Appendix

Development of High-Definition Television in

Japan, Europe, and the United States

Plans for the practical implementation of improved television systems in the three
principal TV regions of the world have been influenced at least as much by con-
siderations of economic gain and national or company pride as by the scientific
and technological principles discussed in Chap. 8. For this reason, it is necessary
to deal here with many issues that are matters of opinion rather than fact. The
opinions expressed here are those of the author alone. Another view is found in a
report to the French parliament [A.l]. The final outcome of what has turned out
to be a vigorous contest is far from clear at the time of writing (March 1990).
No doubt, many of the contentious arguments now being conducted will even-
tually disappear into well earned oblivion. Nevertheless, enough new knowledge
of the purely technical aspects of providing much better pictures and sound for
the world's living rooms and classrooms has emerged to make it worthwhile to
give an up-to-date account.

At. Introduction

France had an 819-line monochrome system in commercial operation from shortly

after WWII until 1985, but it shifted its principal transmissions to 625-line
SECAM in 1965 when all of Europe adopted color. Thus, the recent history of
HDTV started in Japan in 1970, when the Japan Broadcasting Company (NHK)
set about to design the next generation of television systems. The main systems
work was done by NHK, but the development of particular equipment, such as
cameras, recorders, and displays, was delegated to the large Japanese electronics
companies. The first demonstrations were held in the late seventies. Japanese at-
tempts to have their system accepted as an international standard for production
and international exchange of programs stimulated the Europeans to pool their
efforts in Project Eureka in 1985. A preliminary demonstration with most of the
hardware, but not the final system design, was held in 1988 in Brighton, England.
A more advanced version was demonstrated in 1989 in Berlin. In the US, there
has been no agreement among the interested parties. The system to be used will
be selected in a contest run by the television industry under supervision of the
Federal Communications Commission (FCC).

262
AI.I Compatibility

No aspect of new TV systems excites such intense discussion as the question as

to whether the new signals should be viewable on old receivers. For the sake of
the then-existing 10 million receivers (there are now about 180 million NTSC
receivers and 60 million NTSC VCRs in the US, and more than 500 million TV
sets in the world at present) the US added color in 1953 in a completely com-
patible manner. Old receivers could use the new signals and new color receivers
could use the old signals. This was considered a technological triumph at the
time, although it caused problems that are only now being properly addressed.
On the other hand, when Europe adopted 625-line color in 1965, although it
was compatible in some countries, the new signals were not compatible with old
receivers in France and Great Britain. A decision was made to serve existing sets
by separate transmissions for a period of 20 years. This procedure, called simul-
casting, has generally been viewed with disfavor by commercial broadcasters
because of the extra expense.
Although, to many in the industry in the US and Europe in the 1980s, it
seemed self-evident that HDTV had to be compatible to be successful, com-
patibility removes one of the principal incentives to buy new receivers - the
prospect of seeing attractive programs that are not otherwise available. Viewed
in this light, the near-failure of NTSC compatible color (it took 10 years to reach
the 1% penetration point) may perhaps be attributed to the fact that viewers were
able to see all the new programs on their existing receivers, albeit in black and
white.
Aside from providing more incentive to buy new receivers, a system free of
the constraint of compatibility can surely be a better system, particularly with
respect to the efficiency of spectrum utilization, as has been shown in Chap. 8.
Simulcasting also deals with the "chicken-and-egg" problem, in which viewers
fail to buy receivers because there are few new programs and producers are
reluctant to make programs since there are so few receivers. When one adds to
that the difficulty of conforming different aspect ratios in compatible systems as
well as the fact that simulcasting need not be as expensive as once thought, the
argument for compatibility seems not so overwhelming.

AI.2 Spectrum Considerations

With about twice the horizontal and vertical resolution as NTSC and with a
wider aspect ratio, HDTV transmitted with the same efficiency as NTSC or PAL
would require about five times the bandwidth. This seems impractical in all
existing transmission media, and would probably not be feasible even in fiber-
optic cable. An important goal, therefore, is to compress the required bandwidth
to not much more than is used at present. While bandwidth-reduction schemes
have never been popular with broadcasters since they did not believe it was
possible to maintain quality while reducing bandwidth, the exigencies of the
HDTV situation and the demonstration of some effective compression methods
have changed attitudes considerably.
263
 Aside from reducing the bandwidth required for single HDTV signals, the
other important spectrum consideration relates to the number of stations that
can be accommodated within a given overall allotment. Here the question of
interference, both within a single channel and between adjacent channels, is of
paramount importance. It has now been accepted that much better interference
performance is possible with systems that abandon current transmission formats.
The pressure for using some spectrum now allocated to television for other
services, such as mobile radio and cellular telephones, is very strong, particularly
in the United States. To accommodate these new services without reducing the
current level of TV service is probably only possible with entirely new systems.

Al.3 Transmission Channels

As shown in Fig. A.1, the overall television system, from the scene in front of the
camera to the image on the home receiver, is very complicated. Any new system
design must take account of this. It must be adaptable to every transmission
medium and, if different formats are required in the different media, provide for
adequate transcoding among them.
Even though more than half of American TV homes subscribe to cable, the
over-the-air (terrestrial) channel is likely to remain the primary distribution means
in the US for some time. For one thing, it is by far the cheapest way to distribute
broadband signals to the home. For another, it is free in the US and cheap to
viewers in most countries. All other media involve higher cost to the consumer, so
there is an understandable reluctance to eliminate terrestrial transmission entirely.
As discussed in Chap. 8, delivering really high-quality signals in this medium is
not easy.
Cable, in principal, should give better signal qUality. It should also have fewer
spectrum constraints. However, since it is very important to be able to distribute,
by cable, the same programs that are transmitted terrestrially, in practice the
two media have used exactly the same transmission format. Such proposals as
have been made to make cable the primary distributor for HDTV are based on
commercial rather than technological considerations.
Direct-broadcast satellites (DBS) are already in use. As compared to terres-
trial transmission, the signal quality is normally much better, with substantial
freedom from ghosts and interference, and with an adequate SNR. In addition,
in Europe and Japan, which have far fewer terrestrial channels available than the
US, DBS is a way of providing additional channels at acceptable distribution
costs. The consumer must buy a DBS antenna and receiver, which some predict
may cost as little as $300 at today's prices. Since it has been repeatedly demon-

Fig. A.I. Universal Production and Distribution [Link] the future, TV may be distributed to the
home by four different kinds of channels in addition to tape or disk recorders. Production of programs
is a complicated business involving multiple sources and a good deal of communication. The finished
programs must then be transmitted to the TV stations, cable head-ends, satellite up-links, etc., for
international exchange and/or retransmission to the viewers. If, as anticipated, different transmission
formats are used for each section of the pathway, transcoding is required at many points

264
 AEF

Telecommunications
Network

DEF

I I I • Cable Network

DPF

Postprodllction
System Fiber Network
Notes:
DTF: Digital Transmission Format
ATF: Analog Transmission Format
DEF: Digital Exchange Format
AEF: Analog Exchange Format
DPF: Digital Production Format
APF: Analog Production Format
DVR: Professional Digital Video Recorder
AVR: Professional Analog Video Recorder
VCR: Consumer Video Cassette Recorder
Fig. A.t. For caption see opposite page
N
Ol
tTl
strated that consumers will pay for program choice, such prices would not be an
insurmountable barrier in countries where there is not a well established cable
network already serving this purpose.
Fiber-optic cable is already in use for telephone trunk lines, and there is a
move to put it into every home in order to serve all telecommunication purposes.
However, at present, fiber installations are more expensive than cable. In order for
fiber to the home to become economically viable, viewers must be willing to pay
substantially more for a wider variety of services than they now get from cable.
There is no market evidence that they want these additional services, so fiber
remains an intriguing possibility rather than a certainty. If fiber does materialize,
it will have a large effect, since it can provide a high-capacity high-quality digital
channel. This is especially true if the telephone companies, which are committed
to all-digital transmission, become the main providers of this service.
One medium that will certainly be of great importance in HDTV, as it is
in today's television distribution system, is the video recorder, both erasable,
as the VCR, and not, as the optical disk. It was once thought (feared, actually)
by American broadcasters that enough viewers might watch recorded HDTV to
cause them to lose noticeable audience share. However, since sports are sure
to be among the most popular HDTV subjects, there certainly will be a role
for live high-definition broadcasting. To the extent that magnetic video recorders
remain analog, they can use technology quite similar to that of DBS. If sufficiently
inexpensive digital magnetic recorders are developed for home use, digital coding
schemes such as those suitable for optical disks or optical cable can be used.

AI.4 Transition Scenarios

A final element that influences HDTV strategy is the means by which a new
TV system can be introduced. This, in turn, depends a great deal on whether a
compatible or noncompatible (to some, evolutionary or revolutionary) approach
is taken, and on whether today's TV is to be abandoned or is to exist side-by-side
with HDTV.
For compatible systems, broadcasters must take the first step of transmitting
HDTV programs while there are few HDTV receivers. Much of the cost would be
borne by advertising directed primarily at the installed base. The incentive to bear
the increased cost of HDTV programming would presumably be competition from
HDTV delivered in alternative media. The presumption is that viewers would
start to buy HDTV receivers to see the same programs in higher quality. As this
happened, production values could gradually be changed to take advantage of
the higher definition, with the result that the visual impact on the old receivers
would decline. Presumably, this would provide added incentive to buy new sets.
Eventually, all broadcasting might be in HDTV, but viewers would never be
forced to buy new sets, a situation identical to that of adding color to NTSC. In
the latter case, color receivers have become so cheap, and color is so obvious an
improvement over monochrome, that the market for black-and-white receivers
is quite small, and virtually all broadcasting is in color. Since high definition

266
gives a much smaller increment in visual impact than color, we would expect
the transition to be much longer, and the market for old-style sets to remain
substantially larger. These trends would be intensified if the HDTV sets were to
remain considerably more expensive than today's and if the picture quality on
HDTV receivers continued to be generally impaired by transmission problems.
In the compatible scenario, it has also been suggested that the ultimate picture
quality might be reached in a. series of steps to be taken both by broadcasters
and viewers. The former would start with low-cost steps such as first increasing
the aspect ratio without changing the resolution, thus not requiring much new
studio equipment. Additional channel capacity and new cameras might be added
at a later date to provide higher resolution. Viewers could start by buying new
receivers to match, which presumably would be less expensive than those for
full HDTV. Of course, this requires that each step be compatible with the pre-
vious step, particularly in the case of receivers. We would also expect slower
penetration with this scenario, since the increment of picture quality at each step
would be smaller.
In the case of noncompatible systems that are to supplement the existing
service, rather than to replace it, the scenario is simpler, being comparable to the
coexistence of AM and FM. Broadcasters will start up with almost no audience.
The cost of production and broadcasting in HDTV will be borne in the hope that
an audience will develop because of the possibility of seeing unique programs
that are also of higher quality. Simulcasting will be avoided in most cases so as to
provide the greatest incentive to buy new receivers. This scenario clearly works
much better where an insufficient number of channels is currently available, as
is the case in Europe and Japan. The additional programs available, whether in
HDTV or not, would most likely be the main reason for buying the receivers.
Some have suggested that this scenario lends itself to the introduction of HDTV
via cable or other pay services, rather than as a free service.
For noncompatible systems intended ultimately to replace existing services,
simulcasting will probably be used as a matter of equity to viewers who might
otherwise be deprived of service unless they were willing to buy a new set im-
mediately as well as to provide an audience for the programmers and advertisers
while the number of new sets remained small. The incentive to buy new sets
is initially the better quality, and, eventually, the demise of the old broadcasts.
If announced far enough in advance (10 years or so), this scenario meets the
test of fairness. It also leaves open the option of retaining the old broadcasts for
a longer period, depending on the acceptance of high definition in the market
place.
It is difficult to know the extent to which authorities in the various countries
thought through the implications of the fundamental choices that they have made
in their quite distinct approaches. The desire to capture market share in receivers
and other HDTV equipment, as well as the equal desire not to be captured, have
no doubt figured prominently in the various strategies.

267
[Link]

To many observers, HDTV is simply the new television system developed by

the Japanese. Since they started so early and have spent so much more money
than anyone else, those outside the field sometimes feel that there is no point in
anyone trying to catch up, in particular the Americans. As we shall see, starting
early has its disadvantages as well as its advantages.

A2.1 History of the Japanese Developments

With typical foresight and willingness to invest money in order to secure markets
many years in the future, the Japan Broadcasting Corporation (NHK) began de-
velopment of HDTV in 1970. The program was orchestrated by NHK, which did
the system development and preliminary development of equipment, while the
commercial equipments - cameras, displays, recorders, and ancillary products
- were developed by the major domestic electronics corporations, including Hi-
tachi, Matsushita, Sony, NEC, Toshiba, Ikegami, etc. From the first, the plan was
to implement HDTV in Japan as an entirely new service, delivered to viewers by
DBS, and intended to supplement, rather than replace, the existing over-the-air
(terrestrial) system, which would continue to employ NTSC. 1
Scanning standards of 1125 lines, 60 fields/s, 2:1 interlace, and 5:3 aspect
ratio (later changed to 16:9) were chosen with the intention of making the picture
quality comparable to that of 35-mm motion pictures? These numbers may be
contrasted with NTSC's 525 lines, 59.94 fps, 2:1 interlace, and 4:3 aspect ratio,
and the 625/50/2: 114:3 PAL and SECAM systems used in the 50-Hz countries.
Demonstrations outside of Japan with this system, now called the "studio
system," began in 1981. Experiments were carried out to show terrestrial trans-
mission at 38 GHz and analog transmission in optical fiber. Successful satellite
transmissions were carried out in 1978. Because the system had a baseband-
width about five times that of NTSC, standard satellite transponder channels,
which have 24 or 27 MHz bandwidth, 3 were not adequate, so that a special
transponder of more than 100 MHz bandwidth was required.
Evidently, it was deemed impractical to use such wide transponder channels.
Therefore, a transmission system, MUSE [A.2], was developed so that a compres-
sed version of the signal could be transmitted in a normal satellite channel. This
1 The color TV system originally proposed in 1953 by the National Television System Committee
is used in Japan and in most 6O-Hz countries, including the United States. (Most 50-Hz countries
use PAL or SECAM.) Most of the satellite capacity would be used by NHK, but some would be
made available to independent broadcasters, who presumably would use NTSC or some enhanced
version of NTSC. It was not originally intended to use HDTV for terrestrial broadcasting.
2 This is not a very well defined quality, since movie quality is highly variable and steadily improving.
It should be noted that 35-mm amateur slides have more than twice the area of movie film. The
I 125-line system, as so far demonstrated, does not have nearly the resolution of amateur slides.
3 FM is used today for satellite TV transmission. The rf bandwidth of the satellite channel is usnally
four to five times that of the basebandwidth of the TV signal. It is not out of the question to use
digital transmission for DBS in the future.

268
system was announced in 1984. MUSE has been demonstrated many times by
DBS, and was also demonstrated in terrestrial UHF service in Washington in
January 1987, using two adjacent channels, for a total of 12 MHz rf bandwidth.
A MUSE DBS system is operating one hour per day in Japan at the present time,
and regular broadcasts are scheduled for late 1990 or 1991.
The National Association of Broadcasters (NAB) and the Association of
Maximum Service Telecasters (MST) urged NHK to develop systems that were
compatible with (i.e., could be received by) NTSC receivers and that would be
suitable for the 6-MHz channel-allocation scheme used in the US. Several such
systems, called the "MUSE Family," were demonstrated by computer simulation
at the NAB convention in April 1988 and in hardware at NAB in April 1989.
System and equipment developments were paralleled by efforts to have the
"studio" system adopted as an international standard for program production and
international exchange, but this has not yet been accomplished, primarily because
of opposition by European interests. Intense controversy exists at the present time
over this issue.

A2.2 The "Studio" System

A definitive description of the original 1125-line system was given by NHK in
a detailed technical report issued in 1982 [A.3]. It was stated that the system
parameters were chosen to give a more psychologically satisfying viewing ex-
perience in which the viewing angle would be large enough to give some sense
of depth (30 0) and the resolution high enough so that at this viewing distance
(about three times the picture height or 3H) the image would not be noticeably
blurry. A wider screen (5:3 rather than NTSC's 4:3) was felt to be very important
for an improved sense of realism. 4
Actually, as in any system design, the final parameters are the result of
compromises between ideal and affordable values. In this case, widening the
screen and increasing the viewing angle as well as the resolution all lead to
wider video bandwidths. When the bandwidth is increased, camera sensitivity
and signal-to-noise ratio (SNR) go down, costs go up, and fewer programs can
be accommodated within an overall spectrum allocation.
Generally speaking, the NHK choices were reasonable for the intended pur-
pose. There were two questionable decisions, however. One was to use interlace
and the other was to use vestigial-sideband transmission (VSB) for AM trans-
mission. The use of interlace is justified only on the basis that HDTV is to be
a straight-through system without any storage. As a result, the cameras, channel

4 When widescreen movies were introduced in the 1950s, 2.35:1 was the favored aspect ratio, not
only to enhance the viewing experience, but also to be as different from television as possible.
At the present time, l.85:1 is the most common film aspect ratio, partly because aspect ratios
much wider make cropping for 4:3 television much more damaging to the artistic effect. Note that
actual practice varies a great deal. Typical TV viewing angles are much smlIller than typical movie
viewing angles.

269
signal, and display are all synchronous, just as in the NTSC monochrome stan-
dard adopted in 1941. Even in 1970, many would have argued that progressive
scan would give much better results and would simplify transcoding between
various standards. Twenty years later, it appears that all new TV systems will
use frame stores, so that the argument for interlace is now even weaker. 5
Since the "studio" system was originally designed for PM satellite transmis-
sion, it is possible that the decision to use VSB for AM applications simply did
not receive enough attention at the time. At present, it appears to be a simple
error. Double-sideband quadrature modulation, as is used for the color informa-
tion in NTSC, is much more effective. For example, MUSE would require an rf
bandwidth of just over 8 MHz with DSBQM, whereas with VSB, there is con-
siderable doubt as to whether the claim that it will fit within 9 MHz is actually
justified.

A2.3 MUSE
The "studio" system required a bandwidth of 32.5 MHz - 20 for luminance
and 7 and 5.5 for the two chrominance signals. To fit into one normal FM
transponder channel, a bandwidth reduction of about 4: 1 is needed. MUSE ac-
complishes this by two methods. Halving the diagonal resolution by means of
transmitting only every other signal sample, interleaved line to line so as to give
a 45° sampling pattern, provides a factor of 2. Another factor of two is gained
by reducing the frame rate to 15 fps by sending alternate samples on alternate
frames. 6
At the receiver, a clever motion-adaptive interpolation method completely
eliminates the sampling structure due to the effective 4: 1 interlace and also elim-
inates much of the blurring of moving objects that would be expected with
nonadaptive interpolation. The basic scheme is to interpolate temporally in the
stationary image areas and spatially in the moving areas. This technique reduces
the resolution of moving objects by a factor of two. The effect would be most
noticeable when the cameras were panned. The original system used one motion
vector for panning, the receiver performing motion-compensated interpolation.
MUSE pictures have been getting better and better, however, and it is possible
that more motion vectors are now being used. Some believe that an independent
vector is used for each of 9 image segments.

5 Some workers do believe that a progressive-scan HDTV camera is beyond the state of the art at
present, but many others would disagree.
6 MUSE receivers must use frame stores to derive a 6O-field display from the transmitted 30-field
signal. This removes the principal obstacle to using progressive scan in the studio system, making
the use of interlace even more questionable.

270
 1200
2250-line
------------------,
" progressive display I Fixed area: 30 Ips
, \ I
Moving area: 15 Ips
"'- NHK Wideband System I
,, I II
',1125-line I
I------~--. - - - --:-- - -l Ken lactor limit
Interlaced display I
I
f--- 20-MHz bandwidth limit
I
Fixed area I
I
I

' , / Diagonal Ii Iter limit

,,
" ,,
,
o 400 800 1200
Horizontal resolution
Lines (samples) per picture width

Fig. A.2. Resolution of the NHK Wideband System and MUSE. Both systems have about 1070
active lines, so that the resolution on an interlaced display is about 750 Iines/picture height. The
exact horizontal resolution of the wideband system is governed by camera resolution and bandwidth
assigned to each component. At bandwidths of 20,7, and 5 MHz, this comes out to 1007, 353, and
252 for L, Cl, and C2 respectively. The use of a diagonal filter and 4 x subsampling for bandwidth
compression in MUSE lowers the diagonal resolution as shown. The moving-area horizontal reso-
lution is also halved to give the results indicated in the diagram. The static luminance resolution is
further limited by the camera and display, but the other resolutions are set by the system parameters

MUSE pictures look very good. Although there is some loss of sharpness with
motion, the reduction in picture quality is much less than might be expected with
a 4: 1 reduction in the transmission rate of image samples. It may well be that
limited camera-tube resolution is masking some of the loss in resolution. The
lack of interline flicker due to interlace (interline flicker is often noticeable in
NTSC) gives strong evidence for this speculation. If this assumption is true, then
the difference between the "studio" system and MUSE will eventually become
more evident with the development of better camera tubes. Figure A.2 shows the
resolution of both the "studio" system and MUSE.

A2.4 Systems Specifically Designed for US Broadcast Use

Since the original development work was intended for DBS use, and not to
replace NTSC and PAL in terrestrial broadcasting, no attention was given in
Japan to conforming to terrestrial channel widths used in the US. Likewise, the
overwhelming preference of US broadcasters for a compatible system seems not
to have been made evident to NHK until the specific request of NAB and MST

271
to develop such systems.7 As a result, the Japanese are not ahead of the world
in HDTV system design; they are actually several years behind. Specifically,
there is no evidence that they have yet started thinking seriously about systems
designed to work well in today's over-the-air channels.

A2.4.1 Narrow MUSE

This is a version of MUSE in which the transmitted signal is 750 lines, 2: 1
interlace, 60 fps, with a bandwidth of 6 MHz. The main performance difference
from the original version of MUSE (now called MUSE-E) should be a reduction
in diagonal resolution and some loss of vertical resolution. As actually exhibited,
the motion compensation did not seem to be as effective, but no doubt this
deficiency can be removed. The 750-line signal is derived from an 1125 source
and reconverted to 1125 lines in the receiver for display. A receiver capable of
handling NTSC and Narrow MUSE as well as MUSE-E would cost little more
than the normal MUSE receiver.

A2.4.2 NTSC "MUSE"-6

Compatible systems must accommodate the difference in aspect ratio between
NTSC and HDTV, and this may well be one of the most difficult problems.
Certainly it has been the subject of a great deal of shallow thinking. NHK showed
two different methods of accomplishing this. In their "top-and-bottom-mask"
method, 25% of the theoretical picture height is left empty of picture material
on the NTSC receiver, so that the remaining image area is 16:9. Thus the image
content and shape shown on the old and new receivers are the same. The bar
areas can be used for enhancement information. This is the method proposed for
the MIT-RC system and is much like the "letter-box" format used in Europe for
wide-screen movies shown on TV.
The "side-panel" method, like that of the Sarnoff system (ACTV), transmits
information for the extra picture area by hiding it within the NTSC signal. The
extra area encompassed within the side panels is not seen on the NTSC receiver.
It therefore must be devoid of significant picture information, greatly reducing
the visual impact of the wider screen. NHK advocates the bar method rather
than the side-panel method, partly for this reason and partly because the seams
between the side panels and central section are often made visible by channel
impairments.

7 In my opinion, the insistence on receiver compatibility by American and European broadcasters,

although conforming to their short-term interests, is definitely not in their long-term interests.
Compatible systems cannot possibly provide the high channel efficiency or resistance to analog
channel impairments that are now available in completely new systems, since the deficiencies
of NTSC and PAL in this regard are inherent in their design. If broadcast TV cannot compete
in picture quality provided by alternative media, then as these media become able to provide
progranuning of equal attractiveness as that of the broadcasters, the latter will lose audience share
and eventually go out of business.

272
 The receiver-compatible NHK systems are not MUSE systems at all, since
they do not use subsampling. Several methods are used to hide information, in
addition to placing it in the top and bottom bars. Some is hidden in the vertical
and horizontal blanking intervals. High-frequency enhancement information is
transmitted at a lower frame rate, only in the stationary areas of the image, by
multiplexing it with the upper half of the NTSC luminance and chrominance
signals. This is an indirect way of using the "Fukinuki hole," [A.4] the portion
of 3-dimensional frequency space diametrically opposite to that occupied by the
color subcarrier and its side bands.

A2.4.3 NTSC "MUSE"-9

This system is much like the previous one, but in addition uses another 3-MHz
channel for augmentation information. In the version shown, this extra capacity
was used for digital audio and for increasing the vertical and horizontal resolution
in moving areas, which are of quite low resolution in the 6-MHz version.
An interesting point is that NHK itself places, in rising order of quality,
NTSC MUSE-6, NTSC MUSE-9, Narrow MUSE, and MUSE-E. Thus Narrow
MUSE, with a 6-MHz bandwidth, is better than NTSC MUSE-9. This shows
quite clearly the penalty that is paid for building NTSC compatibility into any
new television system.
As exhibited at the 1989 NAB meeting, Narrow MUSE was distinctly inferior
to MUSE-E in resolution, and the moving-area interpolation did not seem to be
working properly. Both NTSC-compatible systems showed such serious defects
in moving areas that it is hard to understand why they were exposed to public
view. Table A.l and Figures A.3 and AA show some aspects of the MUSE
Family [A.5].

A2.S Conclusions: Japan

The 1125/60 "studio" system makes rather good pictures, but the use of interlace
is a major drawback in other than the originally intended applications. MUSE,
although quite sophisticated, is closely tied to the original system. Although
clearly inferior in quality, it makes very good pictures by DBS under conditions
of adequate CNR. Its applicability to terrestrial broadcasting and to cable has yet
to be adequately demonstrated. The 6-MHz and compatible versions developed
specifically for US use are, as of this date, not as good. They certainly cannot
have the resistance to channel impairments and the higher spectrum efficiency
attainable with entirely new designs, and transcoding to NTSC and PAL are likely
to be quite expensive for some time to come.

273
TableA.1. The MUSE Family. There are many variants here because of the necessity of conforming
aspect ratio. NHK showed two different methods. The "top-bottom mask" method is generally called
"Ietterbox" format in this country. Blank bars are left at top and bottom so that the same image
material is shown on the two kinds of receivers. In the "side-picture" method. generally called "side-
panel" here. the sides of the wide-screen (16:9) HDTV picture are cut off for display on the NTSC
(4:3) screen
Resolution HDTVr2

.~
Compati- Aspect ratio NTSC'3
System Bandwidth (lines/£icture ADTV
bility expansion wi th) aspect ratio aspect ratio

HDTV MUSE 9MHz no· 1 1020 16:9 [Link]

NARROW
MUSE
6MHz no· l - 1010 16:9 16:9/4:3

top-bottom 960
16:9
NTSC- mask 900
ADTV MUSE-9 9MHz yes 16:9
side 900 4:3
picture
top-bottom 960
16:9
NTSC- mask 680
MUSE-6 6MHz yes 16:9
side 900 4:3
picture

*1 An experimental low-cost downconverter has already been realized.

'2 Aspect ratio displayed on HDTV and ADTV receivers.
'3 Aspect ratio displayed on NTSC receivers.

A3. Europe

Although there was originally some indication of acceptance of 1125/60 for

studio use by the European Broadcasting Union, the potential economic effects
eventually caused Europe to reject that system emphatically in favor of one more
in harmony with its own view of the future of television. The major broadcasting
and manufacturing interests agreed on a system in 1986, although all details had
not been worked out by then. The principal decisions were that HDTV was to be
delivered exclusively by DBS and that it was to be 'compatible' with PAL and
SECAM. (It is not really compatible, since the signal format is quite different,
but the cost of transcoding is potentially quite low, so that it can be hidden within
the cost of the satellite receiving equipment.) Thus, many viewers are likely to
be watching the new broadcasts on their old receivers for a long time to come.
Work on the system is being carried out under project Eureka 95, which involves
both industrial and government investment.

A3.1 Multiplexed Analog Components

Independently of HDTV, the Europeans had developed, in connection with DBS,

a number of so-called MAC systems in which the composite video formats
of PAL and SECAM were abandoned in favor of transmitting the three color
components in sequence. The time bases of these components are changed so

274
 1125/60
STUDIO STANDARD

HDTV Program Transmission by

Communication Satellite

Program Distribution by
DBS, CABLE, VCR

ADTV Program Distribution by

Terrestrial Broadcast

Fig. A.3. The MUSE Family. NHK has proposed a family of systems that fit within the US 6-MHz
channel and/or are compatible with NTSC receivers. All these systems are derived from the wideband
system by filtering and subsampling. MUSE-T is a higher-resolution version of standard MUSE, now
sometimes called MUSE-E. Narrow MUSE is a 6-MHz version that achieves its compression by
reducing the number of scan lines, and thus, the vertical resolution. It is otherwise quite similar to
MUSE and can be processed almost entirely in standard MUSE coders and decoders. The NTSC-
compatible systems are not MUSE at all, but utilize a variety of methods to add information to
a standard NTSC signal. NTSC MUSE-6 uses a single channel, while NTSC MUSE-9 uses an
additional 3-MHz augmentation channel. NHK rates the quality of NTSC MUSE-9 as slightly below
Narrow MUSE

that they can fit within the line duration in proportion to their relative bandwidth.
MAC systems eliminate luminance!chrominance cross-talk and are also better
suited to the triangular noise spectrum characteristic of FM. One such system,
D2-MAC [A.6], also has provision for four independent channels of digital audio.
Conversion to PAL or SECAM is not very difficult, and would be implemented by
adding some circuitry to the satellite receiver required, in any case, to utilize DBS.
MAC systems are also suitable for digital transmission or storage in accordance
with CCIR Recommendation 601, which calls for a total data rate of 216 Mb/s.

A3.2 Overall European Strategy

A fundamental aspect of European strategy has been to prevent Japanese domi-

nation of the European market, and, preferably, of the American market as well.
The decision to maintain at least the appearance of compatibility with PAL and

275
 510
~ static resolution
c-&,
0·"
." CD
:i.e
-o ..CD
.. ::I
CD . .
.. U

]~
'.;3
.. c
~
moving resolution
CD·"
>c

170 460 580 960

horizontal resolution
(lines/picture width)

(a) NTSC·MUSE·6

690

~
c on NTSC static
0·"
." CD
:;.e 340 -----------------,
I
-o ...
CD
I
.. ::I
"'..,
.. U
I
I static resolution
"Cd'a I
I
uUl I
"..:J I
... c
Q)

'" 0"
>c moving resolution

440 670 1010

horizontal resolution
(lines/picture width)

(b) Narrow·MUSE
Fig. A.4a,b. Resolution of the MUSE Family. These NHK figures show their assessment of the
moving and static resolution of the MUSE Family. Note that Narrow MUSE is much like MUSE·E
except for the lower vertical resolution. The odd shapes of the resolution outlines for the NTSC-
compatible systems is due to hiding various pieces of enhancement infonnation in various places in
the NTSC signal

SECAM may well have been as much related to this point as to concern over the
usefulness of the existing large installed base of receivers. In any event, the Eu-
ropean decisions were as different from Japanese strategy as possible; the HDTV
system was to be compatible and it was to be implemented in an evolutionary
manner. In view of the implication that existing systems are not to be dropped
for many years and since there are no spare terrestrial channels available, a de-
cision that is hard to criticize was to transmit high-definition signals solely by
DBS. Thus, HD-MAC is really compatible with normal-definition MAC, and not
with PAL or SECAM. The latter are accommodated in the same way as when
any MAC format is used. It is a format-conversion process in which everything
about the signal is changed except the line and frame rate.

276
A3.3 Technological Approach
The more obvious methods by which a high-definition system can be made
compatible with existing systems are discussed in Sect. 8.6. One is to hide aug-
mentation infonnation within the existing signal and the other is to utilize an
additional half or full channel. A third method, proposed for the US by Iredale
[A.7] but since abandoned, involved straightforward subsampling of an HDTV
signal to obtain the nonnal fonnat. For example, if an HDTV signal has exactly
twice the number of lines/frame and twice the number of pels/line and the same
frame rate as NTSC, taking half the pels on half the lines would give a signal with
the same line and frame rate as NTSC. (Subsampling is complicated somewhat
by interlace.) The principal problem with this simple-minded approach is that
the subsampled signal exhibits aliasing, as the higher vertical- and horizontal-
frequency components are folded back into the baseband spectrum by sampling
at less than twice the bandwidth. The fonn of the aliasing depends on the sub-
sampling grid. In a grid such as in MUSE, four fields are used, so that 1/4 of
all the pels in a frame are sampled from each field. As a result, IS-Hz flicker
(12.5Hz for PAL) appears around sharp edges. The effect is totally unacceptable
[A.8].
To reduce the aliasing on old receivers, the image can be blurred by low-
pass filtering, which, of course, also reduces sharpness on new receivers. The
sharpness can be restored by high-pass filtering at the expense of an increase
in the noise level. The question is whether there is a compromise that does
not reduce the compatible quality excessively while at the same time provides
worthwhile improvements on new receivers. The answer seems to be "no". The
reason is simple and is independent of the cleverness of the system designer.
The augmentation infonnation required to make the 'HDTV' picture better than
the 'NTSC' picture comprises precisely the signal components that cause the
aliasing on the old receivers.
A possible way out of this dilemma is to use motion-adaptive prefiltering in
combination with motion-compensated interpolation in the new receivers. This
is only a possible solution, as can be seen by considering the zero-motion case.
If the four-field sampling pattern is a 2 x 2 square, the four pels of the high-
resolution picture will appear in the same location on the low-resolution raster.
If these pels are not identical, as required to retain the higher resolution at the
HDTV receiver, flicker must result on the standard receiver. The motion-adaptive
filtering and interpolation in this case can only be parallel to the time axis and
thus cannot affect the result.

A3.4 Three-Branch Motion Compensation

The algorithm now used in Eureka classifies each 16 x 16 block of each frame
as having rapid (more than 12 pels/frame), slow (0.5 to 12 pels/frame), or
zero (less than 0.5 pels/frame) motion. The slow blocks are reconstructed us-
ing motion-compensated interpolation. The velocity vectors are detennined by a

277
phase-matching technique and are transmitted in a side channel with a capacity
of about 1 Mb/s. Different prefilters, postfilters, and sampling patterns are used
for the three branches. The image is reconstructed at the coder independently
from all three branch signals, the one producing the smallest error being chosen
in each block. The frequency response that results is shown in Fig. A.5 [A.9].
Obviously, there are many parameters that can be used for optimizing this tech-
nique, such as the filters, the accuracy of velocity information, the block size,
the subsampling pattern, etc., so that the quality that may ultimately be obtained
is not definite at the present time.
A basic assumption that underlies the use of multiple branches depending on
the amount of motion is that the spatial acuity of the human visual system depends
on velocity. While this must be true to some extent (hard data on this point is
lacking), in the case of tracking of objects moving at constant velocity within the
image, after 200 ms or so the required resolution is not much less than needed
for stationary objects. In the original BBC demonstrations of digitally assisted
TV (DATV, the technique now used in HD-MAC), the halving of resolution of
objects in motion when the motion compensation was turned off was clearly
visible and most annoying [[Link]).

1152

Stationary areas
Slowly moving areas
576
Rapidly moving areas

O~---r---'----r-~'----.---r---'~~--­
o 720 1440
Horizontal resolution
Lines (samples) per picture width

Fig. A.S. Luminance Resolution of HD-MAC. The original image of 1152 active lines and 1440
resolvable elements per line is sampled in an offset (quincunx) pattern. In the stationary areas, all
samples are transmitted in 4 frames, in the slowly moving areas, half the samples are transmitted in
2 frames, and in the rapidly moving areas, one-fourth of the samples are transmitted every frame.
This results in the theoretical spatial~frequency responses shown here. For comparison, see Fig. A.2

278
A3.5 Conclusions: Europe
The spirit of cooperation shown in Europe in agreeing on a common approach and
in coordinating the work of many laboratories in the various countries involved
is most remarkable. It has yet to be demonstrated just how good the images will
be while retaining compatibility. The decision to use a compatible system is not
technological, but is based on an assessment of the nature of the dynamics of the
TV market and on the likely behavior of consumers. Clearly, non-Europeans are
not in as good a position to judge these matters as the Europeans themselves.
However, depending entirely on improved picture quality to motivate prospective
purchasers would seem to be a risky approach anywhere.
For purely technological reasons, the constraint of compatibility must reduce
the image quality that can be attained within a given channel capacity. Since the
subsampling technique used in HD-MAC is rather similar to that used in MUSE
(if it proved to be substantially better, it could be adopted by MUSE), it would
seem that it will be very hard to match the ultimate quality of MUSE, which
does not have this serious constraint.

A4. The United States

The most striking feature of the HDTV scene in the US is that the domestic
consumer-electronics industry, whose research and product-development expen-
ditures would normally be expected to produce systems particularly adapted to
the many unique features of American television broadcasting, has been deci-
mated by foreign competition. 8
The result was to leave the field open to the Japanese. There was no deny-
ing the high image quality produced both by the 1125/60 "studio" system as
well as MUSE, although some did voice objections to motion blurring in the
latter. Consequently, in the absence of American alternatives, there was a sub-
stantial amount of support in the US for adopting 1125/60 as a production and
international exchange standard. The question of transmission in the American
environment was put aside. However, when it appeared that HDTV might first
be introduced by alternative media, the industry petitioned the FCC for a formal
inquiry, which began in September 1987 with the appointment of an Advisory
Committee made up primarily of TV industry executives. Congressional hearings
were also held, at which the trade implications of adopting a foreign standard
surfaced. When transmission began to be considered, the broadcasters' preference
for compatible systems resulted in a turning away from the Japanese approach.

8 Some of this has been ''fair,'' in that it represents the results of superior product design, but the
most important factor was systematic and illegal dumping [AJ!). The departure of a number
of American companies from this low-profit field for more lucrative areas such as military and
medical electronics is another. In any event, at this date, only one large American-owned company,
Zenith, still makes TV receivers, and most of those are assembled in Mexico.

279
A4.1 The FCC Inquiry

Although originally intended simply to assess the probable impact of HDTV on

TV broadcasting in the US, the Inquiry has developed into a selection process
for an American system for terrestrial transmission service. Under the general
supervision of the FCC, the Advisory Committee has set up a testing procedure
using the facilities of the Advanced Television Test Center, which in turn was set
up and financed by the TV industry. Systems are to be tested, in full hardware
implementations only, during a period of about one year beginning sometime in
1990, with the FCC due to make a selection in 1992.
In September 1988, the FCC issued a Tentative Decision in which it was
stated that the government would set standards for terrestrial broadcasting only,
the other media being free to adopt whatever standards they pleased. The most
important part of the ruling stated that, at least for an initial period, all programs
broadcast in HDTV must be made available to NTSC receivers and that no
more than 12 MHz could be used to serve both kinds of receivers. The three
methods available to do this are single-channel compatible systems such as ACTV
[A.12], augmentation systems such as that of Philips [A.13], and single-channel
noncompatible systems that would utilize simulcasting [A.14, 15].
This decision was widely misreported as indicating that the FCC had ruled
in favor of compatible sytems. In fact, nearly the reverse is true. During the
Inquiry, it has become evident that, in order to find enough channel capacity
to accommodate both NTSC and HDTV during a transition period, the HDTV
and/or augmentation signals would have to operate at much lower DIU ratios
than NTSC, which means that they must abandon the NTSC signal format. The
decision did eliminate standard MUSE, which would have required at least 15
MHz for both services.

A4.2 Proposed Systems

At a time when no American systems had been proposed, the Association of Max-
imum Service Telecasters (MST) and the National Association of Broadcasters
(NAB) requested the Japanese to produce versions of MUSE that were compat-
ible with NTSC receivers and/or were compatible with the American spectrum
allocation of 6 MHz for each channel. As a result, the MUSE family, discussed
in Sect. A2.4, was developed. Two of these systems will be presented for testing,
the 6-MHz version of standard MUSE and a compatible system as well. Aside
from the Japanese entries, the largest amount of funding has gone to ACTV, the
single-channel compatible system developed at the Sarnoff Laboratories. ACTV-
II, like the first version but with a second channel for enhancement information,
was orginally proposed but appears to have been dropped. The compatible aug-
mentation system of North American Philips is also a contender. Single-channel
noncompatible systems are offered by NHK, Zenith, and MIT, and Faroudja
Laboratories has proposed an enhanced NTSC system [A.16].

280
 Although noncompatible systems that would require simulcasting were orig-
inally derided by the entire American TV industry except Zenith, the effect of
the Inquiry has been to change many opinions. It is now generally accepted that
true HDTV is possible within 6 MHz if an entirely new system is used, and that
such new systems would require much lower transmitter power than now needed
for NTSC. The importance of achieving much better interference performance
for the purpose of higher spectrum efficiency is now widely understood, espe-
cially within the FCC itself. It seems to be understood that this cannot be done
with NTSC. As a result, in a most remarkable move, North American Philips
and Thomson USA (formerly the GE and RCA consumer-electronics divisions)
have agreed to cooperate with the National Broadcasting Corporation (NBC) to
develop a simulcast system. The effect of this decision on ACTV, ACTV-II, and
the Philips augmentation system were not clear at the time of the announcement.
In March 1990, the FCC issued another ruling that affected the competition
for the US transmission standard. It stated that augmentation systems would no
longer be considered, as they were not sufficiently bandwidth-efficient. 9 It also
said that it would issue a standard for EDTV, if ever, only after an HDTV standard
was selected. Since all the compatible systems then proposed were EDTV, and
not HDTV, that meant that the FCC had concluded that the transmission standard
to be selected would be noncompatible with NTSC, and that simulcasting would
be used to service NTSC receivers. Of course, this decision depends on the
successful operation of the simulcast systems, which has yet to be conclusively
demonstrated.
As a result of these rulings, Faroudja has dropped out of the competition, the
NHK compatible system has been withdrawn, and all the augmentation systems
have been abandoned. However, ACTV remains alive, although its only hope
would appear to rest on the possibility that none of the simulcast systems will
work. The Thomson-Philips consortium has an analog simulcast system on the
table, but it appears that this will be replaced by an all-digital system from the
Sarnoff Laboratory.

A4.3 Digital Broadcasting?

Recently, there has been a great deal of interest in digital terrestrial broadcast-
ing. This is mainly due to the proposal by General Instrument Corporation for
an all-digital HDTV coding and transmission system. Several other digital sys-
tems are likely to be proposed in the near future. The main argument made for
these systems is that they permit the use of much more effective coding methods.
However, that is a question of source coding. The proposal to use digital trans-
mission is a question of channel coding, a quite different matter. In fact, most

9 This should not have taken an FCC ruling. Augmentation systems, which transmit NTSC in one
channel and enhancement information in a second channel, combine all the worst features of new
and old systems. In addition, they preserve the deficiencies of NTSC forever. It is a mystery to
the author how this simple conclusion was not evident to everyone at an early date.

281
of the currently proposed HDTV systems do use digital coding and decoding.
The chosen transmission format in no way limits the efficiency of most source
coding algorithms. tO The practical limit to coding efficiency may well be ilie
associated complexity of receiver signal processing. With the economies of scale
that are possible in manufacturing signal-processing components for millions of
receivers, extremely complex processing will become economical.
A particular feature of terrestrial broadcasting is iliat receivers located at dif-
ferent distances from the transmitter have different SNRs and therefore widely
varying channel capacities. ll In today's analog broadcasting system, this mani-
fests itself by a variation of picture quality with distance. (At very great distances,
viewers tend to install antennas just adequate to receive tolerable pictures; under
these conditions, with NTSC, sound quality is quite good.) Conventional digital
systems have sharp thresholds, and, above threshold, iliey deliver the same num-
ber of bits/s to all receivers in the service area. This means iliat, in ilie central
city, channel capacity that is actually available (due to the higher SNR) and that
could be used, in principle, to improve picture quality or for other purposes,
is not utilized. Unfortunately, iliis is just the area where ilie spectrum shortage
triggered by the rapid growth of cellular telephone services will be most acute.
It follows from this argument that some way must be found to utilize iliis wasted
capacity.
High efficiency in channel utilization, i.e., signalling at a rate even close to
ilie theoretical channel capacity, whether the transmission is analog or digital,
requires iliat the channel be properly equalized and that echoes be substantially
eliminated. This degree of equalization is mandatory for digital transmission at
the Nyquist rate. When this is done, independent analog samples can also be
transmitted at the Nyquist rate, without intersymbol interference.
In the MIT-CC system, a hybrid analog-digital transmission format is used
in which ilie selected sub band samples are transmitted in analog form, super-
imposed on the digital data, which uses a conventional QAM constellation. At
the threshold SNR, only the digital data is recovered. This is used to produce a
noiseless picture of a certain resolution. At higher SNR, the analog data starts
to become useful, and the resolution of the recovered picture improves. Thus, a
soft threshold is provided, much as in today's TV systems. As compared with
all-digital transmission, using source coding systems of comparable efficiency, a
hybrid system of this type gives better quality in the central city (where most of
the viewers reside) and somewhat poorer quality at the boundary of the service
area.

10 As long as the coded data includes selected signal samples or their equivalent, as in the OCT, the
channel coding method does not affect the efficiency of source coding. An exception would be
vector quantization, which is not being used in any of the HDTV system proposals.
11 Channel capacity is not defined solely by bandwidth. The maximum transmission rate is pro-
portional to the product of bandwidth and SNR in dB. To fix the latter, we must make some
assumptions about radiated power, antennas, and receiver noise figure, of course.

282
 It should be noted that all of the digital systems currently proposed transmit
some data that must be carefully protected, and that therefore should be transmit-
ted digitally, and other data, such as OCT coefficients, which is entirely suitable
for analog transmission in the same manner as in the MIT-CC system.
These conclusions, within limits, are independent of the performance of the
compression system. Of course, if someone discovers how to transmit true HDTV
in 1 Mb/s, the argument about transmission efficiency becomes much weaker,
since much less than 6 MHz would be needed for all-digital transmission. How-
ever, when it is proposed to do HDTV in 10 Mb/s in a 6-MHz channel (and no
one has yet claimed to be able to use such a low data rate), a possible regula-
tory response would be to allocate only a 3-MHz channel, and to mandate the
use of 5 Mb/s plus 5 million analog samples/so Such a scheme would provide
much better pictures to a majority of viewers and would completely solve the
spectrum-shortage problem.
The hybrid analog-digital transmission format may not be the only way, or
even the best way, of achieving high channel efficiency in cases where there is
a large variation in SNR among the receivers. It is conceivable that this can be
done with an all-digital scheme. Such a scheme is theoretically possible [A.17].
The most that can be claimed for the hybrid format is that it provides a solution
for a very important problem in spectrum allocation that is not provided by
conventional digital formats.

A4.4 Conclusions: United States

Disarray brought about by the demise of the domestic consumer-electronics in-
dustry is the main feature of the American scene. The realization that over-the-air
broadcasting of HDTV requires a totally new approach has come only within the
last year. Of all the systems proposed that might result in true HDTV, only
the Zenith system (and perhaps also the General Instrument System) is from an
American-owned company that is large enough to marshall the resources required
to become a significant actor. Both Philips and Thomson are foreign-owned, and
even the ACTV system itself will probably eventually fall under foreign owner-
ship. The trade implications of HDTV are quite clear in Congress, although the
federal government thus far has not committed itself to any strong action in the
field.

AS. Overall Conclusions

The bandwidth required to deliver HDTV to homes is not available, except po-
tentially via fiber-optic cable, so compression systems are absolutely required.
In addition, analog channel impairments must be eliminated in order to preserve
the higher quality. Both Japan and Europe have decided to use subsampling sys-
tems and DBS to solve these problems, but the Japanese system is deliberately

283
noncompatible with NTSC, while the European system is effectively compatible
with PAL and SECAM. Only time will tell which of these approaches is more
successful in the marketplace and, indeed, what quality will be obtained in prac-
tice. In the US, terrestrial broadcasting will be a main delivery method. This
requires a totally new system that is not only immune to channel impairments,
but will provide much higher spectrum efficiency. If such a system is actually
implemented in the US, it will be very attractive to use in other countries as well,
since the higher spectrum efficiency will permit many more channels to be made
available within a given overall frequency allocation, and because, unlike both
other approaches, it will be usable for terrestrial transmission as well as DBS.

284
References

Chapter 1

1.1 Image Processing at the Jet Propulsion Lab. See J. Geophys. Res. 82 (28), 3951-4684 (1977)
(entire issue); see also Proceedings of the CALTECHlJPL Conference on Image Processing,
JPL SP 43-30, California Institute of Technology, Pasadena, Califomia, November 3-5 (1976);
W.B. Green: "Computer Image Processing, the Viking Experience", IEEE Trans. CE-23 (3),
281-299 (1977);
J.M. Soha et al.: "IPL Processing of the Mariner 10 Images of Mercury," J. Geophys. Res. 80
(17),2394-2414 (1975)
1.2 J. Allnatt: Transmitted-picture Assessment (Wiley, New York 1983)
1.3 J.M. Barstow, HN. Christopher: ''The Measurement of Random Video Interference to
Monochrome and Color Television Pictures," Trans. AIEE (Comm. and Electron.) 81, Pan
I, 313-320 (1962)
1.4 F.T. Percy, T. Gentry Veal: "Subject Lighting Contrast for Color Photographic Films in Color
Television," J. SMPTE 63,90-94 (1954)
1.5 P. Mem, F. Grey: "A Theory of Scanning," Bell System Tech. J. 13,464-515 (1934)
1.6 D.G. Fink: Principles of TV Engineering (McGraw-Hill, New York 1940)
1.7 W.M. Goodall: "Television by PCM", Bell System Tech. J. 30 (1), 33-49 (1951)
1.8 W.F. Schreiber: "Psychophysics and the Improvement of Television Image Quality," J. SMPTE
93 (2), 717-725 (1984)
1.9 DR. Kelly: "Theory of Flicker and Transient Responses, I: Uniform Fields," J. Opt. Soc.
Am. 61 (4), 537-546 (1971); "II. Counterphase Gratings," J. Opt. Soc. Am. 61 (5), 632-640
(1971); Ph.D. Thesis, University of California, Los Angeles (1960); "Adaptation Effects on
Spatio-temporal Sine-wave Thresholds," VISion Res. 11, 89-101 (1972)
1.10 M.W. Baldwin, Jr.: "Demonstration of Some Visual Effects of Using Frame Storage in Tele-
vision Transmission," IRE Conv. Rec., p. 107 (1958) (abstract only)
1.11 JL. Cunningham: ''Temporal Filtering of Motion Pictures," ScD. Thesis, Massachusetts in-
stitute of Technology, Electrical Engineering Department (1963)
1.12 J.R. Ratzel: "Discrete Representation of Spatially Continuous Images," ScD. Thesis, Mas-
sachusetts Institute of Technology, Electrical Engineering and Computer Science Department
(1980)
1.13 RD. Kell, A.V. Betford, M.A. Trainer: "An Experimental Television System," Proc. IRE 22
(11), 1246-1265 (1934)
1.14 S.C. Hsu: "The Kell Factor: Past and Present," J. SMPTE 9S (2), 206-214 (1986)
1.15 N.T.S.C.: Television Standards and Practice (McGraw-Hill, New York 1943)
1.16 B.M. Oliver, J.R. Pierce, C.E. Shannon: "The Philosophy of PCM," Proc. IRE 36, 1324-1331
(1948)

285
Chapter 2

2.1 A. Chappel (ed.): Optoelectronics, Theory and Practice, Texas Instruments Electronics Series
(McGraw-Hil, New York 1978);
D.E. Gray (ed.): American Institute of Physics Handbook (McGraw-Hill, New York 1957)
2.2 A.C. Hardy: Handbook of Colorimetry (MIT Press, Cambridge, MA 1936)
2.3 W.G. Driscoll (ed.): Handbook of Optics, Optical Society of America (McGraw-Hill, New
York 1978)
2.4 M. Born, E. Wolf: Principles of Optics, 6th ed. (Pergamon, New York 1980)
2.5 A. Papoulis: Systems and Transforms with Applications in Optics (McGraw-Hill, New York
1968)
2.6 R.E. Hopkins, MJ. Buzawa: "Optics for Laser Scanning," Opt. Eng. 15 (2),90-94 (1976)
2.7 J.W. Goodman: Introduction to Fourier Optics (McGraw-Hill, New York 1%8)
2.8 [Link]. Fourier: Theorie analytique de la chaleur (Didot, Paris 1822)
2.9 A. Papoulis: In [2.5], p. 145
2.10 E.R. Kretzner: "Statistics of TV Signals," Bell System Tech. J. 31, 763 (1952)
2.11 E. Dubois, M.S. Sabri, J.Y. Ouellet: "Three-Dimensional Spectrum and Processing of Digital
NTSC Color Signals," J. SMPTE 91, 372-378 (1982)
2.12 D.G. Fink: "The Future of High Definition Television," J. SMPTE 89, 89-94 (1980)
2.13 M. Rittennan: "An Application of Autocorrelation Theory to TV," Sylvania Technologist
70-75 (1952)

Chapter 3

3.1 S.S. Stevens: Handbook of Experimental Psychology (Wiley, New York 1951)
3.2 T.N. Cornsweet: VISual Perception (Academic, New York 1970)
3.3 S. Hecht: "Quantum Relations of Vision," J. Opt. Soc. Am. 32, 42 (1942)
3.4 A. Rose: VISion: Human and Electronic (Plenum, New York 1973)
3.5 M.E. Chevreul: Laws of Color Contrast (Routledge, London, 1868; Imprimerie Nationale,
Paris, 1889; Reinhold, New York, 1967) (first published 1839). This excellent book is an
example of the extent to which detailed knowledge about human vision was available long
before the advent of modern technology
3.6 Koenig and Brodhun data (1884), quoted by S. Hecht: J. Gen. Physiol. 7, 421 (1924)
3.7 R.M. Evans: The Perception of Color (Wiley, New York 1975); "Fluorescence and Gray
Content of Surface Colors," J. Opt. Soc. Am. 49, 1049-1059 (1959)
3.8 S.M. Newhall: "Preliminary Report of the D.S.A. Subcommittee on the Spacing of the Munsell
Colors", J. Opt. Soc. Am. 30, 617 (1940)
3.9 R.M. Evans: An Introduction to Color (Wiley, New York 1948);
G. Wyszecki, W.S. Stiles: Color Science (Wiley, New York 1967) pp.451-453;
[Link]. Bartleson, EJ. Breneman: "Brightness Perception in Complex Fields," J. Opt. Soc. Am.
57, 953-957 (1967)
3.10 B. Hashizume: "Companding in Image Processing," B.S. Thesis, Massachusetts Institute of
Technology, Electrical Engineering and Computer Science Department (1973)
U. Malone: "New Data on Noise Visibility," M.S. Thesis, Massachusetts Institute of Technol-
ogy, Electrical Engineering and Computer Science Department (1977)
3.11 W.F. Schreiber: "Image Processing for Quality Improvement," Proc.1EEE 66 (12),1640-1651
(1978)
3.12 D.H. Kelly: ''Theory of Flicker and Transient Responses, I: Uniform Fields," J. Opt. Soc.
Am. 61 (4), 537-546 (1971); "Adaptation Effects on Spatio-temporal Sine-wave Thresholds,"
Vision Res. 12, 89-101 (1972)

286
3.13 EM. Lowry, JJ. DePalma: "Sine Wave Response of the VISual System: I. The Mach Phe-
nomenon," 1. Opt. Soc. Am. 51 (7), 740-746 (1961); "Sine Wave Response of the Visual
System: II. Sine Wave and Square Wave Conttast Sensitivity," 1. Opt. Soc. Am. 52 (3), 328-
335 (1962)
3.14 DR. Kelly: "Theory of Flicker and Transient Response, II: Counterphase Gratings," 1. Opt.
Soc. Am. 61 (5), 632-(j4() (1971)
3.15 O.R. Mitchell: "The Effect of Spatial Frequency on the VISibility of Unstructured Spatial
Patterns," PhD. Thesis, Massachusetts Institute of Technology, Electrical Engineering and
Computer Science Department (1972)
3.16 A. Rosenfeld, A.C. Kale: Digital Picture Processing (Academic, New York 1976), Chaps.6
and 7
3.17 DR. Kelly: "Image-Processing Experiments," 1. Opt. Soc. Am. 61 (10), 1095-1101 (1971)
3.18 D.E. Troxel, W. Schreiber, C. Seitz: "Wll'ephoto Standards Converter," IEEE Trans. COM-17
(5), 544-553 (1969)
3.19 WF. Schreiber: "Wirephoto Quality Improvement by Unsharp Masking," Pattern Recognition
2, 117-121 (1970)
3.20 l,A,C. Yule: Principles of Color Reproduction (Wiley, New York 1967) p.74
3.21 RM. Evans: "Sharpness and Contrast in Projected Pictures," presented at the 1956 SMPTE
Convention, Los Angeles
3.22 DR. Kelly: Private communication
3.23 G. Sperling: "Temporal and Spatial VISUal Masking," 1. Opt. Soc. Am. 55, 541-559 (1965)
3.24 O. Braddick et al.: "Channels in Vision: Basic Aspects," in Handbook of Sensory Physiology,
Val. 8, ed. by R. Held, H.W. Leibowitz, HL. Teubner (Springer, Berlin, Heidelberg 1978)
3.25 A.N. Netravali, B. Prasada: "Adaptive Quantization of Picture Signals Using Spatial Masking,"
Proc. IEEE 6S (4), 536-548 (1977)
3.26 W F. Schreiber, R.R. Buckley: "A Two-Channel Picture Coding System: II-Adaptive Com-
panding and Color Coding," IEEE Trans. COM-29 (12), 1849-1858 (1981)
3.27 AJ. Seyler, Z. Boudrikis: "Detail Perception After Scene Changes in TV," IEEE Trans. IT-ll,
31-43 (1965)
3.28 W.E. Glenn: "Compatible Transmission of HDTV Using Bandwidth Reduction," videotape
demonstration (Natl. Assn. of Broadcasters, Las Vegas, April 12, 1983)
3.29 L.A. Riggs, F. Ratliff, I.C. Comsweet, T.N. Cornsweet: "The Disappearance of Steadily Fix-
ated Test Objects," J. Opt. Soc. Am. 43, 495-501 (1953)
3.30 R. Fielding (ed.): A Technological History of Motion Pictures and Television (University of
California, Berkeley 1967, 1983)
3.31 R.A. KinchJa, L.G. Allan: "A Theory of VISUal Movement Perception," Psychol. Rev. 76,
537-558 (1969)
3.32 J. Korein, N. Badler: "Temporal Anti-Aliasing in Computer Generated Animation," Comput.
Graph. 17 (3), 377-388 (1983)
3.33 Association for Computing Machinery, MOTION: Representation and Perception,
SIGGRAPH/SIGART Interdisciplinary Workshop, Toronto, April 4-6, 1983. See especially
A.B. Watson, AJ. Ahumada, "A Look at Motion in the Frequency Domain," pp.l-1O

Chapter 4

4.1 SJ. Mason, HJ. Zimmermann: Electronic Circuits, Signals, and Systems (Wiley, New York
1960) p.281;
A. Papoulis: Signal Analysis (McGraw-Hill, New York 1977)
4.2 BM. Oliver, J.R. Pierce, C.E. Shannon: "The Philosophy of PCM," Proc. IRE 36 (11), 1324-
1331 (1948)

287
4.3 IN. Ratzel: "The Discrete Representation of Spatially Continuous Images," Sc.D. Thesis, Mas-
sachusetts Institute of Technology, Electrical Engineering and Computer Science Department
(1980);
W.F. Schreiber, D.E. Troxel: ''Transformation Between Continuous and Discrete Representa-
tions of Images: A Perceptual Approach," IEEE Trans. PAMI·7 (2), 178-186 (1985)
4.4 H.S. Hou, H.C. Andrews: "Cubic Splines for Image Interpolation and Digital Filtering," IEEE
Trans. ASSp·26 (6), 508-517 (1978)
4.5 F.C. Harris: "On the Use of Wmdows for Harmonic Analysis with the Discrete Fourier Trans-
form," Proc. IEEE 66 (1), 51-83 (1978)
4.6 R.W. Grass: "An Image Compression/Enhancement System," M.S. Thesis, Massachusetts In-
stitute of Technology, Department of Electrical Engineering and Computer Science (1978)
4.7 C.U. Lee: "Image Rotation by I-D Filtering," M.S. Thesis, Massachusetts Institute of Tech-
nology, Department of Electrical Engineering and Computer Science (1985)
4.8 P. Roetling: "Halftone Method with Enhancement and Moire Suppression," I. Opt. Soc. Am.
66 (10), 985-989 (1976)
4.9 R. Mersereau: "Hexagonally Sampled 2-D Signals," Proc. IEEE 67 (6), 930-949 (1979)
4.10 L.G. Roberts: "Picture Coding Using Pseudorandom Noise," IRE Trans. IT·8 (2), 145-154
(1962)
4.11 N. Ziesler: "Several Binary Sequence Generators," Massachusetts Institute of Technology
Lincoln Laboratory Tech. Rep. 95 (9) (1955) "Linear Sequences," I. SIAM 7, 31-48 (1959);
S.W. Golomb: Shift Register Sequences (Aegean Park, Laguna Hills, CA 1967)
4.12 DN. Graham: "Two-Dimensional Filtering to Reduce the Effect of Quantizing Noise in Televi-
sion," M.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering Department
(1962);
R.A. Bruce: "Optimum Pre-emphasis and De-emphasis Networks for Transmission of Televi-
sion by PCM," IEEE Trans. CS·12 (9), 91-96 (1964);
E.G. Kimme, F.F. Kuo: "Synthesis of Optimal Filters for a Feedback Quantization System,"
IEEE Trans. CT·I0 (9), 405-413 (1963);
A.E. Post: B.S. Thesis, Massachusetts Institute of Technology, Department of Electrical En-
gineering (1966)
4.13 K.P. Wacks: "Design of a Real Time Facsimile Transmission System," Ph.D. Thesis, Mas-
sachusetts Institute of Technology, Electrical Engineering and Computer Science Department
(1973)

Chapter 5

5.1 P.A. Wmtz: "Transform Picture Coding," Proc. IEEE 60 (7), 809-820 (1972);
see also A.G. Tescher's article on transform coding in
WK. Pratt (00.): Image Transmission Techniques (McGraw-Hill, New York 1979)
5.2 1.0. Limb, C.R. Rubinstein, I.E. Thompson: "Digital Coding of Color Video Signals - A
Review," IEEE Trans. COM·25 (11), 1349-1385 (1977);
R.R. Buckley: "Digital Color Image Coding and the Geometry of Color Space," Ph.D. Thesis,
Massachusetts Institute of Technology, Electrical Engineering and Computer Science Depart-
ment (1982)
5.3 A.N. Netravali, ID. Robbins: "Motion-Compensated TV Coding", Bell System Tech. I. 58
(3), 631-670 (1979)
5.4 WE. Schreiber: "Color Reproduction System," US Patent No.4,500,919 (1985)
5.5 W.-H. Chen, W.K. Pratt: "Scene Adaptive Coder," IEEE Trans. COM·32 (3), 225-232 (1984)

288
5.6 C.E. Shannon, W. Weaver: The Mathematical Theory o/Communication (University of lllinois
Press, Urbana 1949). Everyone interested in statistical coding should read Shannon's landmark
paper
5.7 D.A. Huffman: "A Method for the Construction of Minimum Redundancy Codes," Proc. IRE
40 (9), 1098-1101 (1952)
5.8 WF. Schreiber: "The Measurement of Third Order Probability Distribution of TV Signals,"
IRE Trans. IT·2 (9), 94-105 (1956)
5.9 S. Ericsson, E. Dubois: "Digital Coding of High Quality TV," Paper 35, High Definition
Television Colloquium, Canadian Dept. of Communications, Ottawa (1985)
5.10 D. Bodson, S. Urban, A. Deutermann, C. Clarke: "Measurement of Data Compression in
Advanced Group IV Facsimile Systems," Proc. IEEE 73 (4), 731-739 (1985)
5.11 W.F. Schreiber: "Reproduction of Graphical Data by Facsimile," Quarterly Progress Report,
Research Laboratory of Electronics, Massachusetts Institute of Technology, No. 84 (1967);
R.B. Alps, RL. Erdmann, A.S. Neal, C.E. Schlaepfer: "Character Legibility Versus Resolution
in Image Processing of Printed Matter," IEEE Trans. MMS-I0 (3), 66-71 (1969)
5.12 RN. Ascher, G. Nagy: "A Means for Achieving a High Degree of Compaction on Scan Dig-
itized Printed Test," IEEE Trans. C-23 (11), 1174-1179 (1974);
W.K. Pratt, P. Capitant, [Link], E. Hamilton, R. Wallis: ''Combined Symbol Matching Fac-
simile Data Compression System," Proc. IEEE 68 (7), 786--795 (1980)
5.13 D.R. Knudson: US Patent No. 4,281, 312 (1981)
5.14 S.W. Golomb: "Run Length Encodings," IEEE Trans. IT·lS (7), 399-400 (1969);
1. Capon: "A Probabilistic Model for Run-Length Coding of Pictures," IRE Trans. IT -5 (12),
157-163 (1959);
C.G. Beaudette: "An Efficient Facsimile System for Weather Graphics," in Proceedings of
Symposium on Picture Bandwidth Compression, MIT, ed. by T.S. Huang, OJ. Tretiak (MIT
Press, Cambridge, MA 1969) pp.217-229;
H.G. Mussman, D. Preuss: "Comparison of Redundancy Reducing Codes for Facsimile Trans-
mission of Documents," IEEE Trans. COM·2S (11), 1425-1433 (1977);
T.S. Huang, A.B.S. Hussain: "Facsimile Coding by Skipping White," IEEE Trans. COM·23
(12), 1452-1460 (1975);
T.S. Huang: "Bounds on the Bit Rate ofUnear Run-Length Codes," IEEE Trans. IT-21 (11),
707-708 (1975);
R.B. Alps: "An Introduction and Digital Facsimile Review," in Proc. Int. Conf. on Commu-
nications Vol. 1 (San Francisco 1975) p.7. A number of other interesting papers are included
in this issue;
Kalle Infotec U.K. Ltd.: "Redundancy Reduction Technique for Fast Black and White Fac-
simile Apparatus," CCIrr Study Group XIV, Temp. Document No. 15 (1975);
R.B. Alps: "Bibliography of Digital Graphic Image Compression and Quality," IEEE Trans.
IT·20 (I), 120-122 (1974)
5.15 M.W. Baldwin: "The Subjective Sharpness of Simulated TV Pictures," Proc. IRE 28 (10),
458-468 (1940)
5.16 T.S. Huang: "Run Length Coding and its Extensions," Proceedings of Symposium on Picture
Bandwidth Compression, MIT, ed. by T.S. Huang, OJ. Tretiak (MIT Press, Cambridge, MA
1969) pp.231-264
5.17 F. DeCoulon, M. Kunt: "An Alternative to Run Length Coding for Black and White Fascimile,"
in Proceedings of IEEE International Zurich Seminar on Digital Communication, Paper C-4
(1974);
F. DeCoulon, O. 1ohnsen: "Adaptive Block Scheme for Source Coding of Black and White
Facsimile," Electron. Lett. 12 (3), 61 (1976)
5.18 R.M. Gray: "Vector Quantization," IEEE ASSP Mag. 4-29 (April 1984)

289
5.19 W.F. Schreiber, T.S. Huang, OJ. Tretiak: "Contour Coding of Images," Wescon Teclmical
Papers, paper 813, IEEE (1968); also in Picture Bandwidth Compression (Gordon Breach,
New York 1972); and in N.S. Jayant (ed.): Waveform QUJmIization and Coding (IEEE. New
York 1976);
D. Gabor, [Link]. Hill: "Television Band Compression by Contour Interpolation," Proc. lEE
(England) 108, Pan B, 303--315 (1961);
T.H. Morrin: "Recursive Contour Coding of Nested Objects in Black/White Images," in Proc.
Int. Conf. on Communications, Vol. 1 (San Francisco, 1975) pp.7-17
5.20 [Link]. Coueignoux: "Compression of Type Faces by Contour Coding," M.S. Thesis, Mas-
sachusetts Institute of Technology, Electrical Engineering and Computer Science Department
(1973)
5.21 DN. Graham: "Image Transmission by Two-Dimensional Contour Coding," Ph.D. Thesis,
Massachusetts Institute of Technology, Electrical Engineering Department (1966); also Proc.
IEEE SS (3), 336-346 (1967)
5.22 J.W. Pan: "Reduction of Information Redundancy," SeD. Thesis, Massachusetts Institute of
Technology, Electrical Engineering Department (1963)
5.23 H.S. Hou, H.C. Andrews: ''Cubic Splines for Image Interpolation and Digital Filtering," IEEE
Trans. ASSP-26 (4), 508 (1978)
5.24 R. Hunter, A.B. Robinson: "International Digital Facsimile Coding Standards," Proc. IEEE 68
(7), 854--867 (1980)
5.25 J.R. Ellis: PhD. Thesis, Massachusetts Institute of Technology, Electrical Engineering and
Computer Science Department (1977);
D. Spencer: M.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and
Computer Science Department (1968)
5.26 H. Freeman: "On the Encoding of Arbitrary Geometrical Figures," IRE Trans. EC·IO, 260-
268 (1961);
I.T. Young: "TV Bandwidth Compression Using Area Properties," MS Thesis, Massachusetts
Institute of Technology, Electrical Engineering Department (1966)
5.27 P. Elias: "Predictive Coding," IRE Trans. IT·I (3), 16-32 (1955);
R.E. Graham: "Predictive Quantization of TV Signals," Wescon Conv. Record 2 (4), 147-157,
IRE (1958);
W. Zschunke: "DPCM Picture Coding with Adaptive Prediction," IEEE Trans. COM·2S (11),
1295-1302 (1977);
C.M. Harrison: "Experiments with Linear Prediction in TV," Bell System Tech. J. 31, 764-783
(1952)
5.28 LD. Davisson: "Data Compression Using Straight-Line Interpolation," IEEE Trans. IT·14 (3),
390-394 (1968);
AN. Netravali: "Interpolative Picture Coding Using a Subjective Criterion," IEEE Trans.
COM-2S (5), 503-508 (1977)
5.29 C.C. Cutler: US Patent No.2,605,361 (1952);
J.B. O'Neal: "Predictive Quantizing System (DPCM) for the Transmission of TV Signals,"
Bell System Tech. J. 4S, 689-721 (1966)
5.30 J. de Jaeger: "Deltamodulation, a Method of PCM Transmission Using a I-bit Code," Philips
Res. Rep. 7 (6), 442-466 (1952)
5.31 J. Max: "Quantizing for Minimum Distortion," IRE Trans. IT·6 (3), 7-12 (1960);
AN. Netravali: "On Quantizers for DPCM Coding of TV Signals," IEEE Trans. IT-ll (3),
360-370 (1977);
T. Berger: ''Optimum Quantizers and Permutation Codes," IEEE Trans. IT-18 (11), 759-776
(1972)
5.32 H.G. Mussman: "Predictive Coding of TV Signals," in Image Transmission Techniques, ed.
by W. Pratt (Academic, New York 1979)

290
5.33 D.K. Shanna, A.N. Netravali: "Design of Quantizers for DPCM Coding of Picture Signals,"
IEEE Trans. COM-2S (11), 1267-1274 (1978)
5.34 B. Prasada: Private communication
5.35 A. Habibi: "Comparison of nth Order DPCM Encoder with Linear Transformations and Block
Quantization Techniques," IEEE Trans. COM-19 (6), 948-956 (1971);
DJ. Connor, R.F.W. Pease, W.G. Scholes: "TV Coding Using Two-Dimensional Spatial Pre-
diction," Bell System Tech. J. SO, 1049-1061 (1971)
5.36 P.A. Ratliff: "Digital Coding of the Composite PAL Colour Television System for Transmission
at 34 Mbits/sec," BBC Eng. No. 115, 24-35 (1980)
5.37 A. Habibi: "Survey of Adaptive Image Coding Techniques," IEEE Trans. COM-2S (11), 1257-
1284 (1977);
B. Prasada, A. Netravali: "Adaptive Companding of Picture Signals in a Predictive Coder,"
IEEE Trans. COM-26 (1), 161-164 (1978)
5.38 N.S. Jayant: "Adaptive Quantization with a One-Word Memory," Bell System Tech. J. 52,
1119-1143 (1973)
5.39 A.P. Zarembowitch: "Forward Estimation Adaptive DPCM for Image Data Compression,"
M.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and Computer
Science Department (1981)
5.40 C.C. Cutler: "Delayed Encoding: Stabilizer for Adaptive Coders," IEEE Trans. COM-19 (6),
898-907 (1971)
5.41 A.N. Netravali, B. Prasada: "Adaptive Quantization of Picture Signals Using Spatial Masking,"
Proc. IEEE 6S (4), 536-548 (1977)
5.42 S.K. Goyal, J.B. O'Neal: "Entropy Coded DPCM systems for TV," IEEE Trans. COM-23
(6), 66()..{)6(j (1975)
5.43 E.G. Kimrne, F.F. Kuo: "Synthesis of Optimum Fiters for a Feedback Quantization Scheme,"
IEEE Trans. CT-I0 (9), 405-413 (1963)
5.44 W.F. Schreiber, D.E. Troxel: US Patent No.4, 268,861 (1981)
5.45 D.E. Troxel: "Application of Pseudorandom Noise to DPCM," IEEE Trans. COM-29 (12),
1763-1167 (1981)
5.46 E.R. Kretzmer: "Reduced Alphabet Representation of Television Signals," IRE Nat. Conv.
Rec.4 (4), 140-153 (1956)
5.47 W.F. Schreiber, C.F. Knapp, N.D. Kay: "Synthetic Highs, An Experimental TV Bandwidth
Reduction System," J. Soc. Motion Picture Television Eng. 68 (8), 525-537 (1959);
D.N. Graham, "Image Transmission by Two Dimensional Contour Coding," Proc. IEEE 55
(3), 336-346 (1967)
5.48 D.E. Troxel, W. Schreiber, P. Curlander, A. Gilkes, R. Grass, G. Hoover: "Image Enhance-
ment/Coding Systems Using Pseudorandom Noise Processing," Proc. IEEE 67, 972-973
(1979);
D.E. Troxel, W.F. Schreiber, R. Grass, G. Hoover, R. Sharpe: "Bandwidth Compression of
High Quality Images," International Conference on Communications, 31.9.1-5, Seattle (1980);
also "A Two-Channel Picture Coding System: I - Real-Time Implementation," IEEE Trans.
COM-29 (12), 1841-1848 (1981)
5.49 W.F. Schreiber, R.R. Buckley: "A Two-Channel Picture Coding System: II - Adaptive Com-
panding and Color Coding," IEEE Trans. COM-29 (12), 1849-1858 (1981)
5.50 R.B. Sharpe: "Statistical Coding of the Highs Channel of a Two-Channel Facsimile System,"
M.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and Computer
Science Department (1979)
5.51 M. Kocher, M. Kunt: "A Contour-Texture Approach to Picture Coding," in Proc. ICASSP-2,
Paris (IEEE, New York 1982) pp.436-440
5.52 P J. Bun, E.H. Adelson: "The Laplacian Pyramid as a Compact Image Code," IEEE Trans.
COM-31 (4), 532-540 (1983)
5.53 J. Alnatt: Transmitted-picture Assessment (Wiley, New York 1983)

291
Chapter 6

6.1 HJ.P. Arnold: William Henry Fox Talbot (Hutchinson Benham, London 1977)
6.2 R. Higonnet, L. Moyroud: US Patent No.3, 188,929 (1965)
6.3 C. Bigelow: "Technology and the Aesthetics of Type," Seybold Repon on Publishing Systems
10 (24) (1981); "The Principles of Digital Type," ibid. 11 (11) (1982)
6.4 D.E. Troxel, W. Schreiber, S. Goldwasser, M. Khan, L. Picard, [Link], C. Turcio: "Automated
Engraving of Gravure Cylinders," IEEE Trans. SMC·l1 (9), 585-596 (1981)
6.5 ''Lasergravure'' (product literature of Crosfield Electronics)
6.6 J.A.C. 'fule, D.S. Howe, J.H. Alunan: "The Effect of the Spread·Function of Paper on Halftone
Reproduction," TAPPI SO (7), 337-344 (1967)
6.7 W.P. Schreiber: "An Electronic Process Camera," in Proceedings, Tech. Assn. of the Graphic
Arts (1983) pp. 101-127
6.8 Yao-ming Chao: "An Investigation into the Coding of Halftone Pictures," Ph.D. Thesis, Mas·
sachusetts Institute of Technology, Electrical Engineering and Computer Science Depanment
(1982)
6.9 C.N. Nelson: "Tone Reproduction", in Theory of the Photographic Process, ed. by T.H. Jones
(Macmillan, New York 1977) pp.536-560
6.10 J. Sturge (ed.): Neblette's Handbook of Photography and Reprography (Van Nostrand Reinhold,
New York 1977). (This is an excellent reference for many of the subjects of this chapter. It
contains a number of review anicles with good bibliographies.)
6.11 B.E. Bayer: "An Optimum Method for Two-Level Rendition of Continuous·Tone Pictures,"
ICC 73, IEEE I, 26.11-26.15
6.12 J.P. Jarvis: "A Survey of Techniques for the Display of Continuous Tone Pictures by Bilevel
Displays," Comput. Graph. Image Proc. 5, 13-40 (1976)
6.13 R.W. Floyd, L. Steinberg: "An Adaptive Algorithm for Spatial Grayscale," Proc. Soc. for
Information Display 17 (2), 75-77 (1976)
6.14 B. Woo: "A Survey of Halftoning Algorithms and Investigation of the Error Diffusion Tech·
nique," B.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and Com-
puter Science Department (1984)

Chapter 7

7.1 R.M. Evans: The Perception of Color (Wiley, New York 1975)
7.2 W.T. Wmtringham: "Color TV and Colorimetry," Proc. IRE 39 (10), 1135 (1951)
7.3 D.L. MacAdam: "VIsual Sensitivities to Color Differences in Daylight," J. Opt. Soc. Am. 32
(5), 247-274 (1942)
7.4 W.F. Schreiber: "A Color Prepress System Using Appearance Variables," J. lmag. Tech. 12
(4), 200-210 (1986)
7.5 M. Baldwin: "Subjective Sharpness of Additive Color Pictures," Proc. IRE 39 (10),1173-1176
(1951)
7.6 U.F. Gronemann: "Coding Color Pictures," PhD. Thesis, Massachusetts Institute of Technol·
ogy, Electrical Engineering Depanment (1964)
7.7 A.V. Bedford: "Mixed Highs in Color Television," Proc. IRE 38 (9), 1003-1009 (1950)
7.8 R.R. Buckley: "Digital Color Image Coding and the Geometry of Color space," PhD. Thesis,
Massachusetts Institute of Technology, Electrical Engineering and Computer Science Depan.
ment (1981)
7.9 Second Color Television Issue, Proc. IRE 42 (1) (1954) (entire issue)
7.10 E. Dubois, W.P. Schreiber: "Improvements to NTSC by Multidimensional Filtering," J.
SMPTE 97, 446-463 (1988)

292
7.11 N.M. Nasrabadi, R.A. King: "Image Coding Using Vector Coding: A Review," IEEE Trans.
COM-36 (8), 957-971 (1988)
7.12 W. Bender: "Adaptive Color Coding Based on Spatial/femporal Features," SPSE Electronic
Imaging Devices and Systems Symposium, Los Angeles (1988)

Chapter 8

8.1 W.F. Schreiber: "Psychophysics and the Improvement of Television Image Quality," J. SMPTE
93 (2), 717-725 (1984)
8.2 E.F. Brown: "Low-Resolution TV: Subjective Comparison of Interlaced and Non-Interlaced
Pictures," Bell System Tech. J. 46 (1), 199-232 (1967);
T. Fujio (ed.): "HDTV," NHK Tech. Monograph 32 (1982)
8.3 B. Wendland: ''Extended Definition TV with High Picture Quality," J. SMPTE 92 (10), 1028-
1035 (1983)
8.4 T.G. Schut: "Resolution Measurements in Camera Tubes," J. SMPTE 92 (12), 1270-1293
(1983)
8.5 S.C. Hsu: "The Kell Factor: Past and Present," J. SMPTE 95 (2), 206-214 (1986)
8.6 D.H. Kelly: "Visual Response to Time-Dependent Stimuli," J. Opt. Soc. Am. 51 (4),422-420
(1961)
8.7 T.N. Comsweet: VISual Perception (Academic, New York 1970)
8.8 E.M. Lowry, JJ. DePalma: "Sine Wave Response of the Visual System," J. Opt. Soc. Am. 51
(10), 474 (1961)
8.9 A.N. Netravali, B. Prasada: "Adaptive Quantization of Picture Signals Using Spatial Masking,"
Proc. IEEE 6S (4), 536-548 (1977)
8.10 W.F. Schreiber, R.R. Buckley: "A Two-Channel Picture Coding System: II - Adaptive Com-
panding and Color Coding," IEEE Trans. COM-29, 12 (1981)
8.11 L.G. Roberts: "Picture Coding Using Pseudo-Random Noise," IRE Trans. IT-8 (2), 145-154
(1962)
8.12 D.G. Fink: IT Standards and Practice (McGraw-Hill, New York 1943)
8.13 Proc. IRE 39 (10) (1951) and 42 (1) (1954) (entire issues)
8.14 U.F. Gronemann: Ph.D. Thesis, Massachusetts Institute of Technology, Electrical Engineering
Department (1964)
8.15 A.V. Bedford: "Mixed Highs in Color Television," Proc. IRE 38 (9), 1003 (1950)
8.16 F. Gray: "Electro-Optical Transmission System," US Patent No.1, 769,920 (1929)
8.17 E. Dubois, W.F. Schreiber: "Improvements to NTSC by Multidimensional Filtering," J.
SMPTE 97 (6), 446-463 (1988)
8.18 M.W. Baldwin: "The Subjective Sharpness of Simulated Television Pictures," Proc. IRE 18
(10), 458-468 (1940)
8.19 Y. Faroudja, J. Roizen: "Improving NTSC to Get Near-RGB Performance," J. SMPTE 96 (8),
750-761 (1987)
8.20 "Hierarchical High Definition Television System Compatible with the NTSC Environment,"
North American Philips Corp., Briarcliff Manor, NY (1987)
8.21 W.E. Glenn, K.G. Glenn: "HDTV Compatible Transmission System," J. SMPTE 96 (3), 242-
246 (1987)
8.22 W.R. Neuman: MIT Audience Research Facility, private communication
8.23 J.S. Wang: "Motion-Compensated NTSC Demodulation," Ph.D. Thesis, Massachusetts Institute
of Technology, Electrical Engineering and Computer Science Department (1989)
8.24 Y. Ninomiya, Y. Ohtsuka: "A Single Channel HDTV Broadcast System, - The MUSE," NHK
Lab. Note 304, 1-12 (1984)
8.25 W.R. Neuman: MIT Audience Research Facility, private communication

293
8.26 T. Fukinuki, H. Hirano: "Extended Definition TV Fully Compatible with Existing Standards,"
IEEE Trans. COM-32 (8), 948-953 (1984)
8.27 M. Isnardi, J. Fuhrer, T. Smith, J. Koslov, B. Roeder, W. Wedam: "A Single-Channel Compat-
ible Wide screen EDTV System," 3rd HDTV Colloquium, (Canadian Dept. of CommUnication,
Ottawa 1987)
8.28 Y. Yasumoto, S. Kageyama, S. Inouye, H. Uwabata, Y. Abe: "An Extended Definition Televi-
sion System Using Quadrature Modulation of the Picture Carrier with Inverse Nyquist Filter,"
Consumer Electronics Conference, Chicago (1987)
8.29 J.M. Barstow, H.N. Christopher: "The Measurement of Random Video Interference to
Monochrome and Color Television Pictures," Trans. AlEE (Communication and Electronics)
81 (1), 313-320 (1962)
8.30 W.F. Schreiber et al.: "A Compatible High-Definition Television System Using the Noise-
Margin Method of Hiding Enhancement Information," J. SMPTE 98 (12), 873-879 (1989)
8.31 W J. Butera: "Multiscale Coding of Images," B.S. Thesis, Massachusetts Institute of Technol-
ogy, Media Laboratory (1988)
8.32 W.F. Schreiber: "Improved Television Systems: NTSC and Beyond," J. SMPTE 86 (8),734-
744 (1987)
8.33 W.F. Schreiber, H. Lippman, A. Netravali, E. Adelson, D. Staelin: "Channel-Compatible 6-
MHz HDTV Distribution Systems," J. SMPTE 98 (1), 5-13 (1989)
8.34 A. Toth, J. Donahue: "ATV Multiport Receiver - Prelintinary Analysis," EIA Multiport Re-
ceiver Subcommittee (1989)
8.35 A. Hirota, S. Hirano, H. Kitamura, T. Tsushima: "Noise Reducing System for Video Signal,"
US Patents No.4,607,285 (1986);
A. Hirota, T. Tsushirna: "Noise Reducing System for Video Signal," US Patent No.4, 618,893
(1986);
L. Pham Van Cang: "Circuit for Processing a Color Television Signal," US Patent No.4, 007,
483 (1977)
8.36 P.P. Vaidyanathan: "Quadrature Mirror Filter Banks, M-Band Extensions, and Perfect-Recon-
struction Techniques," IEEE ASSP Mag. 4-20 (July 1987)
8.37 E.R. Kretzmer: "Reduced-Alphabet Representation of Video Signals," IRE Convention Record,
Part 4, 140--146 (1956)
8.38 W.F. Schreiber, A.B. Lippman: "Reliable EDTV/HDTV Transmission in Low-Quality Analog
Channels," J. SMPTE 98 (7), 496-503 (1989)
8.39 Spectrum Compatible HDTV System (Zenith Electronics Corp., Glenview, IL 1988)
8.40 W. Luplow: Zenith Electronics Corp., private communication
8.41 B.G. Schunck: "The Image Flow Constraint Equation," Compo Vision, Graph. Image Proc. 35,
20-46 (1986)
8.42 E. Dubois, S. Sabri: "Noise Reduction in Image Sequences Using Motion-Compensated Tem-
poral Filtering," IEEE Trans. COM-73 (7), 502-522 (1984);
D. Maninez, I.S. Lim: "Implicit Motion-Compensated Noise Reduction of Motion Picture
Scenes," Proc. Inti. Conf. on Acoustics, Speech, and Signal Processing (1986) pp.375-378
8.43 E.A. Krause: "Motion Estimation for Frame-Rate Conversion," Ph.D. Thesis, Massachusetts
Institute of Technology, Electrical Engineering and Computer Science Department (1987);
D.M. Martinez: "Model-Based Motion Estimation and its Application to Restoration and Inter-
polation of Motion Pictures," Ph.D. Thesis, Massachusetts Institute of Technology, Electrical
Engineering and Computer Science Department (1986)
8.44 M.A. Krasner: "The Critical Band Coder - Digital Encoding of Speech Signals Based on the
Perceptual Requirements of the Auditory System," ICASSP (IEEE, New York 1980) pp.327-
331
8.45 R.E. Crochiere et al.: "Digital Coding of Speech in Subbands," Bell System Tech. J. 55 (10),
1069-1085 (1976)

294
8.46 N. Ziesler: "Several Binary Sequence Generators," Massachusetts Institute of Technology
Lincoln Laboratory Tech. Rep. 95 (9) (1955); "Linear Sequences," I. SIAM 7,31-48 (1959);
S.W. Golomb: Shift Register Sequences (Aegean Park, Laguna Hills, CA 1967)
8.47 A.B. Watson: "The Window of Visibility: A Psychological Theory of Fidelity in Time~Sampled
Visual Motion Displays," NASA Tech. Paper 221 I (1983); "A Look at Motion in the Frequency
Domain," MOTION: Representation and Perception, SIGGRAPH/SIGGART Workshop, As-
sociation for Computing Machinery, Toronto 1-10 (1983)
8.48 E. Chalom: "Video Data Compression Using Quadrature Mirror Filters and Adaptive Coding,"
M.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and Computer
Science Department (1989);
W.P. Schreiber et aI.: "Robust Bandwidth-Efficient HDTV Transmission Formats," I. SMPTE
Conference, Los Angeles (1989)
8.49 C.E. Shannon, W. Weaver: The Mathematical Theory of Communication (University of llIinois
Press, Urbana 1949)
8.50 W.F. Schreiber: "6-MHz Single-Channel HDTV Systems," HDTV Colloquium, Canadian
Dept. of Communications, Ottawa (1987)
8.51 W.P. Schreiber: "The Role of Technology in the Future of Television," Telecommun. I. 57
(11),763-774 (1990)
8.52 W J. Butera: "Multiscale Coding of Images," M.S. Thesis, Massachusetts Institute of Tech-
nology, Media Laboratory (1988)
8.53 CL. Ruthroff: "Computation of FM Distortion in Linear Networks for Bandlirnited Periodic
Siguals," Bell Syst. Tech. I. 47 (6), 1043-1063 (1968)
8.54 W.F. Schreiber, I. Piot: "Video Transmission by Adaptive Frequency Modulation," IEEE
Commun. Mag. MCOM 26 (11), 68-76 (1988)

Appendix

A.l R. Forni, M. Pelchat: Deputies, "La Television a Haute Definition," Rapport de l'Office par-
lementaire d' evaluation des choix scientifiques et technologiques, Assemblee Nationale-Senat
(Ed. Economica, Paris 1989)
A.2 Y. Ninomiya, Y. Ohtsuka: "A Single Channel HDTV broadcast System, - The MUSE," NHK
Lab. Note 304, 1-12 (1984)
A.3 T. Fujio et al.: "High-Definition Television," NKH Technical Monograph 32 (1982)
A.4 T. Fukinuki, H. Hirano: "Extended Definition TV Fully Compatible with Existing Standards,"
IEEE Trans. COM-32 (8), 948-953 (1984)
A.5 T. Nishizawa et al.: "HDTV and ADTV Transmission Systems: MUSE and its Family,"
Unpublished oral presentation, Nat!. Assn. of Broadcasters, Las Vegas, April 11 (1988);
Y. Tanaka, K. Enarni, H. Okuda: "Compatible MUSE Systems," NHK Laboratories Note 375
(January 1990)
A.6 M.D. Windram, GJ. Tonge, R.C. Hills: ''The D-MAC/Packet Transmission System for Satellite
Broadcasting in the United Kingdom," EBU Tech. Rev., No. 227 (1988)
A.7 RJ. Iredale: "A Proposal for a New High Definition NTSC Broadcast Protocol," IEEE Trans.
CE·33 (1), 14-27 (1987)
A.8 E. Chalom: B.S. Thesis, Massachusetts Institute of Technology, Electrical Engineering and
Computer Science Department (1987)
A.9 F.W.P. Vreeswijk, M.R. Haghiri: "HDMAC Coding for MAC Compatible Broadcasting of
HDTV Signals," Third Int!. Workshop on HDTV, Torino, Sept. 1989
A.1O R. Storey: "Motion-Compensated DATV Bandwidth Compression for HDTV," IBC-88, Brigh-
ton, England (1988)
A.11 M.L. Dertouws, R.K. Lester, R.M. Solow: Made in America (MIT Press, Cambridge, MA
1989)

295
A.12 M. Isnardi et al.: "A Single-Channel Compatible Widescreen EDTV System," 3rd HDTV
Colloquium, Canadian Dept. of Communication, Ottawa (1987)
A.13 "Hierarchical High Definition Television System Compatible with the NTSC Environment,"
North American Philips Corp., Briarcliff Manor, NY (1987)
A.14 W.F. Schreiber, A.B. Lippman: "Reliable EDTV/HDTV Transmission in Low-Quality Analog
Channels," J. SMPTE 98 (7), 496--503 (1989)
A.15 Spectrum Compatible HDTV System (Zenith Electronic Corp., Glenview, II.. 1988)
A.16 Y. Faroudja, J. Roizen: "Improving NTSC to get Near-RGB Performance," J. SMPTE 96,
750-761 (1987)
A.17 T M. Cover: "Broadcast Channels," IEEE Trans. IT-18 (I), 2-14 (1972)

296
Bibliography

Chapter 2

C.E.K. Mees, T.H. James: Theory of the Photographic Process (Macmillan, New York 1977)
A. Rose: VISion: Human and Electronic (Plenum, New York 1973)
W. Thomas (ed.): Handbook of Photographic Science and Engineering, SPSE (Wiley-Interscience,
New York 1973)

Chapter 5

H.S. Hou: Digital Document Processing (Wiley, New York 1983)

T.S. Huang, OJ. Tretiak: Picture Bandwidth Compression (Gordon and Breach, New York 1972).
Proceedings of a conference held at MIT in April 1969. A good collection of papers on the then
state of the art
T.S. Huang, W. Schreiber, O. Tretiak: "Image Processing," Proc. IEEE 59 (11), 1586-1608 (1971).
Extensive references
International Conference on Communication, ICC-79, Vols.l,2, Boston (IEEE, New York 1979).
This is a yearly conference, and subsequent issues contain many useful articles on image coding
N.S. Jayant (ed.): Waveform Quantization and Coding (IEEE, New York 1976)
A.N. Netravali, B.G. Haskell: Digital Pictures, Representation and Compression (Plenum, New York
1988)
A.N. Netravali, J.O. Limb: "Picture Coding, a Review," Proc. IEEE 68 (3), 366-406 (1980). This
paper contains an excellent bibliography
W.K. Pratt: Digital Image Processing (Wiley, New York 1978). Extensive references
W.K. Pratt (ed.): Image Transmission Techniques (Academic, New York 1979). This book contains
a very perceptive article on transform coding by A.G. Tescher
W.F. Schreiber: "Picture Coding," Proc. IEEE 55, 320 (1967)
Special Issues:
Special Issue on Redundancy Reduction, Proc. IEEE 55 (3) (1967)
Special Issue on Digital Communications, IEEE Trans. COM-19 (6), Part I (1971)
Special Issue on Digital Picture Processing, Proc. IEEE 60 (7) (1972)
Special Issue on Two-Dimensional Signal Processing, IEEE Trans. C-21 (7) (1972)
Special Issue on Digital Image Processing, IEEE Trans. C-7 (5) (1974)
Special Issue on Digital Signal Processing, Proc. IEEE 63 (4) (1975)
Special Issue on Image Bandwidth Compression, IEEE Trans. COM-25 (II) (1977)
Special Issue on Digital Encoding of Graphics, Proc. IEEE 68 (7) (1980)
Special Issue on Picture Communication Systems, IEEE Trans. COM·29 (12) (1981)
Special Issue on Visual Communication Systems, Proc. IEEE 73 (4) (1985)
I.C. Stoffel (ed.): Graphical and Binary Image Processing and Applications (Anech House, London
1982). A good collection of articles

297
Chapter 6

W.H. Banks (ed.): Halftone Printing (pergamon, New York 1964)

Eastman Kodak Co.: "Kodak Handbook of Newspaper Technique," Q165. (There are many other
useful Kodak Data Books in the Q Series. Some are available in large camera stores.)
RL. Gregory: The Intelligent Eye (McGraw-Hill, New York 1970)
R.W.G. Hunt: The Reproduction of Color (Fountain, Tolworth, England 1987)
W.M. Ivins, Jr.: Notes on Prints (MIT Press, Metropolitan Museum of Art, New York 1930)
W.M. Ivins, Jr.: How Prints Look (Beacon, Boston 1958)
W M. Ivins, Jr.: Prints and VISual Communication (MIT Press, Cambridge, MA 1965)
Journal of the Optical Society of America (a continuing series)
C.E.K. Mees, T.H. James: The Theory of the Photographic Proces (Macmillan, New York 1977)
A.H. Phillips: Computer Peripherals and Typesetting (Her Majesty's Stationery Office, London 1970)
Pocket Pal: International Paper Co. (Available in many graphic arts and artists' supply stores)
Proceedings, Technical Association of the Graphic Arts (Rochester, New York) (a continuing series)
TAPPI: The Journal of the Technical Association of the Pulp and Paper Industry (a continuing series)
J.A.C. Yule: Principles of Color Reproduction (Wiley, New York 1967). (Excellent references)

Chapter 7

F.W. Billmeyer, Jr.: "Survey of Color Order Systems," Color Res. Appl. 12 (4), 173-186 (1987)
Color as Seen and Photographed, Kodak Publication E-74; Kodak Filters for Scientific and Technical
Uses, Kodak Publication B-3; there are many other valuable Kodak publications in this field
Committee on Colorimetry, The Science of Color (Optical Society of America, Woodbury, NY 1963)
R.M. Evans: An Introduction to Color (Wiley, New York 1948)
R.M. Evans, W.T. Hanson, W L. Brewer: Principles of Color Photography (Wiley, New York 1953)
A.C. Hardy: Handbook of Colorimetry (MIT Press, Cambridge, MA 1936)
R.W.G. Hunt: The Reproduction of Colour in Photography, Printing, and Television (Fountain, Tol-
worth, England 1988) (distributed by Van Nostrand-Reinhold, NY)
R.S. Hunter: The Measurement of Appearance (Wiley, New York 1975)
Fl. In der Smitten: "Data-Reducing Source Encoding of Color Picture Signals Based on Optical
Chromaticity Classes," Nachrichtentech. Z. 27, 176-181 (1974)
D.B. Judd: "Basic Correlates of the Visual Stimulus," in Handbook of Experimental Psychology, ed.
by S.S. Stevens (Wiley, New York 1951) p.811
D.B. Judd, G.S. Wyszecki: Color in Business, Science, and Industry (Wiley-Interscience, New York
1975)
J. Limb, C. Rubinstein, J. Thompson: "Digital Coding of Color Video Signals - A Review," IEEE
Trans. COM·2S (11), 1349-1385 (1977)
D.L. MacAdam: Sources of Color Science (MIT Press, Cambridge, MA 1970)
D.L. MacAdam: Color Measurement (Springer, Berlin, Heidelberg 1985)
S. Newhall, D. Nickerson, D.B. Judd: "Final Report of the O.S.A. Subcommittee on the Spacing of
the Munsell Colors," J. Opt. Soc. Am. 33 (7), 385 (1943)
M. Southworth: Pocket Guide to Color Reproduction (Graphic Arts, Livonia, NY 1979)
M. Southworth: Color Separation Technique (Graphic Arts, Livonia, NY 1979)
L. Stenger: "Quantization of TV Chrominance Signals Considering the Visibility of Small Color
Differences," IEEE Trans. COM-2S (11) 1393-1406 (1977)
G. Wyszecki, W.S. Stiles: Color Science (Wiley, New York 1967)
J.A.C. Yule: Principles of Color Reproduction (Wiley, New York 1967)

298
Chapter 8

W.F. Schreiber: "Improved Television Systems: NTSC and Beyond," I. SMPTE 93 (8), 734-744
(1984)
W.F. Schreiber: "Advanced Television Systems for the United States: Getting There from Here," I.
SMPTE 97 (10), 847-851 (1988)

Appendix

I. Freeman: "A Cross-Referenced, Comprehensive Bibliography on High Definition and Advanced

Television Systems, 1971-1988," I. SMPTE 99,909-933 (1990)

299
Subject Index

Aberrations 28 - Shannon/Fano 106-109

Adaptation, visual 58 - sink 8
Aliasing - source 8
- spatial 40, 91, 92 - spline fitting 116
- temporal 14,79-83 - statistical 106-113
Aperture 28 - synthetic highs 128, 129
Aperture effect 36 - television 105,240-260
Astigmatism 28 - transform coding 105
- two-channel system 128-132
- vector 202
Bandwidth 72, 89
- waveform coding 121-132
Black body 20
Color, Chap. 7
- additive mixing 169
Camera 6, 12,48 - additive reproduction 182-184
- electronic process camera 145 - after-images 206
- process camera 145 - amplitude resolution 199,200
Cathode-ray tube, see CRT - analog representation 200-202
Channel capacity 12 - appearance 168,197
Coding 8,9, Chap. 5 - black printer 193-195
- block 108 - chromaticity coordinates 173, 174
- block artifacts 125 - CIE primaries 176-178
- buffering 106 - coding 198-203
- camera, effect of 126 - color mixture curves 171,183
- channel 9 - colorimeter 169
- color 105,198-203, Chap. 7 - colorimetry 170-184
- continuous-tone 120-131 - correction 186-192
- contour 116, 128 - correction by computer 191,192
- contour-texture 132 - digital representation 198
- delta modulation 122 - editing 192,196-198
- DPCM 121-128 - graphical representation 171-180
- DPCM/pcM comparison 126-128 - gray balance 186
- entropy 106-113 - Grassman's laws 169
- facsimile 117, 118 - masking 188
- graphics 113-118 - matching 169
- Huffman 119,129 - L*a*b* space 182
- Laplacian pyramid 132 - lookup table 191,194,195
- lossless 106-118 - MacAdam ellipse 179
- noise, effect of 110 - Maxwell triangle 172, 173
- performance 116-118 - mixed highs 199
- predictive coding 120-128 - Munsell space 181,182
- READ code 119 - nonprintable colors 193
- run-length 114, 115 - NTSC system 200,201
- sampling density, effect of 110,112 - opponent colors 196

301
- perception 168,203,204 Flicker 14-17,78,79
- primaries 173, 205 Focal length 25-28
- proofs 188 Fourier optics 30-35
- purity 179 Fourier series 31,36, 85-89
- reproduction with inks 189-194 Fourier transform 34
- saturation 179, 197
- scanners 189 Gamma correction 18
- simultaneous contrast 205, 206 Gibbs phenomenon 33, 88
- spaces 179-182 Graphic arts, Chap. 6
- spatial resolution 198,199 Gravure 135,143-145
- spectral composition 205
- subtractive mixing 169
Halftone process 136,145-167
- subtractive reproduction 184-195
- electronic screens 154-167
- temperature 19-22
Huygen's principle 28
- tristimulus values 169-172
- UCR 194
- Uniform Chromaticity Scale 180 lllumination 6
- vector coding 202, 203 llluminance 24
- vectorscope 201,202 lmage(s)
Companding 69-71,104 - analog transmission 17
Compression, see Coding - approximation 3
Cone cells 58 - continuous representation, Chap. 4
Contrast 64-69 - descriptions 2
Correlation function 40-45 - digital representation, Chap. 4
Cosine4 law 27 - enhancement 4
CRT 10, 12, 139 - processing 1-10
- quality 5,6, 10, 12,79-84
Data compression, see Coding - reconstruction90-92
Decoding 9,10 - transmission 4
Decoding channel 9 Information capacity 9,106-108
Differential PCM, see DPCM Interlace 13, 39
Diffraction 28, 29 Interpolation, Chap. 4
Displays 10 - spatial 90-93
Dot, see Halftone - temporal 14
DPCM 121-128 Intersymbol interference 91
Dynamic range 1O,65--{j9
Jones diagram 161
Edge rendition 69-73
Electron multiplication 55-57 Kell factor 17
Encoding, see Coding
Enhancement 73-75 Letterpress 140-142
Entropy, see Information capacity Light sources 19-22
Error correction 9 Lithography 142,143
Eye movements 78 Luminance 24,25
Eye tracking 80-83 Luminosity curve 23,59,60

//number 29 Mach bands 72,73

Filter(ing) 8,81, Chap.4 Masking 77, 78
- adaptive 82 Mixed highs 199
- Gaussian 90-93 Modulation transfer function 30-45,71
- ideal 91 Moire patterns 89,93
- presampling 93 Motion rendition 14-17,79-83
Fixation 14,80 Munsell system 67

302
Noise Roberts' method 98-101,126,129
- amplifier noise 48-52 Rod cells 59,60
- Johnson noise 48
- operating spot noise figure 51 Sampling, Chap. 4
- pseudorandom 98 - hexagonal 95-97
- quantization 14,97-103 - spatial 17
- randomization of quantization 98-101 - temporal 13,63-79
- shot noise 46,53 - theorem 13,18,85-89
- spectrum 50 Scanning 8
- visibility 5,18,69,87,102 - standards 12
NTSC 39,45, see also Television, Color Scattering 21
Nyquist rate 123, 126 Screen, see Halftone
Sharpness 5,73-79
Offset printing, see Lithography Shot noise 46
Optical character recognition 113,140 Signal-to-noise ratio, see SNR
Optical systems 6,25-30 SNR 5,6,8,18,46,55,69
Optical transfer function 30 Spatial frequency response 5, 8,71
Spatial transient response 71
Page composition 140 Spectrum
Pattern recognition 4 - of images 31-36,79-83
Perception 5, Chap. 3 - two-dimensional 89,90
- adaptation 68 - visible spectrum 19-23
- color 203, 204, Chap. 7 Stabilized retinal image 78
- contrast sensitivity 61-69 Stefan-Boltzmann law 19
- dynamic range 66 Stroboscopic effect 16,79-83
- lightness 67-69
- logarithmic characteristic 63-65 Television 12, Chap. 8
- motion 79-83 - ACTV 280,281
- simultaneous contrast 71,205,206 - adaptive modulation 243,246-248
- thresholds 60 - band-sharing 221,232
- Weber-Fechner fraction 63-65,78 - broadcasting 280-283
Phi phenomenon 15,79 - cable transmission 264
Photodetectors 8,45,46 - camera 211
Photoemission 55 - color subcarrier 221
Photometry 22-25 - comb filter 221,225
Photon 45-47 - compatible color 218,221
Planck's law 19 - compatible HDTV 226-228,263,272,277
Plate 10,140-145 - constant-luminance principle 223
Point-spread function 30 - cross-color 221,224
Poisson distribution 53,60 - cross-luminance 221,224
Printing 11,134-145 - data-under method 251
Probability distributions 110 - DBS 264-266
Psychophysics 5, Chap. 3 - digital broadcasting 281-283
- display 211
- Eureka program 277
Quantization, Chap.4
- Federal Communications Commission (FCC)
- adaptive 129
218,262,279-281
Quantum phenomena 45-52
- fiber-optic transmission 266
- field-sequential system 217-219
Radiation 19 - frequency modulation 255,256
Resolution 5, 12 - General Instrument system 281
- Rayleigh's criterion of 28 - HD-MAC 276
Retina 59,60 - HDTV in Europe 274-279

303
- HDTV in Japan 268-274 - simulcasting 281
- HDTV in USA 279-283 - spatiotemporal redundancy 240,241
- high-definition (HDTV), Appendix - spectrum efficiency 235, 263,264
- interference 236, 237 - subband adaptive selection 257-260
- interlace 210,217, 222 - subband coding 239,242,243,252-260
- Japan Broadcasting Corporation, - system design, Chap. 8
see Television, NHK - temporal aliasing 210
- Kell factor 217 - transcoding 236,252
- MAC 274,275 - transition scenario 266-268
- mixed highs 220 - vestigial-sideband modulation 222,232
- modulation methods 216 - Zenith system 239
- motion adaptation 248-250,277, 278 Temporal frequency response 5,8,15,16,78,
- motion-compensated interpolation 250, 277, 79
278 Text editing 139
- motion estimation 250 Tone rendition 5,67-69, 146-154
- multiplexed analog component system, Type composition 137-140
see Television, MAC Typography 119, 122
- MUSE system 268-273
- NHK 268
- NTSC improvement 224-228 Unsharp masking 73
- NTSC limitations 222, 223
- NTSC spectrum 225 Video signals 37-40
- NTSC system 208,215-217 Video spectrum 38-40
- noise-margin method 227-229 Vignetting 27
- open-architecture receiver 231 Visual perception, see Perception
- quadrature modulation 221, 222, 227,255
- satellite transmission, see Television, DBS
- scrambling 245-248 Wien displacement law 19

304