 
 differential amplifier was built using matched transistors and resistors. 
 2N36U2 transistor is a constant current source for the differential ampli- 
 fier. The output level of the differential amplifier is controlled by- 
 varying the 10K potentiometers. 
 
 Originally, the high-voltage output from the monitor was used as 
 the constant high-voltage supply for the screen voltage switch. It was 
 contemplated to employ a set of width control coils in parallel and series 
 that could be switched in and out of the circuit such that the yoke cur- 
 rent could be switched but the load to the flyback would remain constant. 
 That would keep the high voltage constant. Subsequently, this approach 
 was discarded because of its complexity in favor of an independent high 
 voltage supply. Thus, a separate high voltage power supply (10KV-20KV) 
 was obtained from CPS Inc. This allows the horizontal size to be indepen- 
 dently controlled without interacting with the monitor high voltage output. 
 
 6. 3 Focus Voltage Switch 
 
 With the change in the beam voltage from 11KV to l6KV, the electron 
 beam would be defocussed unless the focus voltage is also switched along 
 with the beam voltage. The focus voltage required is directly proportional 
 to the beam voltage. In the present case the focus voltage switch was 
 designed to switch between about l600V to 2U00V. The focus voltage switch 
 (Figure 6.3) is similar to the screen voltage switch but is less sophisti- 
 cated. The MOSFET differential amplifier output drives the base of U032T 
 transistor which in turn controls the plate voltage output of 7235 tube. 
 The 10K potentiometers control the high and low voltage levels. 
 
57 
 
 3KV 
 
 FOCUS 
 OUT 
 
 ■wv vw- 
 
 10 K IK 
 
 Figure 6.3. Focus Voltage Switch 
 
58 
 
 6.k Horizontal Deflection Control 
 
 1/2 
 The deflection sensitivity is inversely proportional to V for 
 
 magnetic deflection where V is the beam voltage. To take care of this 
 change in the sensitivity with the change of the beam voltage, the horizon- 
 tal deflecting signal must be controlled to maintain the same raster size 
 over the two fields. This is achieved by a switching attenuation circuit 
 that controls the amplitude of the horizontal signal input to the horizontal 
 output amplifier. This horizontal size control circuit is shown in Figure 
 6.k. The circuit switches the amplitude of the large horizontal signal to 
 either of two levels. These levels can be adjusted by R . The 2N5^l6 
 transistor is connected as an emitter follower and isolates the analog 
 input. The opto-coupled-isolators (OCl) are connected as multiplexed 
 analog gates and by turning on either of the two opto-isolator branches, 
 the output point is connected to the corresponding amplitude level. Out- 
 puts from the TTL circuits control the switching and they are completely 
 isolated from the large horizontal signal. 
 
 The opto-isolators Monsanto MCT-2, have typical ratings of 
 
 BV CEO = 65V ' BV CBO = l65V and BV ECO = ll|V * When 0CI_1 and 0CI - 2 are on 
 and OCI-3 is off there is a reverse voltage across OCI-3 which should not 
 
 exceed the BV^-,^ rating. To take care of the smaller BV.,_,_,_ rating, a 
 ECO ECO 
 
 1N3070 diode is connected in series with OCI-3 so that most of the reverse 
 
 drop is across the diode. When OCI-3 is on, OCI-1 and OCI-2 are off and 
 
 the voltage across either OCI-1 or OCI-2 should not exceed the BV rating. 
 
 iIjCU 
 
 Two opto-isolators were put in series to increase the overall voltage rating. 
 The offset voltage is equal to V-^C saturated) and is only 0.1V. The off 
 resistance is hundreds of Mego-ohms corresponding to a leakage current, 
 I , of a few nanoamps. 
 
 ! 
 
59 
 
 1N3070 
 
 Figure 6.1*. Horizontal Size Control 
 
60 
 
 The speed at which the output level can be changed is dependent 
 
 on the photo-transistor time constants (t = 5 US, t __ = 25 ys in the 
 
 on oir 
 
 saturated mode) and the output resistor R p . The time in which the output 
 changes from the high amplitude level to the low level is slower compared 
 to the other transistion and corresponds to the time that OCT-1 and OCI-2 
 take to discharge through R . This time is about U00 ys and is quite 
 adequate for our applications. 
 
 It should be noted that an alternative to the use of the opto- 
 isolators is to use bipolar analog switches which is difficult for such 
 large signals. Similarly, MOSFET and JFET analog switches cannot handle 
 such large amplitude signals. Photo-coupled resistors typically have 
 response times of few milliseconds and cannot meet the speed requirements. 
 
 6. 5 Vertical Size Control 
 
 To compensate for the change in the vertical deflection sensitivity, 
 the vertical deflecting signal must be controlled to maintain the raster 
 size constant over the two alternating fields. This is achieved by varying 
 the input signal to the vertical output amplifier with the help of the 
 switching attenuation circuit shown in Figure 6.5- It employs two diamond 
 type analog gates to switch the amplitude of the output vertical signal 
 between two amplitude levels. 
 
 The direct coupled emitter followers isolate the previous vertical 
 tube circuits and bring down the output impedance of the analog source. 
 This enables the analog source to drive the high resistance load through 
 the gate and minimizes the error introduced by the analog switch. Four 
 matched 1N3070 silicon diodes form the diode bridge. Two transistor con- 
 stant current sources T and T in their common base connection supply the 
 current for the bridge. The resistor network biases the zener diode so 
 
6l 
 
 _ a 
 o E 
 '1 < 
 
 -P 
 
 •H 
 
 O 
 
 U 
 
 •H 
 O 
 
 o 
 
 U 
 
 -p 
 
 c 
 o 
 
 a> 
 
 CO 
 
 > 
 
 a; 
 
 •H 
 
 En 
 
62 
 
 that transistor T can be conveniently controlled by the gate voltage. The 
 220K resistors are used to back-bias the diode bridge when T and T are 
 off. The 10K potentiometers are used to match the current sources. This 
 analog gate can handle analog signals of magnitude +50V, is fast and has 
 negligible offset voltage. 
 
 6.6 Video Analog Switch 
 
 To compensate for the change in the electron beam voltage, the video 
 signal also must be switched. The video analog switch circuit is shown in 
 Figure 6.6. A voltage follower using a fast operational amplifier (yA715C) 
 isolates the previous video stage and has low output impedance to drive 
 the analog gates. The analog gate employs a 2N5638 n-channel junction FET 
 and a bipolar circuit using the diode drive technique. The analog gates 
 have little offset voltage and the driving source uses a non-saturated 
 transistor so that this switch is capable of being operated at very high 
 speed. It can handle analog voltages of about +10V which is quite adequate 
 for this application. When the color signal is high, the 2H36kh is off 
 and the collector voltage is -15V. The INU1U8 diode is forward biased 
 and the JFET gate is kept reverse biased for negative analog signals of 
 amplitude corresponding to -15V minus the pinch-off voltage. The gate 
 leakage current flows through the driving diode and clamps the gate to -15V. 
 The capacitor is put across the diode to speed up the turn on time so that 
 the gate -to- source capacitance C is discharged quickly. 
 
 When the color signal is low, the 2N36UU is turned on and the vol- 
 tage at the collector is +15V. The gate junction is forward biased with 
 only the leakage current of the reverse-biased lNUllj-8 diode flowing through 
 it. The gate to source voltage is zero. It is noted that the source to 
 
63 
 
 o 
 
 -p 
 
 •H 
 > 
 
 CO 
 
 bO 
 o 
 H 
 
 o 
 
 > 
 
 VD 
 
 VD 
 
 ft, 
 
6k 
 
 drain resistance r^_, is minimum for V = and that a constant impedance 
 
 Do (ib 
 
 is presented to the analog source irrespective of the amplitude of the 
 analog signal. 
 
 The two JFET analog gates alternately connect the output point to 
 the corresponding video signal level. This is then fed to the succeeding 
 video amplifier stages in the Conrac monitor. 
 
 6.7 Brightness Control and Blanking Circuits 
 
 The red phosphor emits 21 foot-lamberts of light whereas the white 
 phosphor emits k2 foot lamberts. To compensate for this mismatch in the 
 luminance level and to have control over the relative intensity of the two 
 fields, the brightness level is changed every field. This is achieved by 
 the brightness control and blanking circuits shown in Figure 6.7. The 
 opto-isolators controlled by the color frame indicator signals generate 
 sixty-cycle square wave voltage whose lower and upper levels can be ad- 
 justed with potentiometers R and B. . The use of the opto-isolators 
 affords an easy way of controlling the voltage levels from to 75 volts 
 and they are fast enough for the present application. The sixty-cycle 
 square wave voltage is applied through an emitter follower to the collector 
 load resistor of the final stage transistor, T . The horizontal and ver- 
 tical blanking signal derived from the receiver circuits is level shifted 
 through the zener diodes and drives the base of transistor T . When 
 driven by the blanking signal the collector of T is clamped to -75V and 
 at all other times T is off and the collector voltage is the voltage se- 
 lected by the opto-isolator circuit. The collector output is now capaci- 
 tively added to the brightness dc level selected by the brightness control 
 potentiometer in the Conrac monitor circuit. This brightness dc level can 
 
65 
 
 +75 
 
 35K><- 
 
 35K>«- 
 
 MCT-2 
 
 l 1 
 
 \J 
 
 10 
 Fl> 
 
 -w- 
 
 I I 
 
 11 
 
 Fl> 
 
 100£2 
 
 MCT-2 
 
 -W- 
 
 i i 
 
 MCT-2 
 i 1 
 
 \jT 
 
 IN 3070 
 — « 
 
 L I 
 
 ->+5 
 
 1N3070 
 — W 
 
 J 
 
 <» ♦■ 
 
 100K 
 
 +75 
 
 40327 
 Tl 
 
 10/i 
 
 i" 
 
 -75V 
 
 H*C 3 
 
 SYNC > " 
 
 IN 
 5267B -j> 
 
 IN 
 
 50 K 
 
 < 
 
 2.0 M , 200V 
 
 40327 
 T2 
 
 jVe * 1N3070O 
 
 .-^w BRIGHT . 
 
 a IN >" 
 
 5.6K 
 
 1M& 
 ■AAAr- 
 
 ■AAAr 
 8K 
 
 150K 
 
 25K> 
 
 51V 
 
 BLANK HI 
 
 +75 
 
 IN 2 
 
 BRIGHT 
 OUT 
 
 Figure 6.7. Brightness Control 
 
66 
 
 be varied from OV to 120V. Thus we have the freedom to control the 
 brightness in many ways. A SPDT switch is provided which capacitively 
 couples either the brightness and blanking signal generated in the pre- 
 sent circuit or the high blanking signal obtained from the Conrac monitor, 
 as desired. 
 
 6.8 Vertical Sync Separator and Multivibrator 
 
 The vertical sync separator and multivibrator circuit was built to 
 separate the vertical sync from the video signal and then derive synchro- 
 nous field signals. The circuit is shown in Figure 6.8. T and T ampli- 
 fy the composite video to drive the sync clipper T . T and T are biased 
 such that they almost eliminate the video. The clipped sync is applied to 
 phase splitter T, . The emitter output is sent through a vertical integra- 
 tor and then to an emitter follower that drives the multivibrator. 
 
 6.9 Receiver Control Circuits 
 
 All the receiver switching circuits (Figure 6.9) are controlled 
 by the color frame indicator signal. The field indicator signal from the 
 Pentecost Camera system is brought by cables and drives the inverter gates 
 and the monostable multivibrators through the 7^37 gates. The 7*+06 open- 
 collector drivers produce 15 volt signal swings whereas the 'jkok drivers 
 produce TTL swings of 5 volts. The 7^123 monostables produce short field 
 indicator signals whose width can be varied. 
 
67 
 
 o 
 3 
 
 ft 
 
 <L> 
 
 03 
 
 CO 
 
 o 
 
 •H 
 -P 
 
 > 
 
 CO 
 
 •H 
 
11 
 
 7437 
 
 10 " 
 
 7406 
 
 — iH^o^4 — V\A/ >+15 
 
 .^^iT^i— 
 
 5.1K 
 ■AM >+15 
 
 <►— <► 
 
 
 5.1K 
 -AAA/ > +15 
 
 5.1K 
 AAA/ > +15 
 
 5.1K 
 AAA/ > +15 
 
 
 5.1K 
 AAA > +15 
 
 MX 
 
 H> 
 
 47/. 
 
 7404 
 
 tr 
 
 10K>«J 
 
 (°>M>^D[ 
 
 © j m>-[3 
 
 rrn 
 
 10K>«-I ;± ¥■ 
 
 I 
 
 68 
 
 -> FIH 
 
 -> FIH 
 
 -> FIH 
 
 8 
 
 -» FIH 
 
 ■> FIH 
 
 10 
 -> FIH 
 
 -> Fl 
 14 
 
 I 5 _ 
 
 -> Fl 
 
 16 _ 
 -> Fl 
 
 17 
 -> Fl 
 
 18 
 
 — > Fl 
 
 19 
 -> Fl 
 
 V.1N660 
 
 16 6 7 
 5 
 
 74123 
 
 12 
 
 20 
 — > FISP 
 
 13 
 
 -> FISP 
 
 16 14 15 
 
 13 
 74123 
 
 12 _ 
 — > FISP 
 
 21 
 
 -> FISP 
 
 Figure 6.9. Receiver Control Circuits 
 
69 
 
 7. PATTERN DISPLAY SYSTEM 
 
 The color perceived in a two-color television system is influenced 
 "by factors like the average illumination of the scene and the relative lumi- 
 nance of the projected fields. The other most important factor that in- 
 fluences the color perception is the stochastic structure of the image. It 
 has been qualitatively observed that very many colors are observed when 
 the picture has a jumbled set of different colored objects. Land also 
 observed that the color perceived in an image is not changed when a Mon- 
 drian (random) pattern is placed in front of the projector. To test these 
 observations, we designed a random pattern display system that produces a 
 random l6 x l6 checkerboard pattern that can be displayed on the screen 
 either by itself or superimposed with the image. 
 
 7-1 Organization of the Pattern Display System 
 
 The pattern display system is shown in Figure 7-1. The white field 
 memory and the red field memory store the checkerboard patterns (l6 x l6 
 raster with h bits of gray levels) for the respective fields. The pseudo- 
 random binary sequence generator produces the random sequences which are 
 written into the memory units. The checkerboard pattern is read syn- 
 chronously in the standard television format, converted to an analog signal 
 by the digital-to-analog converter and then mixed with the video if intend- 
 ed. It is then displayed on the monitor. The different pattern display 
 circuits are described in the following sections. 
 
 7 «2 Pseudo-random Binary Sequence Generator 
 
 The pseudo-random binary sequence generator produces a cyclic 
 sequence of binary bits. The usual way of producing such a sequence is to 
 
70 
 
 H> 
 
 I 
 
 RANDOM 
 
 PATTERN 
 
 GENERATOR 
 
 WRITE 
 LOGIC 
 
 READ 
 LOGIC 
 
 I 
 
 H 
 V 
 
 MEMORY 
 DRIVER 
 
 a 
 
 R/W GATES 
 
 READ 
 LOGIC 
 
 I 
 
 H 
 V 
 
 MEMORY 
 DRIVER 
 
 a 
 
 R/W GATES 
 
 WHITE 
 
 FIELD 
 
 MEMORY 
 
 ( 16 X 16 X 4 ) 
 
 RED 
 
 FIELD 
 
 MEMORY 
 
 (16X16 X 4) 
 
 D/A 
 
 VIDEO 
 MIXER 
 
 Figure 7.1. Pattern Display System 
 
71 
 
 use shift register chains with linear (modulo-2) feedback logic. For an 
 n-stage shift register, the feedback logic calculates a new term for the 
 last shift register based on the previous n terms. The n-stage shift 
 register shown in Figure 7-2, has 2 possible states. But if it is driven 
 to an all-zero state it remains in that state and hence this state must 
 be excluded. So a linear n-stage shift register can have a maximum length 
 sequence of 2 - 1. The necessary and sufficient condition that the 
 linear n-stage shift register produces the maximum length sequence is that 
 the feedback polynomial be irreducible and primitive. ' These irreducible 
 
 polynomials corresponding to the maximum length sequence have been discussed 
 
 57-59 
 
 in the literature. Table 1 lists the linear feedback logic required 
 
 for the maximum length sequences, from n = 1 to n = l6. 
 
 Using the feedback logic shown in the table, we have built a maxi- 
 mum length pseudo-random generator that is shown in Figure 1.3- The 
 pseudo-random generator is constructed with shift registers consisting of 
 SN7^T^ D-type positive edge-triggered flipflops and linear (modulo -2) 
 feedback logic consisting of SN7^86 Exclusive-OR gates. It has provision 
 for programming n at 2, U, 6, 8, 10, 12, ik or l6 by the help of the 3 
 line-to-8 line decoder. Digital gates were inserted at the feedback points 
 to control the data flow. When the decoder output is low, data from the 
 feedback logic is gated through, but the data from the previous registers 
 is blocked and vice versa. The AMP dual-in-line switches provide an easy 
 way of programming the initial conditions. The switches for the F-Fs 
 which must be preset to 1 are closed allowing them to be set to one 
 initially. Programming the length of the maximum sequence allows us to 
 control the degree of randomness or the auto-correlation of the pattern. 
 
72 
 
 x n = c nn x nn + ^n-2 x n-2 + " +c i x i + Co x o 
 Cj =Oor I 
 
 Figure 7.2. Shift Register with Linear Logic 
 
73 
 
 n 
 
 LOGIC (period 2 n -l) 
 
 n 
 
 LOGIC (period 2 n -l) 
 
 1 
 
 X - X n 
 
 n n-1 
 
 9 
 
 X 
 n 
 
 = X ,_©X _ 
 n-5 ^^ n-9 
 
 2 
 
 X = X . ©X 
 n n-1 ^-^ n-2 
 
 10 
 
 X 
 n 
 
 = X v ©x _ n 
 
 n-7 n-10 • 
 
 3 
 
 X = X _©X , 
 n n-2 v -' n-3 
 
 11 
 
 X 
 
 n 
 
 = X n ©X nn 
 n-9 w n-11 
 
 k 
 
 X = X _©X ,, 
 n n-3 n-4 
 
 12 
 
 X 
 
 n 
 
 = X _©x in ©x ,.©x 
 
 n-2 ^^ n-10 —' n-11^ n-12 
 
 5 
 
 X = X | ©X _ 
 n n-U w n-5 
 
 13 
 
 X 
 
 n 
 
 = X . ©X .. ©X no ©X ._ 
 n-1 ^ n-11 ^^ n-12 w n-13 
 
 6 
 
 X = X c ©x , 
 n n-5 ^-^ n-D 
 
 Ik 
 
 X 
 n 
 
 = X ©x no ©X no ©x . 
 n-2 w n-12 w n-13 w n-1 4 
 
 7 
 
 X = X .©X _ 
 n n-6 ^-^ n-7 
 
 15 
 
 X 
 n 
 
 = X ., ©X nc 
 
 n-lU w n-15 
 
 8 
 
 © X „-8 
 
 16 
 
 X 
 n 
 
 = X , n ©x ,,©x n)l ©x _, 
 
 n-11 w n-13^ n-14 v - / n-16 
 
 Table 1. Linear Feedback Logic 
 
Ik 
 
 <L> 
 
 
 H 
 
 
 £> 
 
 
 'ft 
 
 
 O 
 
 
 ^ 
 
 
 Pl, 
 
 
 ,d 
 
 w 
 
 ■p 
 
 £ 
 
 •H 
 
 o 
 
 > 
 
 •H 
 
 
 -P 
 
 Ih 
 
 •H 
 
 O 
 
 Tj 
 
 •P 
 
 £ 
 
 cd 
 
 O 
 
 Jh 
 
 U 
 
 0) 
 
 
 3 
 
 
 3 
 
 o 
 
 •H 
 
 
 ■P 
 
 flj 
 
 •H 
 
 w 
 
 d 
 
 ■H 
 
 H 
 
 O 
 
 
 !2j 
 
 T* 
 
 fr 
 
 ^ 
 
 Ct) 
 
 ^ 
 
 •9 
 
 -P 
 
 PQ 
 
 a 
 
 
 (1) 
 
 S 
 
 J 
 
 o 
 
 
 t3 
 
 s 
 
 § 
 
 s 
 
 fH 
 
 ■g 
 
 o 
 
 aJ 
 
 T> 
 
 S 
 
 2 
 
 
 0) 
 
 
 CO 
 
 
 CU 
 
 
 on 
 
 0) 
 
75 
 
 7«3 Pattern Memory and Driving Logic 
 
 The chosen random patterns for the white and red fields are written 
 into the respective memories. The memory is organized as l6 x l6 x k hits 
 or 256 words of k hits each. The Y-address selects one of the sixteen 
 SN7^89 memory chips and the X-address selects one of the 16 words of k hits 
 each in the selected chip. The semiconductor memory circuit is shown in 
 Figure J.h. The addressing logic for each memory is shown in Figure 7-5. 
 One of these two memories is selected every field hy enabling the respec- 
 tive SN7Ul5^ decoder chip. 
 
 7-U Read-Write Logic 
 
 The random pattern obtained from the pseudo-random generator is 
 written into the semiconductor memory in the raster format. This is done 
 hy the write addressing logic and the write memory buffer shown in Figure 
 7.6. The SN7^7^ D-type flipflops store the four successive outputs of 
 the random sequence generator. The clock applied to the X-Y counter is 
 H / h, so that after every four horizontal pulses the counter is advanced 
 by one ensuring that h new bits are now stored in the D-storage registers. 
 There is one set of storage registers and counters for the white field 
 memory and another set for the red field memory. The read-logic shown in 
 Figure 7-7 allows the pattern stored in the memory to be displayed syn- 
 chronously in the television raster format. The Vertical Counter CI 
 starts counting at the onset of the vertical sync pulse. After it counts 
 a programmed number of pulses H (say, maximum 16), the carry out pulse 
 clears M2 which starts the counting for C2 and C3. C2 is connected as a 
 divide by 15 counter. C3 is a k bit binary counter. In each field the 
 display memory is read after a maximum delay of l6H lines from the start 
 of vertical sync. After every 15H lines the next row in the memory is read. 
 
76 
 
 A A A A 
 
 * 
 
 UJ 
 
 5 
 A 
 
 CM »-• 
 
 # * 
 
 UJ UJ 
 
 A A 
 
 — i <\j to ^- UJ 
 </)(/) (O (/) z 
 
 zoo(0 
 
 S- CD 
 
 *H JM f> * . 
 
 O Q O O J 
 
 UJ^ 
 
 «-< CJ IO * UJ 
 
 ZCD C\I 
 
 H N IO » W' 
 
 Q Q O Q $ 
 
 y v v v 
 
 m ifi in m 
 + + + + 
 
 ^ n io ^ m 
 ^ «/) tn to 2 
 
 ^ m 
 
 -. N K) * UJ< 
 0$ 
 
 AAAAAAAAA 
 
 *-l CVJ fO ^ UJ 
 
 QOOO$<CDOO 
 
 a 
 
 S 
 
 o 
 -p 
 
 CJ 
 
 § 
 o 
 
 a; 
 
 •H 
 
77 
 
 74154 
 
 4-LINE TO 
 16-LINE 
 DECODER 
 
 14 
 13 
 12 
 11 
 10 
 9 
 6 
 7 
 6 
 5 
 4 
 3 
 2 
 1 
 
 470X1 
 
 Figure 7.5. Addressing Logi, 
 
78 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 7404 
 
 PSEUOO - RANDOM > 1 
 
 GENERATOR 1 
 
 
 
 
 
 
 
 
 
 
 X 
 
 4 
 2 S 
 
 7474 
 
 3 
 
 1 
 
 
 10 
 12 9 
 
 7474 
 
 11 
 
 13 
 
 
 4 
 2 5 
 
 7474 
 
 3 
 
 1 
 
 
 10 
 \2 9 
 
 7474 
 
 11 
 
 1? 
 
 
 OTHER T 
 PATTERN > 1 
 GENERATOR 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ' "" 
 
 
 
 
 
 
 
 
 
 mi 
 
 H 
 
 '« 1.2 T n »« 
 
 111! 
 
 "y •>. , 
 
 
 3 2 C 7 
 
 5 12 
 
 74193 
 
 14 
 
 
 3 2 C 7 
 
 S 12 
 
 74193 
 14 
 
 
 '4 * 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 '404 
 
 
 
 
 
 
 
 
 
 
 
 
 2 4 » 
 
 7474 
 
 S 
 
 1 
 
 
 10 
 12 » 
 
 7474 
 
 11 
 
 U 
 
 
 2 " 3 
 
 7474 
 
 3 
 
 I 
 
 
 12 «° 3 
 
 7474 
 
 11 
 
 13 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 "* w 
 
 
 
 
 
 
 
 
 
 
 
 *«*«*« *m 
 
 741»S 
 14 
 
 Itt Xw *!■ V M 
 
 nil mi 
 
 74193 
 
 14 
 
 -> CARRY " 
 
 Figure 7-6. Write Logic and Buffer 
 
79 
 
 VERTICAL > 
 
 *• CARRY 
 
 HORIZONTAL READ 
 LOGIC 
 
 ♦ CARRY 
 
 Figure 7.7. Read Logic 
 

 80 
 
 The horizontal counter starts counting after a controllable delay 
 by M3. After this delay, Mk is triggered which in turn triggers the cross- 
 coupled monostables M5 and M6. The pulse frequency can be adjusted by 
 varying the resistance R. When the counter has finished counting l6 pulses 
 the carry-out pulse clears MU which in turn inhibits the oscillation of 
 the cross-coupled multivibrators. 
 
 7. 5 Pattern Timing Logic 
 
 Different timing signals are derived in the Pattern timing logic 
 shown in Figure 7-8. The Read-Write DPDT switch selects whether data is 
 read out of the memory or written into it. In the write mode, Ml produces 
 a clear pulse to clear the white field memory write counter. M2 produces a 
 pulse equal to the duration for which the white field memory is to be written 
 into. The clear R pulse clears the red field memory write counter until 
 that time. W enables the write for the white field memory. The chip 
 enable decoder SN7^15^ is also enabled during this period. M3 produces a 
 pulse whose width is equal to the interval for which the red field memory 
 is written into. 
 
 7-6 D/A Converter and Video Mixer 
 
 The four bit digital information read from the semiconductor memory 
 is converted to an analog signal by the digital-to-analog converter. The 
 DAC consists of eight high-speed JFET analog gates described earlier in 
 Section 6.6 and a weighted resistance network connected at the input of a 
 uA715C high speed operational amplifier. The video mixer is another opera- 
 tional amplifier that mixes the image and the checkerboard pattern or out- 
 puts any one of them. The D/A converter and the video mixer are shown in 
 Figure 7-9. 
 
81 
 
 CARRY W > 
 CARRY R > 
 
 -> J 
 
 -[> J 
 
 -»■ A 
 
 -0- 
 
 J 
 
 -t> — On 
 
 SN7404 
 
 LS 
 
 ! 10K > 10K 
 
 lOO^F 
 
 ST 
 
 100 M F 
 
 ST 
 
 SN74L3 
 
 -»• B 
 
 -♦ C 
 
 Fl>- 
 
 Fl>- 
 
 s 
 
 O 8 
 
 ■to 
 
 o 
 
 SN7437 
 
 7400 
 
 -» WR 
 
 -*■ RO 
 
 *■ CLEAR R 
 
 Figure 7.8. Pattern Timing Logic 
 
82 
 
 MSB> 
 
 LSB> 
 
 VIDEO 
 
 Figure J. 9. D/A Converter and Video Mixer 
 
83 
 
 8. PENTECOST CAMERA SYSTEM 
 
 The Pentecost camera system generates a field-sequential color 
 signal using a single imaging tube and a color filter wheel driven "by a 
 synchronous motor. A schematic of the camera system is shown in Figure 8.1. 
 The light from the object passes through the filter wheel and is focused by 
 the lens system onto the face of the pick-up tube. The LED-photo-transistor 
 pair produces a signal indicating the position of the color wheel. This 
 signal is amplified by the photo-detector amplifier and then used to syn- 
 chronize the camera and derive the color field indicator signals. The 
 various components and subsystems of the camera are described in the fol- 
 lowing sections. 
 
 8. 1 Camera Tube 
 
 In a field-sequential color system like ours, it is essential that 
 we have a pick-up tube having fast response so that there is no carryover 
 of the signal from one field to the other. This and other considerations 
 that led to the selection of a Plumbicon camera tube and a black-and- 
 white camera system are described in Appendix B. 
 
 The Plumbicon is a vidicon type television pick-up tube that has a 
 photoconductive layer of lead-monoxide (PbO) instead of antimony trisul- 
 phide. This results in high sensitivity, negligible dark current, fast 
 response and independence from temperature variation. This photosensi- 
 tive layer acts as a p-i-n structure with the red PbO layer behaving as 
 an intrinsic layer. ' ' The vapor deposited PbO layer has crystallites 
 of dimension about ly x ly x O.ly. The small size of the crystals gives 
 
 /To 
 
 rise to numerous surface states and these surface states compensate the 
 
8U 
 
 
 
 
 a: 
 
 I 
 
 
 _l 
 
 UJ 
 UJ 
 
 
 
 
 
 
 * * 
 
 
 
 </) IT 
 
 
 
 
 3 
 
 u 
 
 
 
 
 G 
 
 Z 
 
 D 
 
 
 
 o 
 
 u. 
 
 
 
 
 K 
 
 
 
 
 
 X 
 
 K 
 
 
 
 
 o 
 
 O 
 
 
 
 
 z o 
 
 _J 
 
 
 
 
 >- UJ 
 
 o 
 
 
 
 
 w _l 
 
 o 
 
 
 
 
 
 hi 
 
 
 
 
 
 
 
 kVV WAV V VVVS^IF 
 
 | L2Ai^ 
 
 
 
 K II " 
 
 
 1 
 
 
 
 ^ 1 
 
 
 
 s 
 
 
 
 
 
 1 l\U<J 
 
 
 
 
 
 
 
 
 
 / 
 
 
 a 
 
 
 
 
 
 
 
 o 
 a. 
 
 < 
 
 , 1 
 
 1- 
 
 UJ 
 
 -J 
 
 
 UJ 
 
 
 
 
 
 z 
 
 
 
 
 < 
 
 UI 
 
 
 (- 
 
 UJ 
 
 < 
 
 CD 
 
 a 
 o 
 a. 
 
 £ 
 
 
 
 PHOTO - 
 DETECTOR 
 CIRCUIT 
 
 <r 
 
 »- 
 i 
 
 o 
 
 o 
 
 I 
 
 0. 
 
 z 2) 
 o ^ 
 ffi < 
 5 a: 
 
 u UJ 
 
 -J 2 
 
 a. < 
 
 o 
 
 
 < 
 _l 
 a. 
 
 o 
 
 z 
 1- 
 z 
 
 3 
 
 o 
 
 2 
 
 
 
 ' 
 
 
 
 
 
 
 
 
 
 n 
 
 
 
 
 
 
 " 
 
 
 
 
 
 
 Q 
 
 
 
 co 
 
 
 
 
 i uj <r ui 
 
 
 ,« •- 
 
 
 
 
 1 t- ui »- 
 
 
 o v> Z 
 
 
 
 
 K T 2 to 
 
 
 z •- _ UJ 
 
 
 
 
 y I < >■ 
 m 
 
 
 HOUSI 
 
 FOR 
 
 CIRCUI 
 
 AND 
 
 COMPON 
 
 
 
 
 o 
 
 ^ 4i 
 
 j t 
 
 
 
 o ♦■ 
 
 
 
 
 > 
 
 * 
 
 
 
 
 
 
 z 
 
 
 
 
 
 o 
 
 
 
 
 
 
 (- 
 
 
 
 
 
 
 < (0 
 
 
 
 
 
 
 NK 
 
 
 
 
 
 
 2 3 
 
 
 
 
 
 
 OO 
 
 a a: 
 
 z 
 to 
 
 
 
 
 
 
 
 " 
 
 
 
 
 
 
 K 
 
 
 
 
 
 ftop 
 
 
 
 
 
 
 o _i 5 
 
 -J UJ o 
 
 
 
 
 
 
 
 
 o — — 
 
 
 
 
 
 o u. o 
 
 
 
 
 
 z 
 
 
 
 
 
 ▼ 
 
 
 
 cr 
 
 
 
 ^ O 
 
 
 
 9 »- 
 
 
 
 i < 
 
 
 
 
 
 
 
 
 UJ O 
 
 
 
 
 
 
 P 
 H 
 
 P 
 
 •H 
 
 s 
 
 a) 
 
 -p 
 m 
 
 CO 
 
 cd 
 
 0) 
 
 o 
 ■p 
 
 CO 
 
 o 
 o 
 
 0) 
 
 p 
 
 CD 
 
 H 
 
 CO 
 
 
 
 •H 
 
 Pn 
 
 U.Q 
 
85 
 
 "bulk impurity giving rise to essentially intrinsic behaviour. The PbO 
 layer is deposited on a transparent electrode layer of SnO which acts as 
 the n-region of the p-i-n structure. A thin p layer is formed on the 
 other side of PbO layer. The SnO electrode is usually biased to around 
 UOV with respect to the cathode. The p-i-n layer is thus reverse biased 
 and very little dark current flows . 
 
 It should be mentioned that in the red region the response of the 
 Plumbicon decreases very rapidly. To enhance the absorption in the red 
 region the intrinsic thickness is increased to 10-20y. In red-extended 
 tubes this absorption is enhanced by creating a lead oxy-sulphide layer 
 having smaller bandgap. 
 
 The camera tube selected is a red Plumbicon (Amperex l6 x QRIG). 
 
 It employs a separate mesh construction in a l" diameter envelope. Some 
 
 6k 
 of the salient features of the tube are: 
 
 Signal Electrode Voltage: k5 volts 
 
 Dark current : 3 nA max 
 
 Sensitivity at Color Temperature of Illumination = 2850°K: 
 
 60 yA/lumen 
 
 Gamma Transfer characteristic: 0.95 + 0.05 
 
 Limiting Resolution: 600 TV lines 
 
 Highlight signal current: 0.1 uA 
 
 Lag: (measured with 100% signal current of 0.1 uA and with a 
 
 light source with color temperature 2850°K. Red filter 
 
 inserted) . 
 
 Residual signal after dark pulse of 50m sec: max 5% 
 
 Residual signal after dark pulse of 200m sec : max 2% 
 
86 
 
 Spectral Response: 
 
 Region of max response: U700-5200°A 
 Regions of 50% response: 5800 + 100°A 
 Cut-off: 6200 2. 
 
 8.2 Filter Wheel 
 
 The color filter wheel has two sections of red and green Wratten 
 filters sandwiched between two Plexiglass plates. It is driven by a 
 1/100 H.P. capacitor-start hysteresis synchronous motor- operating at 1800 
 rpm. The color wheel rotates at 30 rps to yield two fields per revolution 
 at the color frequency of 30 fps. As shown in Figure 8.1 the filter 
 wheel is enclosed in a housing on which the synchronous motor is mounted. 
 The whole assembly is mounted on the base plate of the tripod just in 
 front of the camera lens assembly. The filter wheel is large enough to 
 cover the entire camera lens area and sequentially passes the red and 
 green fields. The LED-Photo-transistor pair is mounted on this filter 
 assembly to detect the position of the filter wheel. 
 
 8.3 Photo-detector Circuit 
 
 Refer to Figure 8.2. Light from the red-emitting LED passes 
 through the rotating filter wheel to the photo-transistor. The photo- 
 transistor is connected as an emitter follower. Since the speed require- 
 ments are not stringent , the emitter load is kept high to have larger 
 output signal. The output voltage is amplified and then applied to an 
 output stage through an emitter follower to decrease the loading on the 
 previous stage. The output drives the schmitt trigger to give clean 
 pulses at 30 Hz. 
 
87 
 
 :u8 f 5Sl 
 
 u 
 o 
 -p 
 o 
 a; 
 -p 
 
 £ 
 i 
 
 o 
 +3 
 o 
 
 X! 
 
 PL. 
 
 CM 
 0O 
 
 1 
 
 •H 
 
88 
 
 8.U Camera Synchronization Circuit 
 
 Rather than synchronize the hysteresis synchronous motor with the 
 camera sync signal it was decided to use the field signal from the photo- 
 detector circuit to synchronize the camera system with the motor. This is 
 accomplished by modifying the 2:1 interlace synchronization circuit shown 
 in Figure 8.3- This has a free-running multivibrator having a frequency 
 of 31.5 kHz. Dividing this frequency by two gives the horizontal fre- 
 quency. It is also divided by a chain of ten divide-by-two counters with 
 proper gating to get a division of 525 » thus giving at the output a sixty 
 hertz vertical signal. This 60 Hz signal is compared with 26V A.C. de- 
 rived from the mains and depending upon their phase difference we get an 
 output at the phase comparator which is then integrated and applied 
 through an emitter follower to the base return resistor of the free-running 
 multivibrator, which in turn changes the frequency of the multivibrator. 
 Thus the vertical and the horizontal signals are locked to the power fre- 
 quency. 
 
 To synchronize the filter wheel with the camera we replace the 
 26V A.C. signal with one derived from the photodetector circuit. To accom- 
 plish this we derive sixty-cycle pulses denoting the start of the red and 
 the green fields that are derived from the photo-detector signal. This 
 signal is then applied to the integrator which produces a 60 Hz sawtooth 
 signal that are used in place of the 26V A.C. input. This part of the 
 synchronization circuit is shown in Figure 8.U. R and R allow us to 
 vary the amplitude and the dc-level, respectively, of this sawtooth output. 
 
 8-5 Color Field Indicator 
 
 The color field indicator circuit is -shown in Figure 8.5- At the 
 output of this circuit we get a thirty hertz TTL signal with the high level 
 
89 
 
 5 
 
 g 
 
 s 
 
 
 o 
 
 i 
 
 it 
 
 o 
 
 s 
 
 
 Q 
 
 s 
 
 1^ 
 
 m 
 
 * 
 
 U". 
 
 i_l _LJ_±_!_± J. 
 
 [~k»J» in * r~ • • 
 
 S Si 3i 
 a ;5 Si 
 
 I gs If 
 
 S !* ^5 
 3 iS is 
 
 u 
 tr 
 
 u 
 o 
 
 PQ 
 
 U 
 O 
 
 g 
 
 o 
 
 R 
 
 CO 
 
 00 
 
 ao 
 
 (L) 
 
 
90 
 
 -p 
 
 •H 
 
 o 
 
 u 
 
 •H 
 O 
 
 a 
 o 
 
 •H 
 -P 
 
 ctf 
 
 tsl 
 •H 
 
 d 
 o 
 
 o 
 
 4} 
 
 O 
 
 CO 
 
 
 "+H ft W 
 
91 
 
 _TL 
 
 VSB 
 
 40 K 
 
 IOK 
 -VW- 
 
 + 5 
 
 Lȣ -^ 5 4148 
 IOK 
 
 16 14 19 
 
 74123 
 M I 
 
 FILTER 
 
 POSITION 
 
 SIGNAL 
 
 n 
 
 _n_r 
 
 X> 
 
 30K 
 
 7400 
 
 If 
 
 7400 
 
 o 
 
 -»■ 4 
 
 « »• 16 
 
 Or- 5 
 
 <t^ 
 
 
 
 =D 
 
 6 
 
 vO 
 
 _n_r 
 
 u 
 
 Figure 8.5. Color Field Indicator 
 
92 
 
 denoting the green field and the lov level denoting the red field. These 
 field signals start at the beginning of vertical blanking so that all the 
 switching that is controlled by this signal in the receiver system has the 
 blanking time to complete their transitions. The vertical system blanking 
 obtained from the camera system is level-shifted and then applied to the 
 monostable Ml which produces a blanking pulse whose width can be controlled. 
 The vertical blanking signal selected by the SPDT switch is inverted and 
 then applied to the SN7^-00 gates. The other input to these NAND gates is 
 the filter position signal obtained from the photo-detector circuit. At 
 the output of the upper NAND gate we have a pulse whose negative transi- 
 tion refers to the start of the green field. At the output of the lower 
 NAND gate we have a pulse whose negative transition refers to the start 
 of the red field. These two pulses are respectively applied to the nega- 
 tive going gated input and the clear input. At the Q output of M2 we get 
 a pulse whose high level refers to the green field. The SPDT switch allows 
 us to select the normal color field signal or a complimentary field signal 
 which controls the receiver system such that the camera red field is pro- 
 jected on the white phosphor and the camera green field is projected on 
 the red phosphor. This corresponds to the television analog of color 
 reversal effect in Land two-color projections. 
 
93 
 
 9. ELECTRONICALLY CONTROLLED COLOR FILTERS 
 
 There have been several attempts in recent years at making a 
 color television camera using a single pick-up tube. One such camera, 
 developed by RCA, uses a patterned dichroic filter of red, green and blue 
 strips area-sharing at the front end. Another color camera using a sin- 
 gle pick-up tube is the Apollo-11 color camera which uses a color filter- 
 wheel. The advances in Ferroelectrics and other solid state materials has 
 spurred interest in having a solid state filter that can be controlled 
 electronically and used in place of the color wheel. 
 
 9.1 Solid State Color Filters 
 
 Presently, there are three possible types of such solid state fil- 
 ters. They are: l) Voltage tunable interference filter 
 
 2) Ferroelectric Ceramic filter 
 
 3) Ferroelectric ferroelastic filter 
 
 Voltage-tunable Interference Filter: 
 
 These are voltage-tunable interference filters such as the 
 Optichron. These types of filters are made from some piezoelectric 
 crystal such as Barium Titanate that show changes in the length of the 
 sample upon application of an electric field of the order of 10 volts/cm. 
 The operation of the filter can be explained in terms of the tetrogonal 
 distortion of the crystallites. An unpolarized sample has its crystallites 
 polarized in more or less arbitrary directions. When a field is applied, 
 some of the crystallites become oriented in directions nearer that of the 
 field. If such a sample is used as the element of an interference filter, 
 then the thickness of the sample can be changed with the application of the 
 
9h 
 
 field and consequently the transmission characteristics change with thick- 
 ness. The Optichron passes a different color upon application of a voltage. 
 Cascading two such Optichrons along with application of a suitable voltage 
 to each of them allows us to selectively pass the red, green or "blue por- 
 tion of the spectrum. 
 
 9-2 Ferroelectric-Ceramic Filter 
 
 Another promising possibility is to construct a filter of thin 
 polished plates of hot-pressed rhombohedral lead-zirconate lead-titanate 
 (PLZT) ferroelectric ceramics. Coarse-grained ceramics that have 
 grain diameters greater than about two microns have light scattering 
 properties that can be varied by switching the orientation of the ferro- 
 electric polarization. Fine grained ceramics having grain diameters less 
 than about two microns have an effective birefringence that can be varied 
 either by applying a biasing field and/or by partially switching the 
 ferroelectric polarization. 
 
 When a dc electric field of sufficient strength is applied to a 
 ferroelectric ceramic, domains favorably oriented with respect to the 
 field grow. When the field is removed, the ceramic retains part or all 
 of this polarization. Under these circumstances it is said to be electri- 
 cally poled. The direction of poling is the ceramic polar axis. By 
 poling these fine-grained ceramics, the ceramic polar axis and the optic 
 axis are made coincident and they become optically uniaxial crystals hav- 
 ing refractive indices n and n . n is the refractive index for light 
 
 e o e 
 
 with its electric vector parallel to the optic axis, and n , the refractive 
 
 index for light with its electric vector perpendicular to the optic axis. 
 
 Birefringence, An, is defined as n - n . Retardation R, the path differ- 
 
 e o 
 
 ence between the two components for any incident polarized light is 
 
95 
 
 R = (n -n )t = tAn 
 e o 
 
 for plate thickness t. An is the macroscopic birefringence which is mea- 
 sured to be about 0.6U An. Since An can be varied by orienting the domains 
 with an electric field, the retardation is under electrical control. (The 
 effective binefringence An also depends on the incident light wavelength). 
 A device made of electrically controlled fine-grained ceramics may 
 be configurated as shown in Figure 9*1 where y is the poling axis and ¥ is 
 the polarization angle of incident light. If is the angle at which the 
 analyzer is oriented, then the transmittance for = 90° is 
 
 m _ o- 2 RlT 
 
 T = Sin — r 
 
 A 
 
 for = 0°, T = Cos 2 £f 
 
 A 
 
 The retardation is varied by adjusting the remanent polarization or bias 
 
 field. If 
 
 R = NX, T = for = 90° 
 
 T = 1 for = 0° 
 
 N = 0,1,2.. . 
 R = (N-1/2)a, T = 1 for = 90° 
 
 T = for = 0° 
 
 If white light is incident, those wavelengths for which R = (N-l/2)A 
 
 will be preferentially transmitted when = 90° and those wavelengths for 
 
 which R = NX will be preferentially transmitted when 0=0°. 
 
 As the ceramic polarization is changed either by incremental 
 
 switching or applying a bias field, the resulting change in retardation 
 
 will change the preferentially transmitted color. For example, if the 
 
 ceramic plate transmits red at P D , it may be switched incrementally to 
 
 transmit orange, yellow, green or white as shown in Figure 9.2. 
 
96 
 
 PULSE 
 GENERATOR 
 
 CERAMIC PLATE 
 (POLAR AXIS P 
 
 ANALYZER 
 
 Figure 9.1- Ferroelectric Ceramic Filter 
 
97 
 
 o 1.0 
 
 I 0.5 
 
 < 
 
 o 
 
 Q- n 
 
 > 
 
 5 0.5 1- 
 
 UJ 
 
 or 
 
 1.0 •- 
 
 / 
 
 RED 
 
 ORANGE 
 
 YELLOW 
 
 GREEN 
 
 WHITE 
 
 Figure 9.2. Hysteresis Loop of Ferroelectric Ceramic 
 

 98 
 
 Some of the limitations of fine-grained ceramics are: 
 
 1. Fine-grained ceramics also scatter transmitted light. This 
 is undesirable in that it tends to obscure birefringence. It 
 is thought that the dominant scattering mechanism in fine- 
 grained ceramics is the scattering from inclusions, impuri- 
 ties, voids, and structural discontinuities at grain boundaries. 
 
 2. The transmission ratio is only in the range of to h5% in the 
 visible range. 
 
 3. Domain switching processes can take place in the ceramic in 
 times of the order of microseconds. But questions still re- 
 main of whether large enough amounts of polarization can be 
 switched fast enough with reasonable voltages and without 
 causing the sample to fracture. 
 
 An attempt was made to obtain a sample of a ferroelectric 
 ceramic filter from Bell Telephone Laboratories or Sandia 
 Laboratories, but we were unsuccessful in getting any such 
 sample for our experimentation. 
 
 9.3 Theory of the Ferroelectric-Ferroelastic Color Filter 
 
 70 
 According to Aizu, "A crystal which is simultaneously ferro- 
 electric as well as ferroelastic and in which all polar properties of 
 every even rank tensor can be changed by the application of an electric 
 field or a mechanical stress is said to be ferroelectric-ferroelastic." 
 
 Godolinium molybdate (Gd_(M 0, )„) has been observed to have the 
 
 2 o 4 3 
 
 71—71). 
 
 ferroelectric-ferroelastic effect and belongs to the same species 
 
 as KHgPO^. Above T , Gd (M 0, ) is optically uniaxial with a tetragonal 
 space group, and below T it is optically biaxial. It has the following 
 
99 
 
 optical properties. 
 
 f> P 
 
 n = 1.8U09 + 1.917 X 10 X 
 
 c 
 
 C _p o 
 
 n a n, = 1.7950 + 1.6500 X 10 X with A in A 
 
 Since n > n * n, , Gd p (M 0i )' is optically positive as contrasted with 
 
 KH PO, which is optically negative. The value of "birefringence An = 
 
 n, - n was obtained to be -U.08 x 10 at room temperature. 
 
 The birefringence, An , in the biaxial phase, occurs opposite in 
 
 sign to electric field E, and mechanical stress X . When the electric 
 
 field or the mechanical stress is reversed, An diminishes in value. This 
 
 s 
 
 is shown in Figure 9-3. The existance of bistable states is utilized to 
 make the color modulators in the following way. One birefringent crystal 
 such as Rochelle salt having retardation R is cascaded with two MOG plates 
 having retardation R_ n and R_, . The total retardation R is then 
 
 R = \ t R G1 ± R G2 - 
 
 By properly selecting the plate voltages we could make, 
 R x = R q + R Q1 + R Q2 = 1313my (= | x 525my) 
 
 R 2 = R Q + R G1 - R Q2 = 1138my (= | x U55my) 
 
 R 3 = R q - R G1 - R G2 = 975my (= | x 650my). 
 
 If the angle at which the analyzer is oriented with respect to the 
 
 polarizer is 90°, then the emerging light is 
 
 T T „. 2 Rtt 
 1 = 1 Sin - — . 
 o X 
 o 
 
 So, green (525my), blue (U55my) and red (650my) corresponding to the 
 
 retardations R , R and R may be transmitted. Since R = tAn, knowing 
 
 the required R and An we can calculate the plate thicknesses. Solving 
 
100 
 
 r 
 
 -4 
 
 -4 
 L 
 
 -2 
 
 2 
 
 A nob 
 
 -4 
 XIO 
 
 50 r- 
 
 25 
 
 -25 
 
 -sot 
 
 E(W) s 
 
 2 4X10 
 
 2 
 
 4XI0 9 
 
 X(N/rn ) 
 
 Figure 9.3. Hysteresis. Loop for Gd (m 0, ) 
 
101 
 
 the above equations we find, R = llUUmy , R_ = 195my and R_, = 210my. 
 Plate thicknesses corresponding to these retardation values are 133y, 
 195y and 210y. The next section describes such a filter obtained from 
 Hitachi Ltd. of Japan. 
 
 9.h Hitachi Gadolinium Molybdate Color Modulator 
 
 The Central Research Laboratory, Hitachi Ltd. of Japan generously 
 donated to us two samples of the Gadolinium Molybdate solid state filter 
 
 for experimentation and possible application in our system. The cross- 
 
 75 
 
 section of filter sample No. 1 is shown in Figure 9.h. As noted in the 
 
 previous section, the modulator consists of a polarizer, two Gd_(M0, ) 
 plates whose birefringence is controlled, a retarder plate and an analyzer. 
 
102 
 
 IMPINGING WHITE 
 LIGHT 
 
 READ WIRES 
 
 POLARIZER 
 
 RETARDER 
 ANALYSER 
 
 --*> OUT GOING LIGHT 
 « R ,B AND G 
 
 MOG X'TALS 
 
 READ WIRES 
 
 Figure 9.^. Cross -section of MOG Color Modulator 
 
103 
 
 10. SYSTEM PERFORMANCE AND RESULTS 
 
 In this section we evaluate the PENTECOST system performance and 
 note some of the experimental results. The specimen color modulators are 
 also evaluated. 
 
 10.1 PENTECOST Receiver 
 
 The PENTECOST Receiver is shown in Figure 10.1. The high voltage 
 section of the screen voltage switch and the focus voltage switch are housed 
 below the Penetron tube. The CPS high voltage supply is mounted on the rear 
 near the high voltage switch enclosure. It has provision for varying the 
 output voltage from 10KV to 20KV. A voltage of about 17KV is used as the 
 supply voltage for the screen voltage switch. The pattern generator cir- 
 cuits are mounted on a Printed Circuit Board rack just above the Conrac 
 black-and-white monitor. The gain correction circuits and the high vol- 
 tage switch driver circuits are mounted on another PCB rack below the 
 Conrac monitor. The lower two racks house various low voltage supplies 
 for +5V, +15V, -15V, +75V and -150V. The bottom rack is the focus voltage 
 supply which has provision for varying the output voltage from 0V to 
 3100V. A voltage of 3KV is used as the supply voltage for the focus vol- 
 tage switch. 
 
 The screen voltage switching circuit was very carefully adjusted 
 and then packaged to achieve freedom from high voltage breakdown and 
 corona effects. There was difficulty in achieving the high voltage switch- 
 ing within the vertical blanking period. Difficulties were also encoun- 
 tered in adjusting the vertical and the horizontal registration circuits. 
 One needs to adjust the magnitude of the deflecting signal to control the 
 
PENTECOST 
 
 ioU 
 
 Figure 10.1. Pentecost Receiver 
 
105 
 
 size. But this in turn changes the dc biasing point. A compensating dc 
 voltage was applied every alternate field for proper centering of the red 
 and white displays. Video amplitude switching was necessary for compen- 
 sating the contrast variation for the two anode voltages, so that flicker 
 is minimum. In addition, the Plumbicon tube has rather poor sensitivity 
 in the red region, and hence the video amplitude needs to be compensated. 
 The waveforms of various receiver signals are shown in Figure 10.2. 
 
 PENTECOST Receiver Controls 
 
 SCREEN HV UP : Controls the upper limit of the screen voltage correspond- 
 ing to the excitation of the white phosphor. The screen voltage should be 
 adjusted around l6KV for a white display. 
 
 SCREEN HV LOW : Controls the lower limit of the screen voltage correspond- 
 ing to the excitation of the red phosphor. The screen voltage should be 
 adjusted around 11KV for a red display. 
 
 VIDEO GAIN : The screw driver adjustment sets the relative contrast of the 
 red and white fields. By changing the position of the Video Field Indicator 
 switch just below the video gain control, the relative contrast of either 
 the red or the white field can be controlled. In normal circumstances, the 
 red video signal must be increased to compensate for the smaller signal 
 from the Plumbicon camera and also for the decrease in contrast due to the 
 lower screen voltage. The Video switch above the video gain control either 
 routes the video through the video gain correction circuit or bypasses it 
 completely. 
 
 FOCUS HV UP : Controls the upper limit of the focus voltage corresponding 
 to the excitation of the white phosphor with screen voltage of l6KV. The 
 upper focus voltage should be adjusted around 2200V. 
 FOCUS HV LOW : Controls the lower limit of the focus voltage corresponding 
 
106 
 
 GREEN 
 
 +5 1 
 
 
 _J 
 
 RED 
 
 
 
 U 1/60 SEC. . 
 
 1/60 SEC. j FIELD INDICATOR 
 
 P T 
 
 ~^ SIGNAL 
 
 r»w 
 
 
 
 HORIZONTAL 
 
 
 DRIVE SIGNAL 
 
 -100 V 
 
 
 VERTICAL 
 DRIVE SIGNAL 
 
 BRIGHTNESS 
 
 ft BLANKING SIGNAL 
 
 VIDEO 
 
 DRIVE SIGNAL 
 
 Figure 1Q.2. Receiver Maveforms 
 
107 
 
 to the excitation of the red phosphor with screen voltage of 11KV. The 
 lower screen voltage should be adjusted around 1500V. 
 VERTICAL HEIGHT : Controls the height of the red and green fields. 
 RED VERTICAL GAIN : Controls the vertical size of the red field. 
 RED VERTICAL CENTERING : Controls in the centering of the red field on the 
 white field by the application of a proper dc correction signal. 
 BRIGHTNESS FIELD INDICATOR SWITCH : This SPDT switch applies either FI or 
 FI to control the relative brightness of the red and white fields. If FI 
 is applied the brightness controls are as indicated below. If FI is 
 applied the red and white controls are exchanged. 
 
 BRIGHTNESS RED : This increases the relative brightness of the red field 
 if turned in the CW direction. 
 
 BRIGHTNESS WHITE : This increases the relative brightness of the white 
 field if turned in the CCW direction. 
 
 BRIGHTNESS SWITCH : This switch applies either the brightness signal de- 
 rived in the brightness correction circuit or the brightness signal ob- 
 tained from the Conrac monitor. 
 
 CONRAC MONITOR CONTROLS 
 
 Most of the Conrac monitor controls are used in conjunction with 
 the PENTECOST Receiver controls. The FOCUS control is not used as a 
 separate high voltage focus supply is used. The focus controls are as 
 described earlier in this section. The BRIGHTNESS and CONTRAST Controls 
 are used along with the PENTECOST Receiver controls to achieve the proper 
 relative brightness and contrast levels. The Conrac monitor controls 
 change both the red and white fields whereas the PENTECOST Receiver con- 
 trols selectively control the red and white field displays. 
 
108 
 
 It should be recognized that some of the PENTECOST system controls 
 interact with one another. If one adjusts the "brightness levels, the 
 screen voltage also changes a little and consequently changes the regis- 
 tration to some extent. Thus, the system parameters should be adjusted 
 for the best performance and left as it is. 
 
 10.2 PENTECOST Camera 
 
 The picture of the PENTECOST camera is shown in Figure 10.3. The 
 black-and-white camera with the Plumbicon tube and the filter wheel assem- 
 bly is mounted on a base plate screwed to the tripod. The synchroniza- 
 tion and the field indicator circuits are housed in a box mounted below 
 the base plate toward the rear of the camera. The camera system, includ- 
 ing the filter wheel assembly and the synchronizing circuits, performs 
 quite reliably. The original vidicon camera did not have any gamma control 
 circuit. The Plumbicon has unity gamma characteristics (y = A log V/A log L; 
 V = signal, L = Luminance) whereas the cathode ray tube luminance output is 
 proportional to the square of the signal. So the dynamic range of the 
 picture is limited. 
 
 The Plumbicon tube is fast enough for this application though there 
 is some carryover from the previous field. This carryover from the pre- 
 vious field tends to mix the red and green camera signals and consequently 
 obscure the Land effect by the percentage of the carryover. The sensi- 
 tivity of the Plumbicon in the red region is very limited. This reduces 
 the magnitude of the signal when the red filter is in front of the camera. 
 This is compensated at the receiver by switching the amplitude of the 
 video for alternate fields. 
 
109 
 
 Figure 10.3. Pentecost Camera 
 
110 
 
 There is some halo effect. Total reflection in the glass face- 
 plate of the Plumb icon tube produces an anular brightening of the picture 
 around highlights. The dark current of the Plumb icon is very small. 
 Hence, the noise in the picture is very small even if the video gain is 
 increased to the maximum. 
 
 Camera Controls 
 
 SYNC ADJUST : Adjusts the synchronization of the camera system with the 
 photodetector signal derived from the position of the filter-wheel. 
 FILTER SYNC (POWER SYNC) : This switch chooses either the sync signal de- 
 rived from the filter wheel position or the 26 volt AC 60c/s power fre- 
 quency to synchronize the 2:1 interlace oscillator circuit. 
 VSBR (CAMERA VSB) : This switch chooses either the vertical system blank- 
 ing derived in the auxiliary synchronization circuit or the one derived in 
 the camera system. 
 
 NORMAL (REVERSAL) : This switch selects the Field Indicator signal, (Fl), 
 or its complement, FI, to control the receiver switching circuits. In the 
 NORMAL position FI is sent and it displays the camera red field on red 
 phosphor and camera green field on white phosphor. In the REVERSAL posi- 
 tion FI is sent and it displays the camera red field on the white phosphor 
 and camera green field on the red phosphor. This corresponds to color 
 reversal in the Land color system. 
 
 FIELD INDICATOR : This switch selects either the field signal chosen by the 
 NORMAL (REVERSAL) switch or the signal level (+5, or 0) chosen by the SPDT 
 switch above it. By sending either +5V or as the receiver control sig- 
 nal, one displays the picture respectively on the white phosphor or the 
 red phosphor. 
 
Ill 
 
 10.3 Color Modulator Testing 
 
 The color modulators No. 1 and No. 2 were mounted with stands and 
 white light was directed with the help of lenses through the aperture of 
 the modulators. Light coming out of the modulators was projected on a 
 white screen. The driving signals required for the two specimens are 
 shown in Figure 10. h. For static conditions (x > 10 sec) the value of V 
 is 30V - 50V for Modulator No. 1 and 50V ~ 80V for Modulator No. 2. For 
 dynamic switching the amplitude of the signal applied is unrestricted. 
 We apply + 50V to Modulator No. 1 and + 75V to Modulator No. 2 for d.c. 
 testing. A switching signal of +_ 75V is applied to both the modulators 
 for dynamic switching. The driving circuits are shown in Figure 10.5. 
 
 The blue and green colors were fairly good, but the red color 
 obtained was more purple for Modulator No. 1 and yellowish for Modulator 
 No. 2. The transmitted light depends on the angle of incidence. If one 
 looks through the modulator to an incandescent source the color seen is 
 quite dependent on the angle looked at. If a switching signal is applied 
 to Modulator No. 1, one could see the color boundary moving at the transi- 
 tion point. No. 2 was tested for transition between red and green trans- 
 mittance conditions. An application of a signal of +_ 75V to pin C 
 (A = 75V; B, D = 0V) results in the transmission of either red light or 
 green light. The transition time is in the range of milliseconds. If 
 insufficient time is allowed for the transition from one color to the 
 other, the color remains at a color corresponding to the dc average of 
 the driving voltage. 
 
 The specimen filters were not incorporated into the system because 
 of the smallness of their aperture (lcm dia circle) and the consequent 
 need for an optical assembly to match this aperture to the present system. 
 
112 
 
 DRIVE MODE FOR MODULATOR NO.l 
 
 B=0, C=0 
 
 +V 
 
 -V 
 
 +V 
 
 i 
 i 
 
 < 1, ►!* % ► 
 
 <-. — t, ► 
 
 -v 
 
 RED- 
 
 I i 
 
 ►»*— GREEN -►*- 
 i i 
 
 BLUE 
 
 DRIVE MODE FOR MODULATOR NO. 2 
 
 B=0,D=0 
 
 Figure 10. k. Drive Signals for the Color Modulators 
 
113 
 
 75V > 
 
 -► 75V 
 
 15K 
 
 -► 50V 
 
 47 K 
 
 50V 
 
 -150V > 
 
 ■> -150V 
 
 30K 
 
 8.2K 
 
 10K 
 
 •75V 
 
 CONTROLLING 
 
 SIGNAL FROM 
 
 PULSE 
 
 GENERATOR 
 
 18K 
 
 8K 
 -vw- 
 
 -50V 
 
 -150V 75V 
 
 100K 
 
 75V 
 
 50 K 
 
 40327 
 
 1N3070 
 
 75V -75V 
 
 -►-50V 
 
 ±75V 
 PULSES 
 
 Figure 10.5. Color Modulator Driving Circuits 
 
llU 
 
 There will "be some loss in the transmission through the optics and the 
 modulator. Since the filter transmission colors (red and green) were 
 fixed and no control over individual colors was provided, there is less 
 flexibility than might be desirable. This is due to the bistable charac- 
 teristic of the MOG crystal. From this point of view the ferroelectric 
 ceramics offer more flexibility for controling the range of colors. In 
 MOG modulators one needs to have n number of plates to control the color 
 pass band. This decreases the transmission and also necessitates con- 
 trolling n different plates. Since the ferroelectric ferrostatic modula- 
 tors are single crystals one expects that modulators of very large aperture 
 will be difficult to make. The ferroelectric ceramic materials should not 
 have this size limitation as they are polycrystalline fine grained materi- 
 als. 
 
 10. k Land Color Experiments 
 
 Various Land experiments were made using the PENTECOST system. 
 The color perceived on the CRT was a strong function of the surrounding 
 illumination near the screen and the illuminating light on the objects 
 and their reflectance. The relative intensity of the two fields had to be 
 adjusted carefully. All of the projection experiments of Section k.O were 
 repeated. We observed red, green, and blue colors though the green and 
 blue were not very saturated. The blue perceived was quite grayish, and 
 towards the green end of the spectrum. It is extremely difficult to ob- 
 serve yellow. At very low light levels one could see some faint yellow. 
 The color reversal effect was observed by projecting the camera red scene 
 on the white phosphor and the green scene on the red phosphor. One ob- 
 serves blue-green for red objects and magnet a for green and blue objects. 
 The blue color was more prevalent in the reverse color display. A neutral 
 
115 
 
 density Mondrian kept in front of the camera did not produce significant 
 change in the perceived colors. 
 
 Motion in the picture had no detrimental effect. We did not ex- 
 periment with elaborate moving pictures because of the lack of a proper 
 synchronizing interface with a commerical color TV. However, slightly mov- 
 ing objects were introduced into the scene and they did not obscure the 
 color perceived. Masking different areas of the scene tended to decrease 
 the degree of the Land effect but was not so pronounced as to completely 
 eliminate the effect. The Land effect becomes enhanced when one adapts to 
 the picture. This is particularly true when one knows the colors of the 
 objects. The Land phenomenon is a large field effect where various psycho- 
 physical factors interplay. However, it was observed that a decrease in 
 the field size did not degrade the colors too any extent. By using a h x k 
 checker board pattern, one could observe colors such as green and blue in 
 addition to red by properly adjusting the relative intensity. The color 
 perceived deteriorated for a color grid pattern of l6 x l6. It appeared 
 to be a matrix of different gray squares under these circumstances. For 
 this reason, the pattern display system designed for implementing the 
 l6 x l6 pattern was not employed. There was no means for correlating the 
 color perceived on the screen to real object colors when the pseudo- 
 random patterns were used. 
 
 One could not optimize all of the parameters to perceive all the 
 colors equally well. If one adjusted the parameters to get the best yellow 
 effect, the green, blue and red were degraded. In an adaptive Land system, 
 each spatial region must be independently controlled in order to get the 
 I best color effect. 
 
116 
 
 Summer i zing, the Land color display system provides a fast and 
 flexible tool for conducting various Land two-color experiments "but the 
 degree of Land effect is usually less than in the projection set-up. This 
 decrease in the effect is due to non-ideal electronic devices and the need 
 to match the camera and receiver characteristics closely. 
 
liT 
 
 11. CONCLUSION 
 
 The PENTECOST Land color display system described affords a very 
 convenient set-up for conducting various experiments in two-primary color 
 perception. The receiver spectral characteristics can he controlled to 
 certain degree. If the ferroelectric ferroelastic or the ferroceramic 
 filter is employed, one could electrically control the exact filtering 
 characteristic for the two fields. A good two-color display must employ 
 an optimum choice for the filters, and the receiver characteristics. 
 
 The system performance can be improved by incorporating a gamma 
 control to increase the dynamic range. One could also consider a silicon 
 matrix tube for such an application since during the last two years they 
 have been improved to avoid the occurance of black spots. They have a lag 
 response similar to Plumbicons but their sensitivity and spectral charac- 
 teristics are much better. Their spectral response extends all the way to 
 the far infra-red region. 
 
 The Penetron tube is certainly a viable display media but the 
 switching and registration problems still make it troublesome to adjust. 
 The multigun Penetron and the current sensitive multilayer CRT might allevi- 
 ate some of these problems, though they, too, introduce their own peculiar 
 problems. Penetron color tubes are being increasingly used in information 
 display systems, such as, air-traffic control displays, computer driven 
 displays and computer graphics. In addition to its simple and rugged 
 construction, some important features like the availability of high resolu- 
 tion, freedom from stray magnetic fields, and the insensitivity to shock 
 and vibration make it a very powerful device for information display pur- 
 poses. The ones most widely used are the red-green and red-white two- 
 layered Penetrons. A red-white Penetron display could produce monochrome 
 
118 
 
 pictures in addition to displaying alphanumeric information. One could 
 also employ Land techniques to produce color pictures on such a system. 
 Since low saturation colors abound in information systems, we would have 
 less trouble in reproducing these by Land techniques. 
 
 The receiver characteristics, the camera filtering properties, 
 and the relative brightness and contrast levels of the fields should be 
 automatically adjusted to achieve the best Land color effect for widely 
 different scenes. In addition, the camera and receiver characteristics 
 should be spatially optimized so that one may obtain the best color repro- 
 duction possible in all regions. This, of course, increases the complexity 
 of the system but this is the way our visual system works — it adapts its 
 capabilities and characteristics to suit different conditions. 
 
 Some thorough psychological experiments are required to take care 
 of the difference in observer characteristics. The results can be com- 
 pared with the results derived from applying the Helson-Judd formulation. 
 It appears that color constancy and the Helson-Judd formulation in fact 
 don't answer all the interesting questions arising from Land effect. We 
 are still far from understanding the visual system and work should be 
 pursued vigorously in this field. 
 
119 
 
 APPENDIX A 
 
 CATHODE RAY TUBE SELECTION 
 
 Different possible schemes for implementing the receiver of the 
 two-color TV system are described below. A cost comparison of the dif- 
 ferent schemes is made. Anticipated technical difficulties are noted. 
 
 1. Single-gun Penetron : Voltage switching of the beam is employed 
 to change the color. Saturated colors are difficult to obtain. 
 
 The manufacturers of the Penetron, along with the price and the 
 characteristics of their products are listed below in Table 1. 
 
 2. Two-gun Penetron : The main problem with the single gun penetron 
 tube is switching the anode voltage. A multiple gun CRT with the 
 guns held at different potentials obviates the requirement for 
 voltage switching. A partial deflection-yoke-shielding technique 
 
 on the lower-voltage guns is used to automatically compensate for 
 the increased deflection sensitivity. 
 The potential suppliers are: 
 
 a. Zenith Radio Corporation, Rauland Division (l6" flat tube, 12 - 
 18KV anode voltage, delivery time = 90 - 120 days, price = 
 $1600.00) 
 
 b. RCA (in development stage). 
 
 Focus voltage switching circuits, video gain change and brightness 
 correction circuits will be, of course, required in this scheme. 
 
 3. Current Sensitive Single-gun color CRT : By combining a super- 
 linear phosphor of one color with a linear or sublinear phosphor 
 of a different emission color, a phosphor screen can be obtained 
 which changes color as a function of current density. Two problem 
 
120 
 
 en 
 a 
 o 
 u 
 p 
 cu 
 a 
 cu 
 Pu 
 
 w 
 i 
 
 a; 
 
 H 
 
 bo 
 
 a 
 
 <P 
 O 
 
 G 
 O 
 en 
 
 •H 
 fn 
 cd 
 
 O 
 
 u 
 
 i 
 
 p 
 
 LT\ 
 
 to 
 
 CM 
 
 o 
 
 C\J 
 
 u 
 
 
 
 -ee- 
 
 
 ^~. 
 
 
 <U 
 
 
 ►J -p 
 
 
 <H -H 
 
 a; 
 
 o 
 
 O 3 
 
 
 00 ^ 
 
 
 ■ l-l 
 
 J — 
 
 g 
 
 «H -d 
 
 P 
 
 a; 
 
 ^ 
 
 la « 
 
 
 H — ' 
 
 ^ 
 
 
 
 H 
 
 cu 
 c 
 
 ■g 
 
 •H 
 
 
 rf 
 
 O 
 
 
 H 
 
 > 
 
 > 
 
 
 O 
 
 W 
 
 o 
 
 3 
 
 -* 
 
 o 
 
 C\J 
 
 o 
 
 1 
 
 fe 
 
 o 
 
 
 NO 
 
 
 0O 
 
 o 
 o 
 oo 
 
 
 cu 
 
 
 —* -p 
 
 > 
 
 T3 -H 
 
 xj 
 
 0) xj 
 
 « 5 
 
 bO 
 
 
 55 
 
 PI 
 
 J 
 
 XI 
 
 <H 
 
 cp 
 
 cd 
 
 H 
 
 CO 
 
 CM 
 
 •H 
 
 -a- 
 
 _3" 
 
 cd 
 
 c5 
 
 PI 
 
 PI 
 
 
 u 
 
 «H 
 
 O 
 en 
 
 -3- 
 
 H 
 
 H 
 
 CM 
 
 CM 
 
 cd 
 
 
 H 
 
 en 
 
 H 
 
 ■H 
 
 Sh 
 
 ■a 
 
 a 
 
 O 
 x) 
 
 
 LA 
 
 ft 
 
 H 
 
 • 
 
 CO 
 
 H 
 
 CM 
 
 O 
 
 
 H 
 
 x) 
 ft 
 
 O 
 
 C 
 
 
 
 o 
 
 ■H 
 
 c 
 
 •H 
 
 1 
 
 a 
 P 
 
 
 
 
 *j 
 
 c 
 
 
 
 a 
 
 
 
 
 H 
 
 o 
 
 
 
 o 
 
 
 
 
 D 
 
 ■H 
 
 
 
 ■H 
 
 
 
 
 ■H 
 
 tn 
 
 xj 
 
 
 tn 
 
 
 bO 
 
 
 g 
 
 en 
 
 p 
 
 CO 
 
 en 
 
 
 c 
 
 t3 
 
 03 
 
 ■a 
 
 •H 
 
 w 
 
 en -h 
 
 
 ■H 
 
 OJ 
 
 
 > 
 
 cd 
 
 cu a 
 
 
 H 
 
 TJ 
 
 G 
 
 CO 
 
 
 H 
 
 P> en 
 
 
 cd 
 
 C 
 
 O 
 
 c 
 
 T) 
 
 DO 
 
 cd C 
 
 H Co 
 
 
 cu 
 
 O 
 
 ■H 
 
 cd 
 
 0) 
 
 
 
 CO 
 
 rO 
 
 CO 
 
 U 
 
 •H 
 
 *h 
 
 ft u 
 
 
 
 
 O 
 
 P 
 
 H 
 
 cu 
 
 P 
 
 
 
 X! 
 
 H 
 
 
 ft P 
 
 CU 
 
 
 
 p 
 
 ft ^9. 
 
 ftH 
 
 O ^9. 
 
 
 
 ■H 
 
 a 
 
 t— 
 
 3 
 
 ■H 
 
 cd no 
 
 
 
 > 
 
 ■H 
 
 tin 
 
 tn 
 
 <P. 
 
 <H NO 
 
 
 0) 
 
 
 
 
 
 
 
 
 H 
 
 
 
 
 
 
 
 
 bO 
 
 
 
 
 
 
 
 
 3 
 
 
 
 o 
 
 
 
 o 
 O 
 
 
 
 
 H 
 
 
 
 
 
 
 
 
 <H 
 
 
 
 
 
 
 
 
 0) 
 
 
 
 
 
 
 
 
 Q 
 
 
 
 
 
 
 
 
 CU 
 
 
 
 
 
 
 
 
 n 
 
 cd 
 
 
 
 cd 
 
 
 
 
 ■H 
 
 •H 
 
 
 
 •H 
 
 
 
 
 W 
 
 tJ 
 
 
 
 TJ 
 
 
 
 
 En 
 
 = 
 
 
 
 - 
 
 
 
 
 K 
 
 CM 
 
 
 
 Zt 
 
 
 
 
 CJ> 
 
 H 
 
 P 
 C 
 
 
 
 H 
 
 -P 
 C 
 
 -8 
 
 
 
 
 s 
 
 ft 
 
 o 
 
 H 
 
 
 CM 
 
 ft 
 
 o 
 
 H 
 
 
 
 
 3 
 
 CU 
 
 (LI 
 
 o 
 
 (1) 
 
 CD 
 
 
 ^ 
 
 
 > 
 0) 
 
 g 
 
 -3- 
 
 OJ 
 
 > 
 CU 
 
 s 
 
 
 «l 
 
 
 Q 
 
 P 
 
 CJ 
 
 Q 
 
 ■p 
 
 
 o 
 o 
 
 — « a 
 
 xl 
 
 p p 
 c o 
 
 CM 
 NO 
 CM 
 
 J P 
 
 o 5 
 
 LA— - 
 
 cu 
 -a- K 
 
 o 
 
 CM 
 
 o 
 o 
 
 -3- 
 
 la 
 CM 
 
 
 > 
 
 o 
 
 o 
 
 o 
 
 o 
 
 H -a/. 
 
 1 
 
 + 1 
 
 
 
 
 
 cu 
 
 f — ■» 
 
 -p 
 
 -d 
 
 ■H 
 
 CU 
 
 ^ 
 
 §£3 
 
 
 C 
 
 
 C 
 
 ?H 
 
 o 
 
 h 
 
 O 
 
 CD 
 
 CU IsS. -H. 
 
 0) 
 
 01 tS. -H 
 
 P 
 
 u O en 
 
 p 
 
 (JVO 01 
 
 rH 
 
 cd vo en 
 
 H 
 
 cd t— en 
 
 •H 
 
 u .-a 
 
 •H 
 
 <H -H 
 
 <^< 
 
 <P 
 
 - a 
 
 
 en cu en 
 
 
 tn cu w 
 
 >> 
 
 en p cj 
 
 !>> 
 
 en -p g 
 cd cd cd 
 
 cd 
 
 cd cd cd 
 
 cd 
 
 h 
 
 H H k 
 
 U 
 
 H H !h 
 
 bo 
 
 to ft p 
 
 bO 
 
 bO ft p 
 
 o 
 o 
 
 ON 
 
 CM 
 
 co 
 
 •H 
 
 U 
 
 o 
 
 
 
 w 
 
 Ch 
 
 
 cd 
 
 M 
 
 
 
 •H 
 
 O 
 
 en 
 
 
 T3 
 
 
 >> 
 
 
 
 
 cd 
 
 
 2 
 
 0) 
 
 t3 
 
 
 CO 
 
 p 
 
 
 
 
 ■H 
 
 O 
 
 
 
 § 
 
 CO 
 
 
 
 XI 
 
 
 
 
 <v 
 
 OJ 
 
 
 
 cc 
 
 a 
 
 cd 
 
 
 
 •H 
 
 •H 
 
 
 A 
 
 p 
 
 § 
 
 On 
 
 a 
 
 
 On 
 
 cu 
 
 >» 
 
 > 
 
 CO 
 
 cu 
 
 ^ 
 
 H 
 
 -* 
 
 h 
 
 CU 
 
 >i 
 
 CJ 
 
 O 
 
 > 
 
 •H 
 
 CO 
 
 CO 
 
 ^d 
 
 H 
 
 
 *•— * 
 
 cu 
 
 0) 
 
 ■ 
 
 H 
 
 K 
 
 Q 
 
 ,a 
 
 • — ' 
 
 o 
 CM 
 
 cd 
 
 •H 
 
 H 
 
 CM 
 
 CO 
 
 -3" 
 O 
 CO 
 
 o 
 o 
 
 LfN 
 
 CM 
 
 J 
 
 a 
 
 Cm 
 
 CU 
 
 
 CU 
 
 O 
 
 M 
 
 OJ 
 
 o 
 
 H 
 
 ^-^ 
 
 ^ 
 
 ,-» 
 
 <P 
 
 T3 
 
 
 cu 
 
 LA 
 
 ffi 
 
 CO 
 H 
 
 o 
 
 o 
 
 o 
 
 o 
 
 H J- 
 
 1 
 
 + 
 
 
 a 
 
 ' — " 
 
 CU 
 
 tJ 
 
 CU 
 
 cu 
 
 !h 
 
 « 
 
 O 
 
 ss 
 
 On co 
 
 
 C 
 
 M 
 
 ■&*=. o 
 
 CU 
 
 0> VO -rH 
 
 4-> 
 
 O t- III 
 
 H 
 
 cd en 
 
 H 
 
 Ch '-H 
 
 '■M 
 
 en g 
 
 
 en 0) tn 
 
 & 
 
 cd cd cd 
 
 M 
 
 H H ^ 
 
 bO 
 
 bO ftp 
 
 o 
 o 
 
 cd 
 ■H 
 
 T3 
 
 g 
 
 CM 
 CM 
 00 
 
 EH 
 P-, 
 On 
 
 en 
 cu 
 ■H 
 
 M 
 P 
 
 en 
 
 3 
 
 a 
 
 o 
 
 p> 
 a 
 cu 
 
 H 
 W 
 
 « 3 
 
121 
 
 areas of the present current-sensitive tubes are limited color 
 gamut and simultaneous variation of brightness with color shift. 
 A combination of current-sensitive and voltage-sensitive screen 
 is possible. The manufacturers are: 
 
 a. Special Purpose Tube Company: (delivery time — 2-k months, 
 price = $5,000.00) 
 
 b. Raytheon (under development). 
 
 h. Chromatron CRT : The display screen consists of a series of 
 vertical phosphor strips; in our case, alternating strips of white 
 and red phosphors. Since most of the electrons from the gun im- 
 pinge on the phosphor screen (unlike the shadow mask tube) the 
 grid type tubes display high brightness color information. Two 
 distinct grid types are used in such color tubes: A 'switching 
 grid' and a non-switching type known as a 'post-focus grid.' 
 'Switching grid' types employ one electron gun. Depending upon 
 what relative potential is employed between the grid wires, the 
 electron beam is made to converge on either of the two phosphor 
 strips. In 'post focus grid' type color discrimination is achieved 
 by the angle of incidence of the individual electron beams (from 
 two electron guns) with the plane of the post-focus grid. The 
 different chromatron tubes in the market are described below. 
 
 a. Electro Vision Industries: 
 
 (1) 2-color, IT" dia, price = $5,000, delivery time = 3 months. 
 
 (2) 2-color, 12" dia, price = $3,500, delivery time = 3 months. 
 
 b. Thomas Electronics: 
 
 (l) 2-color, IT" dia, price = $1,000, delivery time = k months 
 or longer. 
 
122 
 
 5. Multicolor Storage Tubes : A change in electron landing-energy 
 or viewing screen potential of several kilovolts (lOKV and 12KV for 
 red and green) will alter the light output of some phosphors from 
 red to green. In phase with this voltage change and resultant color 
 shift, the backing electrode is varied over a range of one volt. 
 This cycle is repeated at 30 cycles per second. 
 
 Hughes Aircraft produces this kind of storage tube. But the price 
 is high. The receiver system must be completely built or a moni- 
 tor system must be purchased which employs this kind of tube. 
 
 6. Entertainment Type Shadow Mask CRT System : Some experiments 
 have already been done with this type of arrangement. Two film 
 tracks corresponding to monochrome pictures through the red and 
 green filters are made. One of the film tracks is used to control 
 all three color channels equally so that a simple black and white 
 picture was produced from it. The signal from the other film 
 track was then added to one of the three color channels (say, red). 
 Mr. Hughs concluded that 'every phenomenon observed with the pro- 
 jected slides was observed with the color television system, 
 although it was more difficult to get saturated colors on televi- 
 sion. ' This scheme is a simultaneous color system whereas all 
 
 the previous schemes were field- sequential systems (chromatron CRT 
 
 and current-sensitive CRT can be adapted for dot -sequential method 
 
 too) . 
 
 In Japan they have utilized the color signals from a conventional 
 color TV receiver to check Land's predictions. 
 
 One can use just the red and green signals in a color receiver 
 to see the Land phenomena. 
 
123 
 
 Since the objective of the project is to make not simply a two- 
 color TV system hut a high resolution two-color TV system (with resolution 
 as good as monochrome tubes), it did not seem appropriate to employ a 
 shadowmask tube in the final project. The chromatron tubes also suffers 
 from this deficiency (though to a lesser extent). 
 
 Conclusion : It is apparent from cost comparison that the Penetron 
 is the cheapest of all the available two-color CRT's. Two-color display 
 systems with Penetron tubes have been successfully employed for alpha- 
 neumeric displays. It was hoped that it would work for the Television 
 application. At the present time the anode voltage switching can be 
 accomplished for field- sequential systems. The only objection to the 
 penetron scheme is the need for voltage switching — and the associated 
 correction circuits. 
 
 A 'two-gun penetron' or a single-gun current-sensitive' screen 
 would remove some of the difficulties encountered with the 'single gun 
 penetron' scheme. But at the present time they are still in development 
 stage and their delivery time may be quite long. They are also expensive. 
 
12U 
 
 APPENDIX B 
 CAMERA SYSTEM SELECTION 
 
 The different forms of camera systems considered are: 
 
 1. SEC (Secondary Electron Conduction) 
 
 2. Plumbicon 
 
 3. Vidicon 
 
 h. Silicon matrix 
 
 For our system we have the requirements that the camera should 
 have very low lag so that it can be used for a field- sequential system. 
 1. SEC Camera System 
 
 The essential mechanism responsible for this is the secondary 
 electron conduction process: One hundred times multiplication of photo- 
 electrons incident on a special dielectric target. 
 SEC Tube Characteristics 
 
 Target gain: Optimum target potential 10-20 volts for a good target 
 
 gain of around 100. Target gain varies with primary 
 energy of the photoelectrons for a constant target 
 
 potential (maxm. value at ~ 8 Kev. ) . 
 
 -5 -2 
 Transfer Curves: Operation from 10 to 10 lumen of illumination. 
 
 Gamma is close to unity. 
 Lag: The lag signal is only 5% of the 1st field signal after 
 
 50 ms. and is due only to discharge lag. 
 Sensitivity: ^ 15,000 yA/lumen 
 Storage Capacity: Can be good. It is desirable to have a large storage 
 
 capacitance as long as discharge lag is avoided. 
 
125 
 
 Integration: 
 
 Resolution: 
 
 Long integration times possible for static, low-light 
 
 level scenes. 
 
 Because of the high resistivity of the target it seems 
 
 likely that its intrinsic resolution capability will be 
 
 limited by electron scattering rather than by leakage. 
 
 Resolution of 3 - h x 10 lines/inch should be possible, 
 Halation: Bloom and halation absent. 
 Additional Char: Rugged, low power requirements. 
 SEC Market Possibilities 
 a. Westinghouse 
 1. STV - 706A Single Piece SEC camera including 
 
 burn resistant tube with lens. $ 9,875.00 
 
 2-piece SEC including WL30691 (burn 
 
 resistant) less lens (25mm SEC tube) 
 
 (lag 10$), Resolution 500 lines 11,900.00 
 
 2-piece SEC including class 2 WL3065U 
 
 (UQmm tube) less lense lU,000.00 
 
 STV - 309H (l" vidicon) or STV - 3lUH (plumbicon) 
 can be used in STV - 606 with minor modification. 
 CBS Labs 
 
 Field sequential camera system using 
 
 SEC 180 fields with lens and tube $19,500 
 
 For 2-color 120 fields/sec 
 
 modification cost 5,000 
 
 *2. STV - 606 
 
 3. STV - 609 
 
 Minicam 
 
 Total 
 
 $2*1,500 
 
126 
 
 2. Plumbic on Cameras 
 
 The Plumbicon is a camera tube of the vidicon type which utilizes 
 a layer of lead oxide (PbO) for the photosensitive target instead of the 
 more usual photoconducting layer of Sb Q S . 
 
 In the Plumbicon the lead oxide layer acts as a p-i-n diode. The 
 n-region is formed at the PbO - SnO interface. The p-region is a thin 
 layer on the electron gun side of the layer. The bulk of the material 
 behaves as if it is intrinsic. The p-i-n layer is biased in reverse 
 direction. 
 Characteristics : 
 Negligibly low dark current. 
 Temperature independence of dark current. 
 Unity gamma 
 Very fast response. 
 
 Lag: In the vidicon, lag is mainly due to emptying of carriers from traps 
 which exist in large numbers in the bulk of the photoconductive 
 layer. In the Plumbicon traps only play a minor role. Instead, the 
 time response is generally limited by capacitance of the layer and 
 the effective resistance of the electron beam. 
 Spectral Sensitivity : 
 
 Because of 1.9 ev bandgap of PbO the absorption edge occurs at 
 about 6500°A and the absorption coefficient increases only slowly through 
 the visible region of the spectrum. 
 
 Phillips has recently introduced a tube by doping the normal PbO 
 layer with sulfur to form a PbO - S layer — which has a higher absorption 
 coefficient for longer wavelengths. 
 
127 
 
 Disadvantages of Plumbicon 
 
 1. The sensitivity is independent of target voltage so that control of 
 signal current at a given light level, usually an excessively bright 
 level, must he controlled by either reducing the face plate illumination 
 by, say, a remote control iris or, if the signal current is not beam- 
 current-limited, by adjusting the gain of one of the video amplifier stages. 
 
 2. The unity gamma of the layer makes it difficult to handle a wide 
 range of light levels, more than, e.g., several hundred to one. 
 
 3. The resolution of rather thin PbO layer is limited to values of 700 - 
 800 TV lines. 
 
 k. The PbO layer is not airstable and is therefore difficult to manufacture 
 5. Spectral response curve is limited in standard Plumbicons. 
 Plumbicon Camera Market Study 
 a. Phillips 
 
 LDH Multipurpose Vidicon/Plumbicon Camera 
 0150/01 1 inch, 525 line, separate mesh wired, 
 less tube 
 
 LDH Multipurpose control unit, Rack mtg. 
 0l60/02 525 line, less sync, generator 
 
 LDH 
 
 b. Phillips 
 
 LDH 
 
 0151/01 
 
 LDH 
 0160/02 
 
 LDH 
 1*300/10 
 
 Multipurpose sync, generator module, 
 525 line, 2:1 interlace 
 
 Sub -total 
 Tube (CCTV 111, or l6Q) 
 
 Total 
 
 Plumbicon Camera, 1 l/V 525 lines, 
 less tube std. mesh wired 
 
 Camera Control 
 Sync generator 
 
 $ 895 
 1685 
 
 220 
 
 $2800 
 850 
 
 $3650 
 
 $1090 
 1685 
 
 220 
 
 Total 
 
 $2995 
 
128 
 
 c. COHU 
 
 Tube X Q 1021, 55875, 
 
 X Q 102U s CCTV101 (30mm) 
 
 3207 Plumbicon Camera with CCTV 101 
 
 2U8U-750 dual Camera control 
 
 $ 850 
 
 d. Westinghouse STV-711 Single piece Plumbicon camera 
 
 including Tube and 50mm F/1.8 lens 
 
 e. Colorado Video 501 /A with CCTV 111 
 3. Vidicon 
 
 $1+995 
 $2500 
 
 Most of the vidicon cameras can accept l" Plumbicon tube with 
 additional $1000 for the cost of the tube. All vidicon tubes have 25- 
 h0% lag and hence, cannot be advantageously .used for our field-sequential 
 system. 
 
 The different products considered in this category were from 
 manufacturers like MTI, Cohu, Phillips, Westinghouse, and G. E. 
 k. Silicon Matrix 
 
 Silicon Matrix tubes have an array of Silicon diodes that act as 
 the sensing element. These tubes have many advantages: 
 Over Sb^S vidicon tubes: High sensitivity, broad spectral response, 
 
 low lag, good resolution. 
 Over Plumbicon: Sensitivity, spectral response and resolution. 
 
 In lag characteristics they are slightly inferior to Plumbicons. 
 
 The market search revealed the following candidates in this 
 category. 
 
 a. Amperex 
 
 RCA 
 
 Conclusion 
 
 S10XQA 
 
 C23136 
 
 Development type 
 
 Silicon Vidicon Camera tube 10% lag 
 after 50 m.s., electrically interchange- 
 able with any separate mesh l" vidicon 
 
 Silicon diode array, 8% lag 200nA signal 
 
 At this time it looks as though a camera system using a Plumbicon 
 tube is about the best that we can have considering the cost and the present 
 
129 
 
 state of technology. SEC's don't have better lag characteristics than a 
 Plumbicon — they are rugged and have "better versatility in sensitivity 
 range. We rule out SEC system because of price considerations — though 
 CBS lab has developed a field sequential color system employing these 
 tubes for the Apollo program. 
 
 A l" tube should be used since most of the developments and re- 
 search is concentrated on l" tube. If we buy a l" tube we can use it 
 for vidicon, plumbicon or silicon matrix tube (with modifications). 
 Silicon matrix tubes are still in development stages. Though they have 
 better resolution and spectral sensitivity than the Plumbicon, they don't 
 have better lag characteristics. 
 
 With the Plumbicon the resolution is limited to about TOO lines 
 which is adequate for our use. The red sensitivity of standard Plumbicon 
 is limited to 6500°A. So we will have less red output. 
 
 The Phillips camera system (l" type) seems to be the most versatile, 
 The camera system has dual outputs — it means we can also use this camera 
 control for a dual camera system. We can use vidicon or Plumbicon for a 
 dual camera system using dichroic mirrors. So it will be convenient to 
 buy this vidicon/Plumbicon camera. 
 
 Addendum : Because of financial considerations, the GE fully contained 
 vidicon camera system was purchased and the vidicon was replaced with an 
 Amperex Plumbicon tube with some modifications. 
 
130 
 
 APPENDIX C 
 
 LAND SLIDES MD PHOTOGRAPHS 
 
131 
 
 C.l Land Black and White Transparency (Longwave Record) 
 
132 
 
 C.2 Land Black and White Transparency (Shortwave Record) 
 
133 
 
 C.3 Land Color Projection of Longwave and Shortwave Records 
 
13U 
 
 C.U Land Color PENTECOST Display 
 
135 
 
 C.5 Land Color PENTECOST Display (Color Reversal) 
 
LIST OF REFERENCES 
 
 1. Land. Edwin H. , "Color Vision and the Natural Image. Part I," Proc. 
 
 National Academy of Science, Vol. U5, p. 115 (1959). 
 
 2. Land, Edwin H. , "Color Vision and the Natural Image. Part II," Proc. 
 
 National Academy of Science, Vol. 1+5, p. 636 (1959). 
 
 3. Land, Edwin H. , "Experiments in Color Vision," Scientific American, 
 
 Vol. 200, p. 81+ (1959). 
 
 k. Land, Edwin H. , "The Retinex," Scientific American, Vol. 52, p. 2l+7 
 (196U). 
 
 5. Polyak, S. , "The Vertebrate Visual System," University of Chicago 
 
 Press (1957). 
 
 6. McCulloch, W. S. , "Why the Mind is in Head," The Hixon Symposium, 
 
 John Wiley & Sons, Inc. (1951). 
 
 7. de Barenne, H. G. Dusser, H. W. Carol, and W. S. McCulloch, "Physio- 
 
 logical Neuronography of the Cortico-striatal Connections," 
 Association for Research in Nervous and Mental Disease, Vol. 21, 
 p. 2I+6 (19U2). 
 
 8. Kahrisky, Mathew, "A Proposed Model for Visual Information Processing 
 
 in the Human Brain," University of Illinois Press (1966). 
 
 9. Young, Thomas, "Lectures in Natural Philosophy," London, Vol. 2, 
 
 p. 315 (1807). 
 
 10. Maxwell, J. C. , "Scientific Papers," Proc. Royal Institute, Great 
 
 Britain, Vol. 6, p. 260 (l87l). 
 
 11. von Helmhotz, H. , Ann. Physik. , Vol. 87, p. 1+5 (1852). Translation, 
 
 Phil. Mag. Series k, Vol. k, p. 519 (1852). 
 
 12. Hering, E. , "Outlines of a Theory of the Light Sense (English trans- 
 
 lation hy L. M. Hurvich and D. Jameson), Harvard University 
 Press, Cambridge, Mass. (196U). 
 
 13. Wallach, H. , "The Perception of Neutral Colors," Scientific American, 
 
 Vol. 208, pp. 107-116, (Jan. 1963). 
 
 Ik. Biernson, George, "A Feedback-Control Model of Human Vision," Proc. 
 of IEEE, Vol. 5b, No. 6, p. 858 (June 1966). 
 
 15. Wald, G. , P. K. Brown, and I. R. Gibbons, "The Problem of Visual 
 
 Excitation," J. Opt. Soc. of America, Vol. 53, pp. 20-35 (Jan. 1963). 
 
 16. Svaetichin, G. , "Origin of the R-Potential in the Mammalian Retina," 
 
 in 'The Visual System: Neurophysiology and Psychophysics. ' 
 
 R. Jung and H. Kornhuber , Eds. Berlin: Springer-Verlag, pp. 61-61+ 
 
 (1961). 
 
137 
 
 IT. Laufer, M. , G. Svaetichin, G. Mitarai , F. Fatehchand, E. Vallecalle, 
 and J. Villegas, "The Effects of Temperature, Carbon-Dioxide, and 
 Ammonia on the Neuron-Glia Unit," ibid. pp. l*57-*+63. 
 
 18. Campbell, F. W. and W. A. H. Rushton, Journal of Physiology, Vol. 130, 
 
 p. 131 (1955). 
 
 19. Brown, P. K. and G. Wald, Science, Vol. ikk , p. 1+5 (196U). 
 
 20. Rushton, W. A. H. , "Visual Pigments in Man," Scientific American, 
 
 pp. 120-132 (Nov. 1967). 
 
 21. Wald, George, "The Receptors of Human Color Vision," Science, Vol. 1^5, 
 
 pp. 1007-1016. 
 
 22. Judd, Deane B. , "Appraisal of Land's Work on Two-Primary Color Pro- 
 
 jections," J. of Opt. Soc. Am., Vol. 50, 3, 25 1 * (i960). 
 
 23. Woolfson, M. M. , "Some New Aspects of Color Perception," IBM Journal, 
 
 p. 319 (Oct. 1959). 
 
 2k. Wilson, M. H. and R. W. Brocklebank, "Color and Perception: The Work 
 of Edwin Land in the Light of Current Concepts." 
 
 25- Judd, Deane B. , "Hue, Saturation and Lightness of Surface Colors 
 with Chromatic Illumination," J. Opt. Soc. Am., Vol. 30, p. 2 
 (19^0). 
 
 26. Helson, H. , "Fundamental Problems in Color Vision," J. Experimental 
 
 Psychology, Vol. 23, p. ^39 (1938). 
 
 27. Pearson, D. E. , C. B. Rubinstein, "Computations of Color Fidelity in 
 
 Two-Primary Television Systems," Proc. of the 2nd Annual Princeton 
 Conference on Information Sciences and Systems, Princeton, p. 109 
 (1968). 
 
 28. Pearson, D. E. . C. B. Rubinstein, G. J. Spirack, "Comparison of Per- 
 
 ceived Color in Two-Primary Computer-Generated Artificial Images 
 with Predictions Based on the Helson-Judd Formulation," J. Opt. 
 Soc. Am., Vol. 50, 5, p. 6kk (1969). 
 
 29. Pearson, D. E. , C. B. Rubinstein, "Range of Perceived Hues in Two- 
 
 Primary Projections," J. Opt. Soc. Am., Vol. 60, 10, p. 1398 
 (1970). 
 
 30. Land, E. H. and J. J. McCann, "Lightness and Retinex Theory," J. of 
 
 Opt. Soc. of Am., Vol. 6l, 1, pp. 1-11 (1971). 
 
 31. McCann, J. J., "Rod-Cone Interactions: Different Color Sensations from 
 
 Identical Stimuli," Science, Vol. 176, pp. 1255-1257, (June l6 , 
 1972). 
 
 32. Shkol'nik-Yarros Ekaterina, G. , "Neurons and Internauronal Connections 
 
 of the Central Visual System," Plenum Press (1971). 
 
138 
 
 33- DeValois, R. L. , "Mechanisms of Color Discrimination," Proc. Inter- 
 national Symposium, pp. 111-llH (1958). 
 
 3U. DeValois, R. L. , R. J. Smith, S. T. Kitai, and A. J. Karoly, Science, 
 Vol. 127, p. 238 (1958). 
 
 35. DeValois, R. L. and A. E. Jones, In: "The Visual System. Neuro- 
 
 physiology and Psychophysics ," Symposium, Berlin, pp. 178-191. 
 
 36. Simmon, A., "EIDOLYZER: A Hardware Realization of Context -Guided 
 
 Picture Interpretation." UIUC Department of Computer Science 
 Report No. UU8, (June 1971). 
 
 37' Yachida, Masuhiko and Saburo Tsuji, "Application of Color Informa- 
 tion to Visual Perception," Pattern Recognition, Vol. 3, pp. 307- 
 323 (1971). 
 
 38. Preparata, F. P. and S. R. Ray, "An Approach to Artificial Non- 
 Symbolic Cognition," Co-ordinated Science Laboratory Report 
 R-U78, U. of 111. , (1970). 
 
 39' Sutro, L. L. and W. L. Kilmer, "Assembly of Computers to Command and 
 Control a Robot," Report 582, Instrumentation Laboratory, M.I.T., 
 Cambridge (1969). 
 
 kO. Wyszecki, Gunter and W. S. Stiles, "Color Science," John Wiley & Sons, 
 Inc., New York, p. 269 (1967). 
 
 kl. Rutland, David, "A Digital Interactive Color Television Display," 
 Information Display, Vol. 7, 8, p. 20, (Oct. 1970). 
 
 k2. Frei, Werner, "Quantization of Pictorial Color Information; Nonlinear 
 Transforms," Proc. of the IEEE, p. U65 , (April 1973). 
 
 1*3. Preparata, F. P. "Pictorial Antificial Intelligence," Class Notes Spring 
 1972, U. of 111., Urbana. 
 
 kk. Narasimhan, R. , "On the Description, Generation, and Recognition of 
 
 Classes of Pictures," in Automatic Interpretation and Classifica- 
 tion of Images, Academic Press, New York (1969). 
 
 U5. Clowes, M. B. , "Trans fermational Grammars and the Organization of 
 Pictures," in Automatic Interpretation and Classification of 
 Images, Academic Press, New York (1969). 
 
 U6. Guzman, A., "Computer Recognition of Three-Dimensional Objects in a 
 Visual Scene," MAC-TR-59 (Thesis), Project MAC, M.I.T. (1968). 
 
 h"J. McCulloch, C. , "Color Adaptation of Edge-Detectors in the Human 
 Visual System," Science, Vol. 1^9, pp. 1115-1116 (1965). 
 
 U8. Stromeyer III, C. F. , "McCulloch Effect Analogs of Two-Color Projec- 
 tions," Vision Research, Vol. 11, pp. 969-967 (1971). 
 
139 
 
 1+9- Festinger, L. , M. R. Allyn, and C. W. White, "The Perception of Color 
 with Achromatic Stimulation" Vision Research, Vol. 11, pp. 591-612 
 (1971). 
 
 50. Roller, L. R. and H. D. Coghill, "Electron Excitation of Bilayer Screens," 
 
 Journal of Applied Physics, Vol. 29, 7, p. 106U (July 1958). 
 
 51. Davis, J. A., "Recent Advances in Cathode Ray Tube Display Devices," 
 
 Recent Advances in Display Media, NASA SP-159. 
 
 52. Passavant, Francis C. , "Multicolor Cathode Ray Tube Displays," Computer 
 
 Design, p. 53 (January 1970). 
 
 53. Robiner, R. C. and M. Fogelson, "Two-Color Display Using Dual-Neck 
 
 Flat-Face Cathode Ray Tube," Zenith Radio Corporation, Rauland 
 Division 
 
 5I+. Adam, R. C. , "Two-Color, Alphaneumeric Display," Information Display, 
 p. 23 (July/Aug. 1970). 
 
 55- Poppelbaum, W. J., "Computer Hardware Theory," McMillan Press, London 
 (1972). 
 
 56. Hoeschele, David F. , Jr., "Analog-to-Digital/Digital-to-Analog Conver- 
 
 sion Techniques," John Wiley & Sons, pp. 102-103 (1968). 
 
 57. Golomb, S. W. , "Shift Register Sequences," Holden-Day Inc., San Frans- 
 
 cisco (1967). 
 
 58. Davies, W. D. T. , "Generation and Properties of Maximum Length Sequences, 
 
 Parts 1, 2, 3," Control, (June, July, Aug. 1966). 
 
 59- Peterson, W. W. , "Error Correcting Codes," The MIT Press and John Wiley 
 and Sons, Inc. (1961). 
 
 60. Levitt, R. S. , "Performance and Capabilities of New Plumbicon TV 
 
 Camera Pickup Tubes," Journal of the SMPTE, Vol. 79, pp. 115-120, 
 (Feb. 1970). 
 
 61. van den Broek, J., "The Electrical Behaviour of Vapour-Deposited Lead- 
 
 Monoxide Layers," Phillips Research Reports, Vol. 22, pp. 367-37*+ 
 (1967). 
 
 62. Stupp, E. H. "Physical Properties of the Plumbicon," Photo Electronic 
 
 Imaging Devices, Vol. II Phenom Press, pp. 275-283 (1971). 
 
 63. Bardeen, John, "Surface States and Rectification at Metal-Semiconductor 
 
 Contact," Physical Review, Vol, 71, 10, pp. 717-727 (19^7). 
 
 61*. Data Sheet Tube CCTVIIIR, Amperex Electronic Corporation (1968). 
 
 I 65. Wagner, T. M.,"The One-Tube Color-Camera for Live or Film Use," 
 RCA Corporation, (1970). 
 
 j 66. Data Sheet and Supplement on 'Optichron,' Electrochome Corporation, 
 New York, (1972). 
 
lUo 
 
 67. Land, C. E. , P. D. Thacher, "Ferroelectric Ceramic Electro-Optic 
 
 Materials and Devices," Proc. IEEE, Vol. 57, 5,'P- 571 (1969). 
 
 68. Land, C. E. , P. D. Thacher, "Electro-Optic Ceramics — New Materials 
 
 for Information Storage and Display," Western Electric Engineer 
 (1970). 
 
 69. Meitzler, A. H. , J. R. Maldonado, and D. B. Fraser, "image Storage 
 
 and Display Devices Using Fine-Grain, Ferroelectric Ceramics," 
 B.S.T.J., pp. 953-967 (July-Aug. 1970). 
 
 70. Aizu, K. , "Possible Species of Ferromagnetic, Ferroelectric, and 
 
 Ferroelastic Crystals," Physical Review, B2, pp. 75^-772 (1970). 
 
 71. Kumada, Akio, "Optical Properties of Gradolinium Molybdate and Their 
 
 Device Applications," Ferroelectric s , Vol. 3, pp. 115-123 (1972). 
 
 72. Smith, A. W. and G. Burns, "Optical Properties and Switching in 
 
 Gd 2 (M 0^) ," Physical Letters, Vol. 28A, p. 7 (1969). 
 
 73. Cummins, S. E. , "Electrical, Optical, and Mechanical Behaviour of 
 
 Ferroelectric Gd (M 0, ) ," Ferroelectrics , Vol. 1, pp. 11-17 
 (1970). 2 o 4 3 
 
 7^. Kumada, Akio, "Ferroelectric and Ferroelastic Gd (M 0, ) . II. Func- 
 tion and Applications of Gd„(M 0> ) ," English Translation, Central 
 Research Laboratory, Hitachi Ltd. 
 
 75* Kumada, Akio, "instruction of MOG Color Modulator," Central Research 
 Laboratory, Hitachi Ltd. 
 
 76. Hughes, W. L. , "Some Color Slide and Color Television Experiments 
 Using the Land Technique," Trans. IRE PGBTR, (July i960). 
 
lUi 
 
 VITA 
 
 Godavarish Panigrahi was born in Orissa, India on March 30, 19U7. 
 He completed the Pre-University Science and the Pre-Engineering degrees 
 respectively in June 1963 and June 196U from Utkal University. He joined 
 Indian Institute of Technology, Kharagpur in July I96U, and received the 
 B. Tech(Hons) degree in Electronics and Electrical Communication Engineer- 
 ing in June 1968. His Bachelor's project and thesis work was on a Unidigit 
 PCM (Pulse Code Modulation) Communication System. 
 
 In September 1968, he joined the University of Illinois, Urbana as 
 a University of Illinois Fellow in Electrical Engineering and continued in 
 that position until September 1970. He did research on MOS (Metal-Oxide- 
 Semiconductor) Devices in the Solid State Electronics Laboratory and re- 
 ceived his M.S. in Electrical Engineering in June 1970. In September 1970, 
 he joined the Computer Hardware and Systems Research Group presently called 
 the Information Engineering Laboratory under Professor W. J. Poppelbaum. 
 He was a Teaching Assistant in the Department of Electrical Engineering 
 from September 1970 to June 1971- He worked the summer of 1971 as a 
 Research Assistant in the Computing Centre of the Department of Psychology. 
 Since September 1971 he has held a Research Assistantship in the Department 
 of Computer Science and has continued to work under Professor W. J. Poppel- 
 baum towards a Ph.D. degree. 
 
 He is the author of a paper on Metal-Oxide-Semiconductor Devices 
 in the Electronics Letters, IEE London and the co-author of a paper on 
 the Numerical Solution of Poisson's Equation in Semiconductor Devices in 
 the Journal of Applied Physics. He is a member of the Computers Group 
 and the Electronic Devices Group of IEEE. 
 
jrm AEC-427 
 
 (6/68) 
 AECM 3201 
 
 U.S. ATOMIC ENERGY COMMISSION 
 
 UNIVERSITY-TYPE CONTRACTOR'S RECOMMENDATION FOR 
 
 DISPOSITION OF SCIENTIFIC AND TECHNICAL DOCUMENT 
 
 ( See Instructions on Reverse Side ) 
 
 AEC REPORT NO. 
 
 COO-li+69-0235 
 
 2. TITLE 
 
 PENETRON LAND COLOR DISPLAY SYSTEM (PENTECOST) 
 AND SOME OBSERVATIONS CONCERNING COLOR PERCEPTION 
 
 TYPE OF DOCUMENT (Check one): 
 
 Q3 a. Scientific and technical report 
 
 _| b. Conference paper not to be published in a journal: 
 
 Title of conference 
 
 Date of conference 
 
 Exact location of conference _ 
 
 Sponsoring organization 
 
 □ c. Other (Specify) 
 
 RECOMMENDED ANNOUNCEMENT AND DISTRIBUTION (Check one): 
 
 (3 a. AEC's normal announcement and distribution procedures may be followed. 
 
 ~Jl b. Make available only within AEC and to AEC contractors and other U.S. Government agencies and their contractors. 
 ] c. Make no announcement or distribution 
 
 REASON FOR RECOMMENDED RESTRICTIONS: 
 
 i SUBMITTED BY: NAME AND POSITION (Please print or type) 
 
 W. J. Poppelbaum 
 
 Professor and Principal Investigator 
 
 Organization 
 
 Department of Computer Science 
 University of Illinois 
 Urbana, Illinois 6l801 
 
 Signature 
 
 Date 
 
 October 1973 
 
 FOR AEC USE ONLY 
 
 'AEC CONTRACT ADMINISTRATOR'S COMMENTS. IF ANY, ON ABOVE ANNOUNCEMENT AND DISTRIBUTION 
 RECOMMENDATION: 
 
 ATENT CLEARANCE: 
 
 U a. AEC patent clearance has been granted by responsible AEC patent group. 
 J b. Report has been sent to responsible AEC patent group for clearance. 
 I U c. Patent clearance not required. 
 
DEC 14 1973 
 
iiLIOGRAPHIC DATA 
 HET 
 
 1. Report No. 
 
 UIUCDCS-R-73~58^ 
 
 H 
 
 ,;'itle and Subtitle 
 
 PENETRON LAND COLOR DISPLAY SYSTEM (PENTECOST) 
 AND SOME OBSERVATIONS CONCERNING COLOR PERCEPTION 
 
 . uthor(s) 
 
 G. Panigrahi 
 
 . erforming Organization Name and Address 
 
 Department of Computer Science 
 University of Illinois 
 Urbana, Illinois 6l801 
 
 2,;ponsoring Organization Name and Address 
 
 US AEC Chicago Operations Office 
 9800 South Cass Avenue 
 Argonne, Illinois 
 
 5. upplementary Notes 
 
 3. Recipient's Accession No. 
 
 5. Report Date 
 
 October I973 
 
 6. 
 
 8. Performing Organization Rept. 
 
 No. K 
 
 10. Project/Task/Work Unit No. 
 
 11. Contract /Grant No. 
 
 US AEC AT (11-1) 1469 
 
 13. Type of Report & Period 
 Covered 
 
 Thesis research 
 
 14. 
 
 i. bstracts 
 
 Dr E H S.!? exa f"es the human color vision mechanism in lisht of 
 experiments. telCT1510n ^P 1 ^ ^™ based on the two-color projection 
 
 y Words and Document Analysis. 17a. Descriptors 
 
 Land effect 
 
 tifiers/Open-Ended Terms 
 
 j'SATl Field/Group 
 A vl lability Statement 
 
 unlimited distribution 
 
 I ls- 3B ( 10-70) 
 
 19. Security Class (This 
 Report) 
 
 UNCLASSIFIED 
 
 20. Security Class (This 
 
 Page 
 UNC I.ASSIF 1 F D 
 
 21. No. of Pages 
 
 152 
 
 22. Price 
 
 USCOMM-DC 4 0329-P7I 
 
JV3L 
 
 z 5 \*w