Refresher Topics −Television Technology

by Rudolf Mäusl
Introduction
Rudolf Mäusl, Professor at the University of applied Sciences Munich, gave a detailed overview of state-of-the-art television technol-
ogy to the readers of "News from Rohde & Schwarz" in a refresher serial.

The first seven parts of the serial were published between 1977 and 1979 and dealt with fundamentals of image conversion, transmis-
sion and reproduction, including a detailed description of the PAL method for colour TV signals. Further chapters on HDTV, MAC and
HD-MAC methods, satellite TV signal distribution and PALplus were added in two reprints.

In 1998, these topics were no longer of interest or in a state of change to digital signal transmission. This background has been fully
taken into account in the current edition of this brochure which also presents a detailed description of digital video signal processing in
the studio, data compression methods, MPEG2 standard and methods for carrier-frequency transmission of MPEG2 multiplex signals to
DVB standard.

An even more detailed discussion of the subject matter as well as of state-of-the-art technology and systems is given in the second
edition of the book by Rudolf Mäusl "Fernsehtechnik - Übertragungsverfahren für Bild, Ton und Daten" published by Hüthig Buch Ver-
lag, Heidelberg 1995 (only in German).
Contents
1 Transmission method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.1 Scanning
1.2 Number of lines
1.3 Picture repetition frequency
1.4 Bandwidth of picture signal
2 Composite video signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Blanking signal
2.2 Sync signal
3 RF transmission of vision and sound signals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.1 Vestigial sideband amplitude modulation
3.2 Sound signal transmission
3.3 TV transmitter and monochrome receiver
3.4 TV standards
4 Adding the colour information. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1 Problem
4.2 Chromatics and colorimetry
4.3 Luminance and chrominance signals, colour difference signals
5 Transmission of chrominance signal with colour subcarrier. . . . . . . . . . . . . . . . . . . . . . . . . 18
5.1 Determining the colour subcarrier frequency
5.2 Modulation of colour subcarrier
5.3 Composite colour video signal
5.4 NTSC method
5.5 PAL method
5.6 SECAM method
6 Colour picture pickup and reproduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6.1 Principle of colour picture pickup
6.2 Colour picture reproduction using hole or slot - mask tube
7 Block diagram of PAL colour TV receiver. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
8 PALplus system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
8.1 Spectrum of PAL CCVS signal
8.2 Colour plus method
8.3 Compatible transmission with 16:9 aspect ratio
9 Digital video studio signal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
10 Data compression techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
10.1 Redundancy reduction for the video signal
10.2 Irrelevancy reduction for the video signal
10.3 MUSICAM for the audio signal
11 Video and audio coding to MPEG2 standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
11.1 Definition of profiles and levels in MPEG2 video
11.2 Layers of video data stream
11.3 Layers of audio data stream
11.4 Packetized program and transport stream
12 Transmission of DVB signal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
12.1 Error correction
12.2 Satellite channel
12.3 Cable channel
12.4 Terrestrial transmission channel
13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
1 Transmission method
The principle of TV transmission with a brightness of the pattern. This signal cur- where α = αo = 1.5’ and tan αo = 4 x10-4
view to reproducing black-and-white pic- rent, which may contain components of the following approximation formula for
tures can be summarized as follows: the very high frequency due to fine picture calculating the minimum line number is
optical image of the scene to be transmit- details, must be applied to the receiver obtained:
ted is divided into small picture elements without distortion. This requirement
2500
(pixels). determines the essential characteristics L = ----------------- (2)
of the transmission system. D⁄ H
For D/H = 5, this means a number of
1.2 Number of lines L = 500 visible lines [1]. In accordance
The quality of the reproduced picture is with CCIR, the complete raster area has
determined by the resolution, which is been divided into 625 lines, 575 of which
reading writing the better the higher the number of lines, are in the visible picture area due to the
beam beam
a minimum number being required to vertical flyback of the beam (525 lines in
opt. opt.
ensure that the raster is not disturbing to North America and Japan with about
electr. electrical signal electr.
the viewer. In this context, the distance of 475 active picture lines).
converter converter
the viewer from the screen and the acuity
of the human eye have to be considered. 1.3 Picture repetition frequency
Fig 1 When determining the picture repetition
Principle of TV transmission. frequency the physiological characteris-
tics of the eye have to be considered. To
An opto-electrical converter, usually a reproduce a continuous rapid motion, a
camera tube, consecutively translates the H=L lines certain minimum frame frequency is
individual elements into electrical infor- α required so that no annoying discontinui-
mation depending on their brightness. ties occur. 16 to 18 frames per second, as
This signal is then transmitted at its are used for instance in amateur films,
actual frequency or after modulation onto are the lower limit for this frequency. 24
an RF carrier. After appropriate process- D frames per second are used for the cin-
ing at the receiving end, the information ema. This number could also be adopted
is applied to an electro-optical converter Fig 3 for television; however, considering the
and reproduced in accordance with the Angle of sight when viewing TV picture. linkage to the AC supply frequency, a pic-
brightness distribution of the pattern. ture repetition frequency (fr) of 25 Hz for
Continuous transmission is ensured by The optimum viewing distance is found to an AC supply of 50 Hz has been selected
producing a defined number of frames as be about five times the picture height, i.e. (30 Hz for a 60 Hz AC supply in North
in cinema films. D/H = 5 (Fig 3). At this distance, the line America and Japan).
structure should just be no longer visible,
1.1 Scanning i.e. the limit of the resolving power of the However, the picture repetition fre-
The pattern is divided into a number of eye should be reached. quency of 25 Hz is not sufficient for flick-
lines which are scanned from left to right erfree reproduction of the picture. The
and from top to bottom (Fig 1). The scan- Under normal conditions the limit angle is same problem had to be solved for the
ning beam is deflected horizontally and α about αo = 1.5’. From the equation: cinema where the projection of each indi-
vertically, writing a line raster. Synchro- vidual picture is interrupted once by a
⁄ L
tan α= H ≈ α
)

nizing pulses are transmitted to ensure -------------- (1) flicker shutter, thus producing the
D
that the reading and the writing beams impression that the repetition frequency
stay in step, covering the correct, corre- had been doubled.
sponding picture elements.

Scanning converts the individual picture

pattern line brightness distribution
elements from the geometrical into the in lines
Fig 2
1
time domain. Fig 2 gives a simplified rep- 2 1 2 3 4 5 6
Waveform of signal current in
resentation assuming that the scanning y 3 case of line-by-line scanning of
4
beam returns to the lefthand picture mar- 5 t pattern.
signal current
gin within a negligible period of time. In 6
x
general, the signal current obtained is a scanning direction
white
is
train of multishape pulses of varying black
t
mean value, corresponding to the mean line 1 2 3 4 5 6

Refresher Topics - Television Technology 5

This method cannot be used for televi- The resulting line or horizontal frequency In accordance with CCIR, the flyback
sion; here the solution found has been is intervals are defined as follows:
interlaced scanning. The lines of the com-
plete raster are divided into two fields, fh= 25 x 625 = 50 x 312½ = 15 625 Hz. Tf = 0.18 x Th = 11.52 µs
h
which are interlaced and transmitted
consecutively. Each field contains L/2 The period of the horizontal deflection is tf = 0.08 x Tv = 1.6 ms
v
lines and is swept within the interval Th = 64 µs, that of the vertical deflection
Tv/2. This means that lines 1, 3, 5 etc are Tv = 20 ms. The horizontal and vertical Thus, for transmitting the picture infor-
in the first field and lines 2, 4, 6 etc in the frequencies must be synchronous and mation, only the line interval Th x(1−0.18)
second field (geometrical line counting) phase-locked. This is ensured by deriving = 52.48 µs of the total line period Th and
(Fig 4). the two frequencies from double the line the portion Lx (1 −0.08) = 575 lines of the
frequency (Fig 5). L-line raster (= 2 Tv) can be used, the
raster area available for the visible pic-
line 1st field
1.4 Bandwidth of picture signal ture being reduced (Fig 8).
1
The resolution of the picture to be trans-
3
5
mitted is determined by the number of
line
1
2 . lines. With the same resolution in the hor- ^
=Th
3
4 . izontal and vertical directions, the width =tfh
5 line
.
.
of the picture element b is equal to the
2
.
. 4
line spacing a (Fig 6).
complete picture visible picture
. =2·Tv
.

2nd field

Fig 4 =2tfv

Division of complete raster for interlaced scanning. L lines

a b
Fig 8
When reproducing the two fields it is Raster area reduced due to flyback intervals.
essential that they be accurately inter-
laced since otherwise pairing of lines may Fig 6 For optical and aesthetic reasons a rec-
cause the field raster to appear in a very Resolution of pattern by line raster. tangular format with an aspect ratio of
annoying way. In a system using an odd 4:3 is chosen for the visible picture.
line number, for instance 625, the transi- At the end of a line, the scanning beam is
tion from the first to the second field returned to the left. After sweeping a With the same horizontal and vertical res-
takes place after the first half of the last field, it is returned to the top of the raster. olution, the number of picture elements
line in the first field. Thus no special auxil- per line is:
iary signal is required to ensure periodic
4
offset of the two fields. This detail will be --- x 625(1−0.08) = 767
3
discussed in section 2.2.
Ih tfh and the total number of picture elements
Th
t
in the complete picture:
4
--- x 625 x (1−0.08) x 625 x (1−0.08)
2 · fh 3
Iv tfv
= 440 833
2 625
Tv
1 1 t This number of picture elements is trans-
mitted during the time interval:
fh fv

64 µs x (1−0.18) x 625 x (1−0.08)

Fig 5 Fig 7 =30.176 ms
Coupling of horizontal and vertical deflection Periods of horizontal and vertical deflection
frequencies in case of interlaced scanning with flyback intervals. Thus the time TPE available for scanning
according to CCIR. one element is:
During flyback both the reading and the
30.176 ms
Thus 50 fields of 312½ lines each are writing beams are blanked. The required T PE = --------------------------- = 0.0684 µs
440833
transmitted instead of 25 pictures of 625 flyback intervals, referred to the period Th
lines, the field repetition or vertical fre- of the horizontal deflection and Tv of the
quency being fv = 50 Hz. vertical deflection, are given in Fig 7.

6 Refresher Topics - Television Technology

The highest picture signal frequency is Considering the finite beam diameter, the
obtained if black and white picture ele- vertical resolution is reduced compared
ments alternate (Fig 9). In this case, the pattern with the above calculation. This is
period of the picture signal is: scanning direction expressed by the Kell factor K. With a
TPE
value of K = 2/3, the bandwidth of the
TP = 2 x TPE = 0.137 µs picture or video signal, and the value laid
brightness
down in the CCIR standards, results as:
Due to the finite diameter of the scanning
beam the white-to-black transition is BW = 5 MHz.
rounded so that it is sufficient to transmit picture signal
the fundamental of the squarewave sig- 2TPE
t
nal. This yields a maximum picture signal
frequency of: Fig 9
1 Rounding of picture signal due to finite
f Pmax = ----- = 7.3 MHz beam diameter.
TP

Refresher Topics - Television Technology 7

2 Composite video signal
The composite video signal (CVS) is the 2.2 Sync signal The horizontal sync pulse is separated
complete television signal consisting of Synchronization pulses are required so from the sync signal mixture via a differ-
the scanned image (SI), blanking (B) and that line and field of the picture at the entiating network.
sync (S) components. The scanned image receiver stay in step with the scanning at
signal was dealt with in section 1. the transmitter. These sync signals drive Thus the leading edge of the pulse,
the deflection systems at the transmitter whose duration is 4.5 µs to 5 µs, deter-
2.1 Blanking signal and receiver ends. The sync pulse level is mines the beginning of synchronization,
During the horizontal and vertical beam lower than the blanking level, thus corre- i.e. at the beam return. The front porch
return, the scanned image signal is inter- sponding to a "blacker-than-black" region ensures that the beam returns to the
rupted, i.e. blanked. The signal is main- (Fig 11). lefthand picture margin within the blank-
tained at a defined blanking level which ing interval tbh (Fig 12). The back porch is
is equal to the black level of the video sig- the reference level. But it is also used for
nal or differs only slightly from it. In most transmitting additional signals, such as
cases the setup interval formerly used to 100% white level the colour synchronization signal.
distinguish between blanking level and
P picture signal
black level is nowadays omitted for the
black level
benefit of making better use of the whole 0 blanking level
S sync signal
level range. The signal used for blanking -40% sync level P
consists of horizontal blanking pulses
with the width: Fig 11
Level range of composite video signal. setup
S
tb = 0.18 x Th a b c
h
In this level range, the horizontal and ver- tbh
and vertical blanking pulses with the tical sync signals must be transmitted in a ^ h=64µs
H=T

width: distinctive way. This is why differing

pulse widths are used. Fig 12
tb = 0.08 x Tv Horizontal sync signal.
v
This and the different repetition fre-
Thus the signal coming from the video quency permit easy separation into hori- The vertical sync pulse is transmitted dur-
source is completed to form the picture zontal or vertical sync pulses at the ing the field blanking interval. Its duration
signal (Fig 10). receiver end. of 2.5 H periods (2.5 × 64 µs) is consider-
ably longer than that of the horizontal
sync pulse (about 0.07 H periods). To
white obtain regular repetition of the horizontal
SI
black end of
B 2nd field 1st field

B S
tbh 622 623 624 625 1 2 3 4 5 6 7 23 24

2.5 H 2.5 H 2.5 H

H
– 3H
2
white field-blanking interval (25 H + 12 µs)

P V pulse

black
setup blanking
end of 2nd field
B 1st field
level
t
S

309 310 311 312 313 314 315 316 317 318 319 320 336
Fig 10
2.5 H 2.5 H 2.5 H
Horizontal blanking signal and generation 1 )H
(3+ –
2
of picture signal.
field-blanking interval (25 H + 12 µs)
V pulse

Fig 13
Vertical sync signal with pre- and postequalizing pulses.

8 Refresher Topics - Television Technology

without preequalizing pulses
start of 1st field
with preequalizing pulses the two fields due to the half-line offset.
This in turn might cause pairing of the
raster lines. Therefore five narrow equal-
τo τo
Vswitch izing pulses (preequalizing pulses) are
Vint added to advance, at H/2, the actual ver-
2H + τo 2H + τo
t
tical sync pulse so that the same initial
start of 2nd field conditions exist in each field (Fig 14,
right). In a similar way, five postequaliz-
UC o ing pulses ensure a uniform trailing edge
Vswitch of the integrated vertical part pulses.
Vint
<(2H + H_ + τo) 2H + H
_ + τo
2 2
t
The following explanation of the line
numbering of Fig 13 is necessary. In tele-
Fig 14 vision engineering, the sequentially
Effect of preequalizing pulses: transmitted lines are numbered consecu-
left: integration of vertical sync pulse without preequalizing pulses; tively. The first field starts with the lead-
right: integration of vertical sync pulse with preequalizing pulses. ing edge of the vertical sync pulse and
contains 312½ lines. The first 22½ lines
sync pulse, the vertical sync pulse is horizontal sync pulse from one field to the are included in the field blanking interval.
briefly interrupted at intervals of H/2. At next. After 312½ lines the second field begins
the points marked in Fig 13, the pulse in the middle of line 313, also with the
edges required for horizontal synchroni- Since the vertical sync pulse is obtained leading edge of the vertical sync pulse,
zation are produced. Due to the half-line by integration from the sync signal mix- and it ends with line 625.
offset of the two rasters, the interruption ture, different conditions for starting the
takes place at intervals of H/2. Interlaced integration (Fig 14, left) would result for After the complete sync signal with the
scanning also causes the vertical sync correct level has been added to the pic-
pulse to be shifted by H/2 relative to the ture signal in a signal mixer, the compos-
ite video signal (CVS) is obtained.

Refresher Topics - Television Technology 9

3 RF transmission of vision and sound signals
For radio transmission of the television The problem is eluded by using vestigial carrier amplitude and the sync pulse to
signal and for some special applications, sideband amplitude modulation (VSB maximum carrier amplitude (Fig 18).
an RF carrier is modulated with the com- AM) instead of SSB AM. In this case, one
posite video signal. For TV broadcasting complete sideband and part of the other
and systems including conventional TV are transmitted (Fig 15, bottom).
receivers, amplitude modulation is used,
whereas frequency modulation is
employed for TV transmission via micro-
-1 1 2 3 4 5 6 MHz
wave links because of the higher trans-

-1.25
-0.75
5.5 MHz
mission quality. picture transmitter
fvision fsound
frequency
response fvision

Nyquist
receiver
slope 1.0
passband
LSB USB characteristic
AM
0.5
fvision
demod. CVS
frequency
response -1 1 2 3 4 5 6 MHz

SSB AM USB 5.5 MHz

fvision fsound
0
f
Fig 17
VSB AM VSB USB Fig 16 CCIR standard curves for picture transmitter
fvision f Correction of frequency response in vestigial frequency response (top) and receiver passband
sideband transmission by Nyquist filter. characteristic(bottom).
Fig 15
RF transmission of CVS by modulation of vision carrier: At the receiver end it is necessary to
top: amplitude modulation, carrier with two side- ensure that the signal frequencies in the 100% sync level
bands; region of the vestigial sideband do not black level

center: single sideband amplitude modulation; appear with double amplitude after
10% white level
bottom: vestigial sideband amplitude modulation. demodulation. This is obtained by the 0 carrier zero
Nyquist slope, the selectivity curve of the
3.1 Vestigial sideband amplitude receiver rising or falling linearly about the
modulation vision carrier frequency (Fig 16).
The advantage of amplitude modulation
is the narrower bandwidth of the modula- In accordance with CCIR, 7 MHz bands Fig 18
tion product. With conventional AM the are available in the VHF range and 8 MHz Negative amplitude modulation of RF vision carrier
modulating CVS of BW = 5 MHz requires bands in the UHF range for TV broadcast- by CVS.
an RF transmission bandwidth of BWRF = ing. The picture transmitter frequency
10 MHz (Fig 15, top). In principle, one response and the receiver passband char- A residual carrier (white level) of 10% is
sideband could be suppressed since the acteristic are also determined by CCIR required because of the intercarrier
two sidebands have the same signal con- standards (Fig 17). In most cases, both sound method used in the receiver. One
tent. This would lead to single sideband modulation and demodulation take place advantage of negative modulation is opti-
amplitude modulation (SSB AM) (Fig 15, at the IF, the vision IF being 38.9 MHz and mum utilization of the transmitter, since
center). the sound IF 33.4 MHz. maximum power is necessary only briefly
for the duration of the sync peaks and at
Due to the fact that the modulation sig- The modulation of the RF carrier by the the maximum amplitude occurring peri-
nals reach very low frequencies, sharp CVS is in the form of negative AM, where odically during the sync pulses to serve as
cutoff filters are required; however, the bright picture points correspond to a low a reference for automatic gain control in
group-delay distortion introduced by the receiver.
these filters at the limits of the passband
causes certain difficulties.

10 Refresher Topics - Television Technology

 picture transmitter
picture transmitter 3.3 TV transmitter and monochrome
osc.
multi-
plier
driver
output
stage
output stage
with VSB filter
receiver
AM
The RF television signal can be produced
equa- mod. vision/ to
by two different methods.
CVS sound
lizer ampl diplexer antenna

If the modulation takes place in the out-

sound osc.
multi-
plier
driver
output
stage
put stage of the picture transmitter
FM
sound transmitter
(Fig 19), the RF vision carrier is first
brought to the required driving power
Fig 19 and then, with simultaneous amplitude
Block diagram of TV transmitter using output stage modulation in picture transmitter. modulation, amplified in the output stage
to the nominal vision carrier output
power of the transmitter. The modulation
38.9 MHz
picture transmitter amplifier boosts the wideband CVS to the
IF IF VSB output
level required for amplitude modulation
mixer driver
osc. mod. filter stage in the output stage. The sound carrier is
AM frequency-modulated with a small devia-
equa-
tion at a relatively low frequency. The
CVS
lizer final frequency and the actual frequency
fixed con- channel vision/ to
33.4 MHz tuning verter tuning sound antenna deviation are produced via multiplier
osc. diplexer
IF stages. The picture and sound transmitter
sound 1
FM
osc. output stages are fed to the common
output
adder mixer driver
stage antenna via the vision/sound diplexer.
IF
sound 2
osc.
FM When using IF modulation (Fig 20), first
33.15 MHz sound transmitter
the IF vision carrier of 38.9 MHz is ampli-
Fig 20 tude-modulated. The subsequent filter
Block diagram of TV transmitter using IF modulation in picture and sound transmitters. produces vestigial sideband AM. One or
two sound carriers are also frequency-
modulated at the IF. Next, mixing with a
3.2 Sound signal transmission modulated with the sound information. common carrier takes place both in the
In TV broadcasting the sound signal is The frequency of the intercarrier sound is vision and in the sound channel so that
transmitted by frequency-modulating the constant and is not influenced by tuning the vision/sound carrier spacing of
RF sound carrier. In accordance with the errors or variations of the local oscillator. 5.5 MHz is maintained at the RF. Linear
relevant CCIR standard, the sound carrier amplifier stages boost the vision and
is 5.5 MHz above the associated vision More recent studies have shown further sound carrier powers to the required
carrier. possibilities of TV sound transmission, in level.
particular as to transmitting several
The maximum frequency deviation is sound signals at the same time. A second The advantage of the second method is
50 kHz. Due to certain disturbances in sound channel permits, for instance, mul- that the actual processing of the RF tele-
colour transmission, the original sound/ tilingual transmission or stereo operation. vision signal is carried out at the IF, thus
vision carrier power ratio of 1:5 was at a lower frequency, and band- and
reduced to 1:10 or 1:20 [2]. Even in the With the dual-sound carrier method, an channel-independent. However, for fur-
latter case no deterioration of the sound additional sound carrier 250 kHz above ther amplification, stages of high linearity
quality was apparent if the signal was the actual sound carrier is frequency- are required, at least in the picture trans-
sufficient for a satisfactory picture. modulated, its power level being 6 dB mitter.
lower than that of the first sound carrier.
As mentioned above, the intercarrier A multiplex method offers further possi-
sound method is used in most TV receiv- bilities by modulating an auxiliary carrier
ers. The difference frequency of 5.5 MHz at twice the line frequency or using the
is obtained from the sound and vision car- horizontal or vertical blanking intervals
rier frequencies. This signal is frequency- for pulse code modulation.

Refresher Topics - Television Technology 11

Table 1 Frequency ranges, vision/sound carrier spacing, channel width, Reproduction of the image on the
sound modulation receiver screen is based on proper ampli-
fication of the RF signal arriving at the
Standard CCIR OIRT French VHF FCC (USA) antenna (Fig 21). To this effect, the
B, G D E M incoming signal is converted into the IF in
VHF, band l / MHz 47 to 68 48.5 to 100 50 to 70 54 to 88 the VHF/UHF tuner, where also standard
VHF, band III / MHz 174 to 230 174 to 230 160 to 215 174 to 216 selection by the Nyquist filter and the
UHF, band IV/V / MHz 470 to 853 470 to 890 required amplification are provided. The
Vision/sound carrier spacing 5.5 MHz 6.5 MHz 11.15 MHz 4.5 MHz subsequent demodulator generates the
Channel width 7 MHz (B) 8 MHz 13.15 MHz 6 MHz CVS and the 5.5 MHz sound IF.
The latter is limited in amplitude, ampli-
8 MHz (G)
fied and frequency-demodulated. The
Sound modulation, FM deviation FM, 50 kHz FM, 50 kHz AM FM, 25 kHz
CVS is applied to the video amplifier; after
separation of the sync component, the
signal is taken to the control section of
Table 2 Composite video signal the CRT via the video output stage. The
sync signal brought out from the video
Standard CCIR OIRT French VHF FCC (USA)
amplifier by way of a sync separator is fed
B, G D E M
to the horizontal deflection system via a
Number of lines 625 625 819 525 differentiating network and to the vertical
Field-repetition frequency 50 Hz 50 Hz 50 Hz 60 Hz deflection system via an integrating net-
Line frequency 15 625 Hz 15 625 Hz 20 475 Hz 15 750 Hz work.
Video bandwidth 5 MHz 6 MHz 10.6 MHz 4.2 MHz
Line duration H 64 µs 64 µs 48.84 µs 63.5 µs The line deflection frequency is produced
Field duration 20 ms 20 ms 20 ms 16.667 ms in the horizontal oscillator and compared
to the incoming horizontal sync pulses in
a phase discriminator. A control circuit
ensures that the correct frequency and
sound IF 5.5 MHz
sound
FM AF phase relation to the transmitter sync sig-
demod. ampl.
IF ampl. nal is maintained. In the horizontal output
VHF/
from
UHF
IF video stage, the required deflection power is
antenna tuner ampl. demod.
produced and the high voltage for the
video video SI
driver
output CRT is obtained from the line retrace
CVS stage
control
voltage
pulses. The vertical oscillator is synchro-
gen. B nized directly by the vertical sync signal.
sync
sep.
The blanking pulses required for the
S
HB VB beam retrace are derived from the hori-
HS sync VS vert.
zontal and vertical output stages.
vert.
pulse output
sep. osc. stage
3.4 TV standards
hor.
The characteristics of the television sig-
phase hor.
discr. osc.
output
stage HV
nals mentioned in sections 1, 2 and 3
refer to the CCIR standard. Various other
standards are in use; for the differences
in the standard specifications see Tables
Fig 21
1 and 2.
Block diagram of monochrome TV reveiver.

12 Refresher Topics - Television Technology

4 Adding the colour information
To reproduce a colour image of the pat- (Fig 22). The chrominance signal cannot
tern, additional information on the colour be obtained directly from the pattern.
content, i.e. the "chromaticity" of the indi- Instead the three primaries (red, green, 1.0

vidual picture elements, must be trans- blue) are used in accordance with the 0.8
h
mitted together with the brightness or three-colour theory (Helmholtz). The red,
0.6
luminance distribution. This requires first green and blue signals are also required
the extraction of the colour information for reproducing the colour picture. Thus 0.4

and then a possibility of reproducing the the scheme of compatible colour trans- 0.2
colour image. mission is established by the luminance
signal Y and the chrominance signal F 0
400 500 600 700 nm
4.1 Problem (Fig 23). blue cyan green yellow red λ
The problem of colour transmission con-
sists in maintaining the transmission Fig 24
method of black-and-white television and Brightness sensitivity characteristic of human eye.
luminance signal Y
broadcasting the additional colour infor- B&W
camera
B&W
picture tube
mation as far as possible within the avail- In monochrome television, where only
able frequency band of the CVS. This colour
R
G
R
G colour-
the luminance distribution of a coloured
coder decoder B
means for any colour TV system that a camera B picture tube pattern is transmitted, this sensitivity
colour broadcast is reproduced as a per- characteristic of the eye has to be taken
fect black-and-white picture on a mono- chrominance signal F into account. This is done by using the
chrome receiver (compatibility) and that a spectral sensitivity of the camera tube
colour receiver can pick up a mono- Fig 23 and, if required, correction filters in con-
chrome broadcast to reproduce a perfect Principle of compatible colour transmission. nection with the colour temperature of
black-and-white picture (reverse compat- the lighting.
ibility). 4.2 Chromatics and colorimetry
Light is that part of the electromagnetic
radiation which is perceived by the
more
saturated green
luminance human eye. It covers the wavelengths less
from about 400 nm (violet) to 700 nm
colour stimulus

hue colourless
(equal-energy white)
coloured
chrominance (red). The light emitted by the sun con-
pattern saturation sists of a multitude of spectral colours
merging into each other. Spectral colours
Fig 22 are saturated colours. Mixing with white 400 500 600 700 nm
Representation of coloured pattern by luminance light produces desaturated colours. λ
and chrominance components.
Coloured (chromatic) light can be charac- Fig 25
These requirements can be met only if terized by its spectral energy distribution. Colour stimulus at different degrees of saturation.
The radiation of the wavelength λ causes
– information on the luminance distribu- the sensations of brightness and colour in The colours of objects are those colours
tion and the eye. The sensitivity to brightness of that are reflected from the light to which
– information on the colour content the human eye as a function of the wave- the object is exposed. The colour stimulus
length is expressed by the sensitivity curve shows the associated spectral dis-
are obtained from the coloured pattern characteristic or luminosity curve (Fig 24). tribution (Fig 25). In most cases, the
and then transmitted. object colours are not spectral colours but
This characteristic indicates how bright rather mixtures consisting of a number of
The chromaticity is characterized by the the individual spectral colours appear to closely spaced spectral colours or of sev-
hue – determined by the dominant wave- the eye when all of them have the same eral groups of spectral colours.
length of light, for instance for distinct energy level. It can be seen from this
colours such as blue, green, yellow, red – characteristic that certain colours appear This is an additive process. White (colour-
and by the saturation as a measure of dark (e.g. blue) and others bright (e.g. less) can also be produced by mixing.
spectral purity, i.e. of colour intensity green). Fig 26 shows typical examples of additive
with respect to the colourless (white) colour mixing.

Refresher Topics - Television Technology 13

Investigations into the colour stimulus primary stimulus must be added to the curves in an r-g diagram, the locus of all
sensitivity of the human eye have shown colour stimulus. spectral colours is plotted (Fig 28).
that a colour sensation is produced by
mixing the part sensations caused in the In a colorimetric representation of a col-
primaries red, green and blue. This leads our stimulus, the three primaries yield a
to the conclusion that any colour appear- space vector. The direction of the colour Y

ing in nature can be obtained by combin- vector in space determines the chroma- g
ing the corresponding portions of the pri- ticity, its length being a measure of the
2
maries red, green and blue. In accord- brightness.
ance with the Helmholtz three-colour the-
gr
ee
ory, Grassmann (1854) found the follow- n
1
ing law: 0.4
G
_

blu
_
r(λ)

red
e-g
b(λ)

ree
F = R (R) + G (G) + B (B). (3) 0.3 Wo

n
colour stimulus
_ Z R
g(λ)
0.2 -2 -1 B 1 2
X
r
0.1

Fig 28
0
red
Colour surface in r-g diagram.
-0.1
400 500 600 700 nm
yellow purple λ Due to the negative portion of the r (λ)
white
colour mixture curve, here again negative
Fig 27 values are obtained. Coordinate transfor-
green cyan blue
Colour mixture curves b(λ), g(λ), r(λ), referred to mation referring to a new set of fictive,
R, G, B. i.e. non-physical primaries X, Y and Z,
yields a curve which comprises only posi-
However, a three-dimensional coordinate tive colour values [3]. When using the fic-
Fig 26 system is not convenient for graphic rep- tive primaries (standard reference stimuli
Additive colour mixing using three primaries R, G, B. resentation. But since brightness and X, Y, Z), the relationship according to the
chromaticity are independent of each equation (4) still holds, expressed by the
This means that a distinct colour stimulus other, the tristimulus values can be standard reference summands x, y, z:
F can be matched by R units of the spec- standardized to the luminance compo-
tral colour red (R), G units of the spectral nent: x + y + z = 1. (6)
colour green (G) and B units of the spec-
tral colour blue (B).
F
---------------------------- = γ 1.0

Monochromatic radiations of the wave- R+G+B

520
y spectrum locus
lengths 0.8
R(R ) G( G) B(B )
δ = ---------------------------- + ---------------------------- + ----------------------------
G

λR = 700 nm, λG = 546.1 nm, and R+G+B R+G+B R+G+B 560

0.6
λR = 435.8 nm 500 570
580
=1 (4) 0.4 600
have been determined as the standard R
W 700 nm
spectral colours, called primary colours or or, using the chromaticity coordinates: 490
0.2
stimuli. None of the three primaries must 480 line
ple
pur
be obtainable from the other two by mix- r + g + b = 1. (5) B
0
ing. 0 0.2 0.4 0.6 0.8 1.0
These reduced values no longer contain x
Based on the equation (3), colour mixture any luminosity information but merely the
curves were plotted, showing the portion chromaticity. Fig 29
of each primary stimulus required for the Standard colour diagram; colour surface in
different spectral colours (Fig 27). The Since, however, the sum of r, g and b is x-y diagram.
ordinate scale refers to equal-energy always unity, one of the three coordi-
white. As can be seen from the curves, nates can be omitted when specifying the The two-dimensional representation of
negative amounts or tristimulus values chromaticity so that a two-dimensional chromaticity in the x-y coordinate system
are associated with some components. system, the colour surface, is obtained. is called the standard colour diagram in
This means that for matching certain When entering the chromaticity coordi- accordance with CIE (Commission Inter-
spectral colours a specific amount of a nates found from the colour-mixture nationale de l’Eclairage) or, briefly colour

14 Refresher Topics - Television Technology

triangle (Fig 29). The area of colour stimuli high luminance −which can be produced curves using their spectral sensitivity and
that can be realized by additive mixing is economically −are required. In accord- additional colour filters (see Fig 23). The
enclosed by the spectrum locus and the ance with recent EBU (European Broad- output voltage of the colour camera in the
purple line. The line connecting the white casting Union) specifications, the three channels must have the same rela-
point W (equal- energy white) of x = 0.33 receiver primaries Rr, Gr and Br are used; tionship as the tristimulus values. With
and y = 0.33 to the position of any colour their colour coordinates are given in Fig standard illuminant C, the three output
F yields the dominant wavelength, i.e. the 30. signals must be equal, even at different
hue, when extended to its point of inter- luminance values.
section with the spectrum locus. The colour mixture curves plotted with
the aid of the primary stimuli R, G and B 4.3 Luminance and chrominance sig-
The ratio of the distance between the hue are based on equal-energy white W. In nals, colour difference signals
F and the white point W to the distance colour TV technology, standard illuminant For reasons of compatibility, the colour
between the spectrum locus and the C is used as reference white, correspond- camera has to deliver the same signal-
white point W on the connecting line ing to medium daylight with a colour tem- from a coloured pattern, i.e. the lumi-
passing through the colour position gives perature of about 6500 K, the standard nance signal to the monochrome
the colour saturation. The closer the col- tristimulus values being: receiver, as the black-and-white camera.
our point to white, the weaker the colour The spectral sensitivity of a black-and-
saturation. The position of a mixture col- xC = 0.310, yC = 0.316, zC = 0.374. white camera corresponds to the bright-
our is situated on the line joining the loci ness sensitivity curve of the human eye to
of the two colours mixed or, if three col- If now the tristimulus values of the ensure that the black-and-white picture
ours are added, within the triangle receiver primaries Rr, Gr and Br are found tube reproduces the different colour stim-
formed by the connecting lines. for all spectral colour stimuli of equal uli as grey levels of the same brightness
radiation energy and plotted as standard- as perceived by the eye.
ized tristimulus values as a function of the
spectrum locus
wavelength λ (maximum of the curve
0.8
G referred to 1), the colour mixture curves 1

y
used in TV engineering are obtained (540/0.92)
h
0.6
Gr (Fig 31).
0.5
range of (610/0.47)
object
colours (465/0.17)
0.4 0.2
1.0 _ _ _
W br(λ) gr(λ) rr(λ)
C Rr R 0.75 Br Gr Rr
colour stimulus

400 500 600 700 nm

0.2 λ
0.5

Br
0.25 Fig 32
B
0
0
Brightness sensitivity of human eye to receiver pri-
0 0.2 0.4 0.6 0.8
x maries.
-0.25
400 500 600 700 nm
Fig 30 λ The colour camera, however, delivers
Colour coordinates of receiver primaries and three signals with spectral functions
representable colour range. Fig 31 matching the colour mixture curves. To
Colour mixture curves br (λ), gr (λ), rr (λ), referred to obtain one signal whose signal spectral
When defining the tristimulus values in a receiver primaries Rr, Gr, Br. function corresponds to the sensitivity
colour TV system, it is essential to con- curve of the eye, coding is required. To
sider the question of how the primary Here again negative colour values appear this end, the three colour values repre-
stimuli can be realized at the receiver due to the colour stimuli located outside sented by the functions rr (λ), gr (λ) and
end. On the one hand, the requirements of the triangle formed by Rr, Gr and Br. br (λ) are multiplied by the relative lumi-
regarding the receiver primaries are nosity coefficients hr, hg and hb and then
determined by the necessity of providing Therefore, the colour mixture curves are added up. Except for the proportionality
as wide a range of representable mixture slightly modified for practical purposes constant k, the result must be identical to
colours as possible, i.e. the chromaticity (dashed line). The signals produced by the sensitivity function of the human eye
coordinates of the receiver primaries the camera tubes in the red, green and h(λ).
should be located on the spectrum locus. blue channels of the colour camera must
On the other hand, primaries of especially be matched with these colour mixture h(λ) = k[hr x rr(λ) + hg x gr(λ)
+ hb x br(λ)]. (7)

Refresher Topics - Television Technology 15

The relative luminosity coefficients hr, hg Table 3 Signals of standard colour bar The equation
and hb are obtained by normalization sequence
from the corresponding values h(Rr), h(Gr) VY = 0.30 x VR + 0.59 x VG + 0.11 x VB (10)
and h(Br) of the eye (Fig 32): Pattern R G B Y
White 1 1 1 1.00 VR −VY = 0.70 x VR −0.59 x VG
h(Rr) = 0.47 Yellow (R + G) 1 1 0 0.89 −0.11 x VB (11)
h(Gr) = 0.92 Cyan (G + B) 0 1 1 0.70 and
h(Br) = 0.17 Green 0 1 0 0.59 VB −VY =−0.30 x VR −0.59 x VG
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ Purple (R + B) 1 0 1 0.41 + 0.89 x VB (12)
Σh = 1.56
Red 1 0 0 0.30
The colour difference signals only carry
Blue 0 0 1 0.11
The following figures are calculated for information on the colour content. For a
Black 0 0 0 0
the relative luminosity coefficients: monochrome scene (VR = BG = VB) they
h ( R r ) 0.47 are equal to zero.
h r = --------------- = ----------- = 0.30 To reproduce a colour pattern, the three
Σh 1.56 tristimulus signals R, G and B are The amplitude of the colour difference
h ( G r ) 0.92 required. For compatibility, however, col- signals shows the departure of the hue
h g = --------------- = ----------- = 0.59 our transmission uses the luminance sig- from the colourless, which is a measure
Σh 1.56 nal Y and the chrominance signal F. The of colour saturation.
h( B r ) latter cannot be obtained directly from
0.17
h b = --------------- = ----------- = 0.11 the tristimulus values but only by way of The hue is determined by the amplitude
Σh 1.56 the colour difference signals: ratio and by the polarity sign of the colour
Due to normalization: difference signals. Transformation of the
R −Y, G −Y, B −Y. orthogonal coordinate system (B −Y and
hr + hg + hb = 1. (8) R −Y) into polar coordinates (Fig 34) yields
These are the colour values minus the the colour saturation from the vector
Thus for equation (7) luminance component. length A:

h(λ) = k [0.30 x rr(λ) + 0.59 x gr(λ)

+ 0.11 x br(λ)]. (7’) VR 1|1|1 A = ( B – Y )2 + ( R – Y )2 (13)
colour VG 1|1|0 colour-picture
camera VB 1|0|0 tube

is obtained or, written in a simplified way, and the hue from the angle α:
pattern
for the luminance signal Y corresponding white | yellow | red
(R – Y)
to the sensitivity curve of the eye α = arc tan --------------------- (14)
B&W V
Vy
B&W picture
(B – Y)
camera tube
1|0.89|0.30 1|0.89|0.30
Y = 0.30 x R + 0.59 x G + 0.11 x B. (9)

This equation is one of the most impor-

tant relations of colour TV engineering. Fig 33
Producing luminance signal VY from tristimulus R-Y

Technically the luminance signal VY is voltages VR, VG, VB and compatibility relationship. A

obtained from the tristimulus signals VR, α

VG and VB via a matrix (Fig 33). The sig- The chrominance signal carries informa- B-Y
nals listed in Table 3 below result for a tion on hue and saturation. Therefore two
pattern consisting of eight colour bars colour difference signals are sufficient to Fig 34
(standard colour bar sequence) −the describe the chrominance component. Representation of chromaticity as a function
three primaries plus the associated com- For this purpose, the quantities R −Y and of colour difference signal.
plementary colours and the colourless B −Y were selected [4].
stripes white and black.
referring to the voltage of the luminance
signal derived from the coder yields these
two colour difference signals as:

16 Refresher Topics - Television Technology

Investigations have shown that the reso- 2. Driving colour picture tube with The advantage of this method is that the
lution of the eye is lower for coloured pat- colour difference signals (Fig 37) bandwidth of the final amplifier stages
tern details than for brightness varia- The third colour difference signal VG −VY for the colour difference signals is smaller
tions. Therefore it is sufficient to transmit is obtained in a matrix from the two quan- than that of the stages for the tristimulus
only the luminance signal with the full tities VR −VY and VB −VY, based on the fol- signals and that a black-and-white pic-
bandwidth of 5 MHz. The bandwidth of lowing equations ture appears if the colour difference sig-
the chrominance signal can be reduced nals fail. The disadvantage, is that higher
to about 1.5 MHz by taking the two colour VY = 0.30 x VR + 0.59 x VG + 0.11 x VB voltages must be produced in the final
difference signals via lowpass filters. VY = 0.30 x VY + 0.59 x VY + 0.11 x VY stages associated with the colour differ-
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ence signals. Fig 38 shows signals for the
VY −VY = 0.30 x (VR −VY) standard colour bar pattern.
VR VY=0.30xVR+0.59xVG+0.11xVB + 0.59 x (VG −VY)
colour
camera
VG
VB Coder
VR-VY=0.70xVR–0.59xVG–0.11xVB
VB-VY=0.30xVR–0.59xVG+0.89xVB
+ 0.11 x (VB −VY) = 0 (15)
pattern white yellow cyan green purple red blue black
or, after rewriting,
1.0
Fig 35
Producing luminance signal VY plus colour VG −VY = −0.51 x (VR −VY) R 0

difference signals VR −VY and VB −VY. −0.19 x (VB −VY) (16). 1.0

G 0
In the coder, the signals VR, VG and VB 1.0
produced by the colour camera are con-
-VY B 0
verted into the luminance component VY VR
1.0 0.89 0.70
1.0
and the colour difference signals VR −VY 0.59
0.41 0.30
0.11
and VB −VY (Fig 35) and, in this form,
0
Y 0
VR–VY
applied to the reproducing system. How- VG 0.59 0.70
ever, the tristimulus values are required 0
0 0.11 -0.70 -0.59 -0.11 0 1.40
G-Y
for unbalancing the red, green and blue matrix R-Y 0.89
0.59
beams. Two different methods are com- VB 0.30
0 -0.89 -0.59 -0.30 0
1.78
monly used for restoring these colour val- 0
B-Y
ues: VB–VY
0.41
0.11 0.30
0 -0.41 -0.30 -0.11 0
0 0.82
1. Driving colour picture tube with Fig 37 G-Y

RGB voltages (Fig 36) Restoring tristimulus values when driving colour
The tristimulus voltages VR, VG and VB are picture tube with colour difference signals. Fig 38
produced from the luminance component Tristimulus values, luminance component and
VY and the two colour difference signals The colour difference signals are taken to colour difference signals for standard colour bar
VR −VY and VB −VY via matrices and the control grids of the deflection sys- pattern.
applied directly to the control grids of the tems; the negative luminance signal is
colour picture tube, the cathodes being at applied to the cathodes so that the tris-
fixed potential. timulus signals are obtained as control
voltages at the three systems, for
instance:

Ucontr R = (UR −UY) −(−UY) = UR.

0
(17)
VY
R
matrix VR
VR–VY

G
matrix VG

VB–VY
B
matrix VB

Fig 36
Restoring tristimulus values when driving colour
picture tube with RGB.

Refresher Topics - Television Technology 17

5 Transmission of chrominance signal
with colour subcarrier
As explained in the preceding chapter, into the gaps of the CVS frequency spec- the frequency region of the CVS produces
the colour TV signal is transmitted in the trum. This is done by modulating the a bright-dark interference pattern on the
form of the luminance signal Y and the chrominance signal onto a colour subcar- screen.
two colour difference signals R −Y and B − rier whose frequency fSC, and thus also
Y for reasons of compatibility. To transmit the spectrum of the line-repetitive modu- If the colour subcarrier frequency is an
the complete picture information −lumi- lation products, is located between the integer multiple of the line frequency, i.e.
nance plus chrominance −a triple trans- frequency components of the CVS if there is no offset with respect to the
mission channel would be required. Here (Fig 40). line frequency, an interference pattern of
one might think of making multiple use of bright and dark vertical stripes appears,
the TV transmission channel either by fre- their number corresponding to the factor
quency or time multiplex. However, nei- n (Fig 41). As a result of the half-line off-
ther method is compatible with the exist- set the phase of the colour subcarrier
colour subcarrier
ing black-and-white transmission alternates by 180° from line to line of a
method. field. However, because of the odd
amplitude
number of lines, bright and dark dots
m x fh fF – fh (m+1) x fh fF (m+2) x fh coincide after two fields. The interference
f pattern occurring in a rhythm of fv/4 =
fv fv
12.5 Hz would thus be compensated over
Fig 40 four fields (Fig 42). Nevertheless, the
amplitude Spectrum of CVS and modulated colour subcarrier. compensation of the interference pattern
m x fh fSC– fh (m+1) x fh fSC (m+2) x fh on the screen is not perfect due to the
f 5.1 Determining the colour subcarrier nonlinearity of the picture tube character-
frequency istic and the inadequate integrating
Fig 39 One condition for determining the colour capability of the human eye.
Detail of CVS spectrum. subcarrier frequency results from the
symmetrical interleaving of the CVS and If a half-line offset is assumed between
A decisive thought for the colour trans- chrominance signal spectra: the fre- the colour subcarrier frequency and the
mission method actually selected is quency fSC should be an odd multiple of line frequency, the subjective annoyance
derived from spectral analysis of the lumi- half the line frequency fh: can be further reduced by selecting the
nance signal or CVS. It becomes apparent colour subcarrier frequency as high as
f
that only certain frequency components f SC = ( 2n + 1 )x ----h (18) possible. In this way, the interference
occur in the CVS spectrum, these compo- 2 pattern takes on a very fine structure.
nents being mainly multiples of the line In this way the half-line offset is However, the colour difference signals
frequency due to the periodic scanning obtained. modulate the colour subcarrier so that for
procedure. transmitting the upper sideband a certain
Generally, the interference which a col- minimum spacing of the colour subcarrier
The varying picture content amplitude- our subcarrier produces on the black-and- frequency from the upper frequency limit
modulates the line-repetitive pulse white picture can be explained as fol- of the CVS has to be maintained.
sequence producing sidebands spaced at lows, if it is assumed that a sinewave in
multiples of the field frequency from the
spectral components of the line pulse.
Fig 41
Fig 39 shows a detail of the CVS spec-
Interference pattern
trum. Essentially the spectrum is occu-
caused by colour line 1 3 5 7 1st field
pied only at multiples of the line fre- 1 3 5 7
in
3rd field
subcarrier with integer
quency in their vicinity. Between these
relationship between
frequency and groups the spectrum
colour subcarrier
exhibits significant energy gaps. line 2 4 6 2nd field
frequency and line 2 4 6
in
4th field
frequency.
Since the colour information is also line-
line line
repetitive, the spectrum of the chromi- (7)
1 fSC=n x fh
nance signal consists only of multiples of 2 e.g. n=3
the line frequency and the corresponding 3
4
sidebands. Therefore it is appropriate to 5
6
insert the additional colour information 7

18 Refresher Topics - Television Technology

 Fig 42 In the NTSC and PAL methods, double
Compensation of amplitude modulation is used. The 0°
line 1 3 5 7 1st field interference pattern with component of the subcarrier is ampli-
in
1 3 5 7 3rd field
half-line offset of colour tude-modulated by the (B −Y) signal and
subcarrier frequency. its 90° component by the (R −Y) signal,
the subcarrier being suppressed at the
line 2 4 6 2nd field
2 4 6
in
4th field same time. This results in quadrature
modulation (Fig 45). The modulation
line line
(7)
product is a subcarrier frequency whose
f
1 fSC = (2n+1) x _h amplitude and phase are determined by
2 2
3
e.g. n=3 the two colour difference signals. Ampli-
4
5
tude and phase modulation take place at
6 the same time. Compared to the chromi-
7
nance signal represented in Fig 34, the
colour saturation now corresponds to the
Y Y Fig 43 amplitude SC and the hue to the phase
Principle of compatible angle ϕSC of the modulated colour sub-
B, S B, S
R CCVS R colour TV transmission carrier. For this reason, the modulated
G matrix add. sep. matrix G
B B
using luminance and colour subcarrier is also called the
chrominance signals. chrominance signal.
R-Y B-Y

2 2
mod. demod.
SC = (B – Y) + (R – Y) (21)
SC SC
R –Y
ϕSC = arc tan ----------- (22)
Y B –Y
fSC SC fSC

fSC f
(B-Y)
B-Y
mod.

colour
The best compromise has proved to be a standard colour subcarrier frequency in subcarrier
0

colour subcarrier frequency of about the PAL system is chrominance

4.4 MHz. Thus the principle of compatible 90 + signal
colour TV transmission is found using the fsc =283.75 x fh + 25 Hz
90
luminance signal and the chrominance = 4.43361875 MHz (20)
signal modulated onto the subcarrier; this R-Y
(R-Y)
mod.
is the basis of the NTSC system and its The rigid coupling of the subcarrier fre-
variants (Fig 43). For the CCIR-modified quency to the line frequency is ensured
NTSC method, which will be discussed in by deriving the line frequency fh or twice Fig 45
detail later, a colour subcarrier frequency the line frequency 2fh from the frequency Producing chrominance signal by quadrature
of fSC of the colour subcarrier (Fig 44). modulation of colour subcarrier with two colour
difference signals.
f
f SC = 567x ----h = 283.5xf h
2 fSC
SSB f fV
A vector diagram of the chrominance sig-
2f
= 4.4296875 MHz (19) f
nal shows the position of the different
fSC–fV/2
265f colours on the colour circle (Fig 46). Simi-
has been fixed. 1135f f lar to the colour triangle in Fig 29, the
2fh
f 8f complementary colours are located on
A further development of the NTSC opposite sides of the coordinate zero (col-
method is the PAL method, which is Fig 44 ourless). When transmitting a colourless
being widely used today. In the PAL sys- Coupling of colour subcarrier frequency and picture element, the colour difference
tem, one component of the subcarrier is line frequency. signals and the amplitude of the colour
switched by 180° from line to line. How- subcarrier equal zero. In this case, the
ever, this cancels the offset for this sub- 5.2 Modulation of colour subcarrier subcarrier does not cause any interfer-
carrier component so that a pronounced The chrominance signal is transmitted by ence in the black-and-white picture.
interference pattern would appear in the modulating the colour subcarrier with the
compatible black-and-white picture. This two colour difference signals. The modu- To demodulate the chrominance signal,
is avoided by introducing a quarter-line lation method must permit the colour dif- the unmodulated carrier of correct phase
offset of the subcarrier plus an additional ference signals to be extracted separately is required.
offset of fv/2 = 25 Hz. Thus the CCIR at the receiver end.

Refresher Topics - Television Technology 19

 5.3 Composite colour video signal reduction factors for the two colour dif-
The chrominance signal is combined with ference signals, multiplying them by
(R-Y) axis
the composite video signal (CVS) to form
the composite colour video signal (CCVS). 0.49 for the (B −Y) signal
red purple
The CCVS is amplitude-modulated onto and 0.88 for the (R −Y) signal
yellow SC
the RF vision carrier. The full level of the
ϕSC
colour difference signals would cause In this way the reduced colour difference
(B-Y) axis
blue overmodulation of the RF vision carrier by signals U and V are obtained:
the chrominance signal for certain col-
green
cyan oured patterns. This is shown for the sig- U = (B −Y)red = 0.49 x (B −Y)
Fig 46 nals of the standard colour bar sequence = −0.15 x R −0.29 x G +0.44 x B (23)
Vector representation of chrominance signal: (Table 4).
colour circle. V = (R −Y)red = 0.88 x (R −Y)
Overmodulation occurs in both directions = 0.61 x R −0.52 x G −0.10 x B (24)
Synchronous detection is used, evaluat- (Fig 49). In particular, the periodic sup-
ing only the chrominance component pression of the RF carrier and its falling (The values are rounded off.)
which is in phase with the reference car- short of the 10% luminance level would
rier. Since the actual subcarrier is not cause heavy interference. For this reason, The signal values listed in Table 5 are
transmitted, it must be produced as a ref- the chrominance signal amplitude has to obtained for the colour bar pattern with
erence carrier at the receiver end. For be reduced. Αs a compromise between 100% saturated colours. Fig 50 shows the
synchronization with the subcarrier at the overmodulation on the one hand and line oscillogram of the CCVS for a 100%
transmitter end, a reference signal is degradation of signal-to-noise ratio on saturated colour bar sequence.
inserted into each line in the H blanking the other −an overmodulation of 33% in
interval, i.e. the colour sync signal or both directions with fully saturated col- For measurements and adjustments on
burst. This signal consists of about ten ours has been permitted since, in prac- colour TV transmission systems the test
oscillations of the subcarrier at the trans- tice, fully saturated colours hardly ever signal used is the standard colour bar
mitter end and is transmitted on the back occur. This is ensured by using different sequence for which all chrominance sig-
porch (Fig 47).

burst Fig 48 chrominance signal and burst

(B-Y)
B−Y
Synchronization of demod.

H sync pulse reference carrier

−90˚
oscillator at receiver
burst ϕ ref.
end and synchronous carrier fSC, 90˚
amp. comp. osc.
tbh detectors associated (R-Y)
R−Y
demod.
with two colour dif- integr.
fh
Fig 47 ference signals.
Colour sync signal (burst).

In the NTSC system, the burst phase is at

180° compared to the 0° reference phase
of the colour subcarrier. 100%

In the receiver the burst is separated from RF

vision carrier 75% 0
the chrominance signal by blanking. A amplitude
phase comparator produces a control
voltage from the departure of the refer- Y
ence carrier phase from the burst phase;
this voltage controls the frequency and
10% 1.0
phase of the reference carrier oscillator 0
via an integrating network. The control
voltage becomes zero if the phase differ-
ence is 90°. Coming from the reference
carrier oscillator, the 90° component is
white yellow cyan green purple red blue black
taken directly to the (R −Y) synchronous
detector and, after a negative 90° phase Fig 49
shift, as the 0° component to the (B −Y) Amplitude modulation of RF vision carrier by CCVS without reducing colour difference signals.
synchronous detector (Fig 48).

20 Refresher Topics - Television Technology

Table 4 Overmodulation of RF vision carrier by standard colour bars. nals, except in the white bar, are reduced
to 75% in accordance with the EBU (Euro-
Colour bar Y B −Y R −Y SC Y + SC Y −SC pean Broadcasting Union) standard. In
White 1 0 0 0 1 1 this way, 33% overmodulation by the col-
Yellow 0.89 −0.89 + 0.11 0.89 1.78 0 our subcarrier is avoided (Fig 51).
Cyan 0.70 + 0.30 −0.70 0.76 1.46 −0.06
Green 0.59 −0.59 −0.59 0.83 1.42 −0.24 To determine the colour subcarrier phase
Purple 0.41 + 0.59 + 0.59 0.83 1.24 −0.42 or the colour points in the (B −Y) (R −Y)
plane, a vectorscope is used. This is an
Red 0.30 −0.30 + 0.70 0.76 1.06 −0.46
oscilloscope calibrated in polar coordi-
Blue 0.11 + 0.89 −0.11 0.89 1.00 −0.78
nates including two synchronous detec-
Black 0 0 0 0 0 0
tors for the U and V components. The
detected colour difference signals are
taken to the X and Y inputs of the oscillo-
scope. Fig 52 shows the vector oscillo-
Table 5 Modulation of RF vision carrier by standard colour bars with reduced colour gram of the standard colour bar
difference signals sequence.

Colour bar Y U V SCred Y + SCred Y −SCred

White 1 0 0 0 1 +1
Yellow 0.89 −0.44 + 0.10 0.44 1.33 + 0.45 V 90
−0.62
120 60
Cyan 0.70 + 0.15 0.63 1.33 + 0.07 red
purple
Green 0.59 −0.29 −0.52 0.59 1.18 0 150 30
Purple 0.41 + 0.29 + 0.52 0.59 1.00 −0.18
Red 0.30 −0.15 + 0.62 0.63 0.93 −0.33 yellow
0
−0.10 −0.33
180
Blue 0.11 +0.44 0.44 0.55 burst blue
U

Black 0 0 0 0 0 0
210 330
green
cyan
240 300
270

33%
1.33 1.33 Fig 50
1.18
Line oscillogram of CCVS for Fig 52
1.00 1.00
picture 1.0 0.93 standard colour bar sequence, Vector oscillogram of standard colour bar sequence.
signal 0.89
colour saturation 100%, colour
0.70
0.59
0.55
difference signals reduced. 5.4 NTSC method
0.45
0.41 The NTSC, PAL and SECAM methods,
0.30
which are mainly used for colour TV
0.11
0 transmission, differ only with respect to
0 0.07
0 the modulation of the colour subcarrier.
-0.18 The NTSC method, named after the
0.4 -33%
-0.33 -0.33 National Television System Committee,
white yellow cyan green purple red blue black constitutes the basis for the improved
variants PAL and SECAM.

The principle of the NTSC method has

1.00 1.00 1.00 Fig 51
1.0 0.87
basically been described in the chapters
0.75 Line oscillogram of CCVS for on modulation of the colour subcarrier
0.66 0.70
picture
standard colour bar sequence, and on the composite colour video signal.
0.52
signal colour saturation 75%
0.44
0.42 However, the original NTSC system (US
0.30
0.22
(EBU test signal). standard) does not transmit the reduced
0.33
0.08
0
colour difference signals U and V but
0 0.05
0
instead the I and Q components, which
-0.13 are referred to a coordinate system
-0.25 -0.25
-0.4 rotated counter-clockwise by 33° (Fig 53).
white yellow cyan green purple red blue black

Refresher Topics - Television Technology 21

In the colour triangle, the I axis corre- The human eye reacts very strongly to a large extent by negative feedback, the
sponds to the axis for which the eye has incorrect hue. The hue of the colour differential phase can be reduced only by
maximum colour resolution (Fig 54). This image reproduced by the picture tube is limiting the driving level. Fig 56 shows
ensures better transmission of colour determined by the phase angle of the this influence on the CCVS and the effect
transitions. chrominance signal referred to the phase in the vectorscope representation.
of the burst. When producing the CCVS in
the studio, it may happen that the 5.5 PAL method
chrominance signal from different The effects of static and differential
V
sources has different delays and thus dif- phase errors are considerably reduced
red purple
ferent phases with respect to the burst. with the PAL method, the occurrent
To correct wrong hue resulting from static phase errors being corrected with rela-
J

Q
phase errors in the transmission path, the tively little extra outlay. The PAL system is
yellow 33
NTSC colour TV receiver is provided with based on the following concept: an exist-
U
blue a control permitting the phase of the ref- ing phase error can be compensated by a
erence carrier to be adjusted. In most phase error of opposite polarity. This is
cases this is done by referring to the hue realized technically by alternating the
green cyan
of a well-known picture detail, such as phase of one of the two chrominance sig-
the flesh tone. nal components, for instance the SCV
component, by 180° from line to line. PAL
Fig 53 However, this hue control does not allow stands for phase alternation line.
I and Q components of reduced colour difference correction of differential phase distortion.
signals in original NTSC system. In accordance with DIN 45 061 differen- If a phase error exists in the transmission
tial phase means the difference of phase path, alternately positive and negative
The modulation signals are thus shifts through a four-terminal network at departures of the chrominance signal
two different points of the transfer char- phase from nominal are produced in the
l = V x cos 33° −U x sin 33° (25) acteristic at the subcarrier frequency. The receiver after elimination of the line-to-
differential gain is defined in a similar line polarity reversal of the SCV compo-
Q = V x sin 33° + U x cos 33° (26) way. nent generated at the transmitter end.
Delaying the chrominance signal for the
or, using the tristimulus matrix equations, Due to the shift of the operating point on duration of a line (64 µs) and subsequent
the transmission characteristic as a func- addition of the delayed and the unde-
l = 0.60 x R −0.28 x G −0.32 x B (27) tion of the Y components of the CCVS, the layed signals cause two phase errors of
chrominance signal suffers a gain change opposite polarity to coincide and thus to
Q = 0.21 x R −0.52 x G + 0.31 x B (28) (because of the change in slope of the cancel each other. It should be men-
characteristic) and a phase change tioned, however, that this method is
(because the transistor input capacitance based on the assumption that the chro-
minimum is dependent on the emitter current and maticity does not change within two con-
Y green colour perception thus on the operating point) when pass- secutively transmitted lines. If horizontal
yellow maximum ing, for instance, through an amplifier colour edges exist, the eye hardly per-
cyan
W
stage with preceding tuned circuit. While ceives a falsification of the colour transi-
red the differential gain can be eliminated to tion even in this case.
is
Q ax purple
blue X

Fig 55 1-µs
J axis
Block diagram of delay

Fig 54 NTSC coder.

R Y
I LP 0.6-µs
I-Q axes in colour triangle. G matrix
Q 1.3MHz delay
I mod.
B

fSC, 123
The two signals I and Q are transmitted add. add.
CCVS
NTSC
with different bandwidth, i.e.
LP
123˚ Q mod.
0.5 MHz
the I signal with 1.3 MHz
and the Q signal with 0.5 MHz.
fSC, 33
33˚

Fig 55 shows the complete block diagram

burst
fSC, 0
of an NTSC coder. Except for the 33° gen.
phase shift, the functioning of the corre-
sponding decoder is essentially explained BI, S
SC signal
in Fig 48. osc. gen.

22 Refresher Topics - Television Technology

Fig 57 shows the compensation of a (R −Y) sync detector. To this effect, the The reference carrier oscillator in the
phase error with the PAL method. The burst is split into two components, one receiver adjusts, via the phase control, to
assumed phase error α affects the being transmitted at 180° and the other 90° with respect to the mean burst
chrominance signal with respect to the at ±90° alternating from line to line in phase. The identification signal (burst
burst on the transmission link. phase with the SCV reversal. This yields a flag) for synchronizing the PAL switch is
swinging burst of 180° ±45° (Fig 58). The derived from the burst phase discrimina-
After elimination of the SCV polarity actual burst reference phase (180°) is tor (see Fig 62). With PAL the reduced col-
reversal (PAL switchover) and addition of obtained by averaging. our difference signals U and V are directly
the chrominance signals in two succes- transmitted, the bandwidth being
sive lines, the phase angle of the result- 135°
1.3 MHz. Limiting the sidebands of the
ing signal SCres is equal to that of the modulated colour subcarrier to different
transmitted chrominance signal, and the line n n = odd number in widths no longer has an annoying effect
1st and 2nd fields
original hue is thus maintained. After thanks to the phase error compensation.
reducing the resulting signal to half its Fig 59 is a block diagram of a PAL coder.
amplitude, this signal exhibits only slight line n+1 n = even number in
Compared with the NTSC coder, the 33°
desaturation. 3rd and 4th fields phase shift of the colour subcarrier com-
225° ponents is omitted, but reversal of the
An additional identification is transmitted subcarrier component for the (B −Y) mod-
with the burst to ensure correct phase Fig 58 ulator and generation of the swinging
reversal to the SCV component in the Swinging burst with PAL method. burst are added.
receiver or of the reference carrier for the
The technical realization of the PAL error
compensation requires special explana-
V2 tion as against Fig 57. For this purpose it
V
is best to start with the group delay
red
purple decoder included in the PAL decoder. In
contrast to the NTSC decoder, the
yellow chrominance signal is not simultaneously
V1 applied to the two sync detectors in the
t U
PAL decoder but is first split into the SCU
blue
and SCV components.

green
This is performed in the group delay
t cyan decoder (Fig 60). At its output, the incom-
ing chrominance signal is divided into
three components. It is taken to the two
outputs via a 64 µs delay network (line
Fig 56 duration) directly and after a 180° phase
Generation of differential gain and phase distortion. shift. At the outputs, signal addition takes
place. The chrominance signal of the pre-
chrominance signal ceding line (SCn) and that of the ongoing
after polarity reversal addition of chrominance signals
ahead of transmission link after transmission link of SCv component in lines n and n+1 line (SCn +1) are added at the SCU output.
SCn
SCn´ SCn´ Successive lines contain the SCV compo-
+SCv
α SCres nent with a 180° phase alternation so
that the SCV component is cancelled
line n ϕ ϕ+α
SCu SCn´ every two lines. Thus the SCU component
x
SC´n+1
of the chrominance signal is constantly
available at this output.
ϕ

x
The input signal is taken to the SCV out-
SC´n+1
put with a 180° phase shift. Addition of
the delayed chrominance signal cancels
SCu ϕ−α
line n+1 the SCU component, and the SCV compo-
−ϕ nent appears at this output, although
α with a 180° phase alternation from line to
-SCv
SCn+1
SC´n+1 line. Based on the vector diagrams in
Fig 61, the functioning of the group delay
Fig 57 decoder can be explained very easily.
Compensation of phase error with PAL.

Refresher Topics - Television Technology 23

 0.4-µs Fig 59 The line-to-line phase alternation of the
delay Block diagram of SCV component can be disabled by a con-
Y
NTSCcoder. trolled switchover. However, it is easier to
R LP
G matrix U
V 1.3MHz
U mod. provide for line-to-line phase reversal of
B
the reference carrier in the (R −Y) sync
fSC 0 add. add.
CCVS detector. Within the complete PAL
PAL
burst decoder this task is performed by the PAL
gen.
switch, which is synchronized by the
LP
swinging burst (Fig 62).
-90 V mod.
1.3 MHz

signal at output SCU output SCV

+2 x SC´v
fSC 90 0 /180
fSC± 90
SC´n α SC´n
–SC´n+1
BI, S
SC
signal α 2 x SC´n
osc.
gen. line n+1

Fig 60 SC´n+1
output
SCu 64 µs
PAL group delay decoder.
SC´n
…, SCn+1+SCn, SCn+SCn+1,…
input
τ = 64 µs α 2 x SC´n
…, –SCn+1+SCn, –SCn+SCn+1,… line n+1
SCn, SCn+1,…
SCn+2 = SCn,…
180° SCv
SC´n+1 α SC´n+1
–SC´n
signal at input output SCU output SCV Fig 61
SCn Division of chromi- –2 x SC´v
+SCV
nance signal into
line n SCU and SCV Fig 63
SCU
components in PAL Effect of phase error on signals with PALgroup delay
group delay decoder.
+2 x SCV
64 µs decoder.
A phase error in the transmission path
SCn
–SCn+1 SCn affects both the SCU and the SCV compo-
SCU 2 x SCU
nents in the same sense (Fig 63). Since,
line n+1
however, only the component in phase
-SCV
with the reference carrier is weighted in
the sync demodulators, the (B −Y)
SCn+1
SCn+1
64 µs
demodulator delivers the signal
+SCV
SCn SCn U’ =  SCU x cos α and the (R −Y)
demodulator the signal
V’ =  SCV x cos α. Both colour differ-
2 x SCU
line n+2
^= line n SCU
ence signals are reduced by the same fac-
SCn+1 –SCn SCn+1
tor so that the ratio V/U or (R −Y)/(B −Y)
remains constant and the hue of the
–2 x SCV
reproduced image is not affected. Desat-
uration, which corresponds to the factor
cos α, becomes significant only with
SCU U B-Y Fig 62 large phase errors.
add.
demod. PAL decoder with
chrominance reference carrier
64µs -90 generation.
signal fSC, 0

±SCV V R-Y
180 add.
demod.

0 /180
control voltage fSC, ±90

fh/2
burst, 135 /225 phase ref.-
carrier
from burst amp. discr. osc. fSC, 90 gen.

synchronization

24 Refresher Topics - Television Technology

5.6 SECAM method Fig 64 Y

As against the NTSC method, the SECAM Simplified block diagram

R
method too brings an improvement with of SECAM coder (above) G matrix
B-Y FM
add.
CCVS
B mod. SECAM
respect to wrong hues caused by phase and decoder (below). R-Y

errors in the transmission path. Like the

fh/2 fSC B, S
PAL method it is based on the assumption
that the colour information does not
essentially vary from line to line or that SCB–Y FM
B–Y
the human eye does not perceive any …, SCB-Y, SCR-Y
demod.

annoyance if the vertical colour resolu-

SCR–Y FM
tion is reduced to a certain extent. 64µs
demod.
R–Y

Therefore, the colour difference signals fh/2

(B −Y) and (R −Y) characterizing the col-
our information need not be transmitted
simultaneously. They can be sent sepa- Fig 64 is a simplified block diagram of a ated by preemphasis at the transmitter
rately in successive lines. In the receiver, SECAM coder and decoder. In the coder, end and boosted by deemphasis at the
the signal content of one line is stored for the (B −Y) and the (R −Y) signals are receiver end. The effect of noise is
64 µs via a delay line and processed applied to the frequency modulator in reduced by video frequency preemphasis
together with the signal of the next line. alternate lines. To ensure that in the and deemphasis.
The short form SECAM derived from decoder the demodulated colour differ-
"séquentiel à mémoire" indicates that ence signals are in synchronism with the SECAM has gone through several phases
this is a sequential colour system with transmitter end, identification pulses in of development. Its latest variant, SECAM
memory. the form of the modulated colour subcar- III b or SECAM III opt., is based on slightly
rier are transmitted during nine lines of different colour subcarrier frequencies for
As the two colour difference signals are the field blanking interval. the (B −Y) and (R −Y) signals, further
transmitted separately, the type of modu- reducing the interference pattern caused
lation can be freely selected. SECAM When frequency-modulating the colour by the colour subcarrier.
uses frequency modulation, which is not subcarrier, the latter is not suppressed. In
very interference-prone. However, the particular with colours of low saturation, As against PAL, SECAM features some
reference frequency of the FM demodula- this would produce an interference pat- system-dependent weak points since fre-
tor must be kept very stable so that the tern on a black-and-white receiver in quency modulation is utilized at its physi-
demodulated colour difference signals spite of the colour subcarrier offset. cal limits [3].
are not falsified. Therefore the colour subcarrier is attenu-

Refresher Topics - Television Technology 25

6 Colour picture pickup and reproduction
Previous explanation was based on an causing a charge compensation at the
electrical picture signal obtained from the corresponding picture elements due to
pattern to be transmitted by optoelectri- the resulting lower blocking resistance.
correcting filter
cal conversion. Below, the converters During the next charging process, elec-
blue
used at the TV transmitter and receiver trons are again bound at these places on VB
ends are briefly looked at and finally the one side of the target plate and on the
reproduction of the colour picture by the other side the same number of electrons red
VR
TV receiver is explained. are set free. These electrons flow across
the external circuit resistance causing a green
signal voltage to be produced. Fig 66 VG

focusing coil
shows the equivalent circuit of a picture colour splitter pickup tubes
glass disk with storage layer electron gun point on the target, represented by a
capacitor shunted by an exposure- Fig 67
dependent resistor.. Splitting of incident light into three primaries
incident
in colour TV camera.
light

6.2 Colour picture reproduction using

R
signal electrode deflection coil aligning coil signal electrode cathode hole or slot mask tube
C For reproducing the brightness pattern,
Fig 65 television uses picture tubes with a phos-
signal
Design of Vidicon camera tube. voltage
RL phor screen which lights up in accord-
ance with the intensity of the electron
6.1 Principle of colour picture pickup beam falling upon it. The electron beam is
+ –
An optoelectrical converter is used to 10…50V deflected within the raster with the aid of
translate brightness variations into an magnetic fields produced by the deflec-
electrical picture signal. Different con- Fig 66 tion currents in the horizontal and vertical
verter systems are available, but only the Equivalent circuit for picture element on storage deflection coils (Fig 68). The intensity of
pickup tubes with a photosensitive semi- plate of Vidicon camera tube. the electron beam is influenced by the
conductor layer are really important for voltage across the control electrode.
TV technology. In line with the basic principle of colour
transmission, three pickup tubes are
In Vidicon tubes a semiconductor layer is required; the image to be televised is pro-
used as the storage plate or target, its jected, using the primaries red, blue and electron gun
blocking resistance varying with the green, onto three photosensitive semi-
intensity of the light falling upon it. The conductor targets via an optical beam
characteristics of the converter differ splitter, called colour splitter, and via cor-
depending on the composition of the recting filters for matching the signals to
phosphor
semiconductor target. Frequently a Plum- the spectral sensitivity of the semicon- deflection
screen
coils
bicon is used; this has a target of lead ductor layers (Fig 67). To make the three
monoxide and, compared to the Vidicon partial images coincide accurately with HV connector, 20 kV
using an antimony-trisulphide layer, fea- their rasters , high mechanical and
tures higher sensitivity and less inertia. electron-optical precision is required. Fig 68
Coincidence errors of the colour rasters Black-and-white picture tube with deflection
Fig 65 shows a Vidicon pickup tube with would cause a loss in definition for the system.
its deflection and focusing coils. The Vidi- luminance signal. For this reason, colour
con works as follows: the electron beam, television cameras with a separate Whereas a homogeneous, whitish-blue
caused to emanate from the cathode by pickup tube for the luminance signal are phosphor screen is used in picture tubes
an electric field, negatively charges the also used. Progressive developments, for monochrome reproduction, the screen
side of the target facing the beam-pro- already partly implemented in portable of the colour picture tube must emit the
ducing system. Positive charge carriers colour TV cameras, point to a single-tube primaries red, green and blue.
are bound on the picture side of the tar- colour TV camera producing the tristimu-
get with the aid of the positive target- lus signals in the red, green and blue However, colour detail resolution should
plate voltage. At the points upon which channels by a multiplex method. go as far as the individual picture ele-
light falls, the incident photons release ments. For this purpose each picture ele-
electrons in the semiconductor layer ment is represented on the screen by

26 Refresher Topics - Television Technology

three luminescent dots in the colours red, To obtain −with common deflection of the the electron beams only strike lumines-
green and blue, called colour triad (Fig three electron beams −a clear assignment cent stripes of the correct colour associ-
69). About three times 400 000 lumines- to the luminescent dots, a hole mask is ated with the corresponding beam pro-
cent dots are accommodated on the placed about 15 mm from the phosphor ducing system.
screen area. screen (Fig 70).
Moreover, in conjunction with a special
deflection field, excellent convergence is
R, G, B achieved, i.e. the correct luminescent
hole mask luminescent dots
G R B G R
areas associated with the picture ele-
B G R B G B
B R G ments are excited. With the delta colour
G R B G R B B tube, greater complexity of circuitry was
R
G
R
required for this purpose.
Fig 69 electron beam phosphor screen
Arrangement of luminescent dots in colour
picture tube. Fig 70 slot mask

Part of hole mask and phosphor screen in delta

B phosphor
The colour picture tube includes three colour picture tube. G screen
beam-producing systems. In the delta R
colour picture tube, which for many More recent colour picture tubes are fit- electron beam
years was practically the only colour pic- ted with a slot mask. Accordingly, the
ture tube in use, the beam-producing sys- luminescent areas on the screen are R
G
tems are arranged at angles of 120° to either oblong or in the form of stripes. B phosphor stripes
one another. The intensity of the emitted With this in-line colour picture tube the
electron beams is controlled by the tri- three beam-producing systems are Fig 71
stimulus signals that are applied. arranged in one plane (Fig 71). This con- Part of slot mask and phosphor screen with in-line
figuration yields high colour purity, i.e. colour picture tube.

Refresher Topics - Television Technology 27

7 Block diagram of PAL colour TV receiver
Detailed signal tracing is often no longer After amplification the chrominance sig- rier. The control voltages for the refer-
possible in the circuit diagram of a mod- nal is applied to the group delay decoder, ence carrier oscillator and the sync signal
ern TV receiver. The reason is the fre- where it is split into the SCU and SCV for the PAL switch are derived from the
quent use of integrated circuits combin- components. As already explained, the phase discriminator. Furthermore, the
ing several functional units. Even though colour difference signals are retrieved in phase discriminator drives the colour
this entails a large degree of standardiza- the two synchronous demodulators. The killer, which inhibits the chrominance
tion in circuit design, there is still a broad tristimulus signals for driving the colour amplifier if the burst is absent so that no
line of integrated circuits being offered picture tube are obtained from the matrix coloured noise occurs when reproducing
for colour TV receivers due to the variety circuit. a black-and-white picture.
of possible combinations of the func-
tional units. Parallel to its application to the chromi- Following the sync separator, the sync
nance amplifier, the burst is taken to the signals are fed separately to the horizon-
For better explanation of the interworking phase discriminator via an amplifier trig- tal and vertical deflection circuits as with
of the functional units included in a PAL gered by the line sync pulse. Here the sig- a black-and-white receiver. Additional
colour TV receiver, the individual units are nal is compared with the reference car- correction signals, which are applied to
shown separately in Fig 72. The dash-dot-
ted lines indicate the possibilities of com-
bination that are feasible and have been
5.5 MHz sound IF sound IF AF
realized in practice by way of including sound IF ampl. demod. ampl.
several functional units in integrated cir- from VHF- IF IF
UHF
cuits. antenna tuner filter ampl.

vision IF CCVS
demod. ampl.
The RF signal coming from the antenna is
control
converted into the IF in the VHF-UHF voltage
gen.
tuner; next it is taken via the Nyquist filter
Y
to compensate for the vestigial sideband delay
H line
component and then boosted in the IF
amplifier to the level required for demod-
ulation. To ensure that there is no inter- ±SCV
group SCV
modulation between the colour subcar- chrom.
ampl.
delay sync
R
output
decoder demod. colour picture
rier and the intercarrier sound carrier, the tube
SCU
IF sound carrier is isolated from the vision matrix
G
output
IF demodulator. In a separate diode cir-
cuit, the 5.5 MHz intercarrier sound IF sig- SCU
colour B
nal is produced from the sound IF and the killer
sync
demod. output HB, VB

vision IF. More recent developments tend

towards the split-carrier method, i.e. the
33.4 MHz sound IF signal is directly ampli- PAL PAL
-90
multivibr. switch
fied and detected.
ref.
The detected CCVS is derived from the burst phase
carrier
HV
ampl discr. osc. gen.
vision IF demodulator. Next the sync sig-
nal is separated and the chrominance sig-
H
nal is brought out selectively via a
VS vert. vert.
4.43 MHz bandpass filter. The remaining osc. output
Y component is delayed by about 1 µs to sync conv.
sep. corr.
match the longer signal delay in the
phase hor. hor.
chrominance amplifier; after further HS discr. osc. output convergence deflection
amplification, the signal is taken to the coils coils

matrix circuit. H

Fig 72
Block diagram of PAL standard colour TV receiver.

28 Refresher Topics - Television Technology

the deflection coils and/or the conver- The D2-MAC method introduced in 1983 So, the PALplus method is the only topic
gence correction coils of the deflection for use in the whole of Europe has been remaining of the extensions to the first
system of the colour picture tube, are superseded by the rapid development of edition, but even this enhanced PAL
derived from the horizontal and the verti- digital TV. Only about 10% of the program method with a wide aspect ratio will in
cal deflection signals. signals presently transmitted via satellite the near future be replaced by the digital
are to D2-MAC standard and the method MPEG2 standard. Because of its compati-
Modular technology and the use of inte- has never been used for terrestrial trans- bility to PAL and therefore to the great
grated circuits lead in part to subassem- missions. For this reason there is no room number of PAL colour TV sets in use, this
blies whose functions extend to other for this method in a compendium on TV analog procedure can also be employed
units. If required, further subassemblies technology and a description of the HD- in parallel with digital TV.
can be added for user controls, for MAC method also becomes superfluous.
instance, multi-standard decoding, time A digital video source signal is required
insertion, picture-in-picture insertion or Broadcasting of TV signals via satellite for both the PALplus and the MPEG2
viewdata and teletext using analog FM will remain important encoder. So a detailed description of the
for some time. However, the satellite processing of digital video studio signals
The chapters added to the reprints of the channels have become so numerous has been added after the revised chapter
refresher serial have been omitted in the since the last edition of this brochure that on PALplus.
present edition for the following reasons: it is not possible to deal in detail with the
coverage of the ASTRA and EUTELSAT
There is no point now in examining the satellites, which are only relevant for
analog HDTV signal as the source signal Europe. Furthermore, the direct-reception
from the TV camera or the film scanner is satellite TV-SAT 2 has been assigned a
immediately subjected to an A/D conver- special application. The literature on sat-
sion and transmitted in the form of a dig- ellite reception and distribution tech-
ital HDTV signal. This is described in the niques has become so extensive by now
chapter "Digital video studio signal" of that this subject will not be dealt with
the present revised edition. either in the following.

Refresher Topics - Television Technology 29

8 PALplus system
In cooperation with renowned industrial Comb filters are used in the receiver for nance effects which are however less dis-
companies, scientific institutes, Euro- separating the spectral components. In a turbing. In simple PAL TV receivers these
pean broadcasting companies and the standard PAL colour TV receiver, the effects are eliminated by limiting the fre-
Institute for Broadcasting Technology comb filter function for separating the FU quency band of the luminance amplifier
(IRT), an enhanced PAL system called and FV components of the chrominance to approx. 3.5 MHz.
PALplus was developed permitting trans- signal is assumed by the group delay
mission with a wide aspect ratio. decoder (see Fig 60). The periodic transfer
function is obtained by adding or sub-
Characteristic features of the PALplus tracting the luminance signal delayed by H(Fu)
system are the line period H, i.e. 284 x TSC = 64.056
−2 x fh −1 x fh fsc +1 x fh +2 x fh
µs, for setting a phase shift of k x 2 π. Fig f
– suppression of cross colour and cross 74 illustrates the transfer function of the
luminance effects for full utilization of notch filter for the two outputs FU and FV.
the 5 MHz luminance bandwidth H(Fv)

– 16:9 aspect ratio With the comb filter transfer function −2 x fh −1 x fh fsc +1 x fh +2 x fh
superimposed on the PAL CCVS it can be f

and planned seen that the chrominance signal is

divided into FU and FV components but Fig 74
– digital sound transmission that high-frequency components of the Transfer function of group delay decoder acting as
– correction of echoing. luminance signal persist in the two signal comb filter for outputs FU and FV in colour subcarrier
spectra (Fig 75). As a result of these range (fSC).
Since PALplus is fully compatible with the cross-colour or cross-chrominance
standard PAL system, PALplus signals effects, an interference colour pattern is Different methods have been developed
can be transmitted in existing 7 MHz or displayed on the screen when the for the suppression of cross-colour
8 MHz channels or in the PAL satellite chrominance components are demodu- effects, some of them being already
channel. Pictures with a 16:9 aspect ratio lated into the primaries R, G, B. implemented in PAL TV receivers. The
are displayed on a standard 4:3 screen suppression of luminance/chrominance
with the black bands at the top and bot- crosstalk artefacts usually involves
tom known from widescreen film repro- a) expensive memories and also adaptive
ductions. measures for eliminating the side effects
of the suppression.
In addition to the planned digital audio Y F

signals, analog FM audio signals will also

be transmitted. The correction of echoing 0 3 MHz 5 MHz
f "Fu"
is another technical feature of the PAL-
plus system. b)
Fu Fu Fu Fu Fu Fu fsc
Fv Fv Fv Fv Fv Fv Fv
8.1 Spectrum of PAL CCVS Y Y Y Y Y Y Y

In the frequency range between approx.

−2 x fh −1 x fh fsc +1 x fh +2 x fh "Fv"
3 MHz and 5 MHz, the PAL CCVS contains f

components of the luminance signal Y

and the chrominance signal F(Fig 73, top). Fig 73 fsc

A detailed analysis of the PAL CCVS spec- Spectrum of PAL CCVS

trum yields interleaved spectra of the Y top: luminance signal Y and chrominance signal F Fig 75
signal, i.e. the CVS, and of the chromi- bottom: interlaced spectra of Y, FU and FV signals Separation of chrominance signal in group delay
nance signal components FU and FV in colour subcarrier range (fSC). decoder into FU and FV components with crosstalk
(Fig 73, bottom). The signal comes as a of Y signal.
frequency spacing multiplex with the Chrominance signal components appear
components completely separated in the in the broadband Y signal between 3.1
undisturbed transmission channel [5]. MHz and 5 MHz, causing cross-lumi-

30 Refresher Topics - Television Technology

8.2 Colour plus method spaced by 312 lines after storing the sig- The time resolution is based on 25 picture
The suppression of crosstalk artefacts in nal received first. motions per second. If the video signal is
the standard PAL system can be consider- obtained by film scanning, the signal
ably improved through the use of the col- – Adding yields a cross-effect-free lumi- source does not allow for more than
our plus method. It is based on a com- nance signal, i.e. the high-frequency 25 movements/s anyway, i.e. the picture
pensation of interfering signal compo- component YH, content of the two fields is referenced to
nents by averaging the signals of two – subtracting yields the cross-effect-free the same motion phase.
consecutive fields with respect to time, chrominance signal F.
which is also refered to as intraframe If the video signal comes from an elec-
averaging (IFA) [6]. Fig 77 illustrates the principle used in a tronic camera, the motion phases in the
colour plus encoder and decoder. In prac- two fields may differ with rapidly chang-
After averaging, the luminance signal Y tice, averaging in the decoder is per- ing pictures. In this case, the crosstalk
and the colour difference signals U and V formed with the demodulated colour dif- artefact is not compensated for and only
in two successive fields are identical. The ference signals using the 312H memory the low-frequency component YL of the
colour subcarrier and therefore the in the chrominance signal path. luminance signal down to about 3.1 MHz
chrominance signal Fare however shifted is transmitted to avoid the chrominance
by 180° after one field, i.e. after exactly signal being disturbed by luminance sig-
312 lines. This is obtained from the colour nal components.
a)
subcarrier period as defined by 0 to 3.1 MHz

∆τ
YL Another variant of the colour plus method
Tsc = 1/fsc Y matched to camera and film mode is the
Y
= 1/4.43361875 MHz
312H YH
= 0.2255 µs motion adaptive colour plus
3.1 to 5 MHz (MACP) method.
over 312 x 64 µs = 19 968 µs with U

U 312H
88 530.5 colour subcarrier periods. PAL F Here a stepwise (4 steps for the lumi-
0 to 1.3 MHz mod.
nance signal) or a continuous switchover
V 312H V (in the colour channel by means of a non-
colour subcarrier linear characteristic) takes place between
time function spectrum b)
0 to 3.1 MHz
Y YL
– transmission of the averaged high-fre-
(CVS) ∆τ quency luminance signal component
chrominance Y
Y
B–Y
signal F
+F
YH and the averaged colour difference
312H YH signals U and V and
R–Y colour subcarrier
1st field
phase 0˚ – elimination of the averaged high-fre-
3.1 to 5 MHz
Y+F U
quency luminance signal components
(CCVS)
3.1 to 5 MHz 312H and transmission of the original non-
Y PAL
de-
averaged colour difference signals as
Y
+F mod.
required by the picture motion [7].
(CVS) 312H V

chrominance
signal F colour subcarrier The picture motion on the screen is deter-
B–Y
mined by a pixel-by-pixel comparison. The
R–Y
2nd field
colour subcarrier Fig 77 averaged colour difference signals are
phase 180˚
Y+F
(CCVS)
Principle of colour plus encoder (top) and applied to a motion detector providing
colour plus decoder (bottom). the control signals for fading in and out in
f the luminance and chrominance chan-
Averaging in the encoder reduces the nels. Fig 78 shows the principle of lumi-
Fig 76 vertical temporal resolution of the high- nance signal and colour difference signal
Spectral components of PAL CCVS with phase frequency luminance signal components processing in the MACP coder. Digital
assignment in 1st and 2nd field. and colour difference signals. signal processing is used for implementa-
tion.
The spectrum of the PAL CCVS with com- The vertical resolution is reduced from
ponents of the (Y)CVS and the F signal 576 to 288 active lines. This is hardly per-
can be shown in magnitude and phase in ceived by the human eye as the eye is
a simplified diagram (Fig 76). The cross- less sensitive to the vertical resolution
talk artefacts within the frequency range than to the horizontal one.
3.1 MHz to 5 MHz are compensated for by
adding or subtracting the two signals

Refresher Topics - Television Technology 31

 upper and lower screen edges (letterbox Reversible line interpolation is performed
format). The picture height is reduced as shown in the simplified diagram of Fig
through vertical filtering and by removing 80. The 16:9 receiver restores the original
0 to 3.1
MHz YL "Y" every fourth line from the total of lines A, B, C, D from lines A’, B’, C’ trans-
∆τ
Y
576 lines visible with PALplus. The mitted in the compatible 432-line picture
remaining 432 lines are pushed together and from lines C in the vertical helper.
YH
IFA causing the picture height to be reduced The coarse vertical subscanning causes
to three quarters of its original size. interference which can however be
largely eliminated by an improved
∆τ
U The luminance information of the method of vertical band splitting into low-
U "U"
removed 144 lines is transmitted in pass and highpass components [9].
IFA
U 72 lines each at the upper and lower
screen edge in such a way that it remains Vertical filtering of the luminance signal
motion
0 to 1.3 MHz detect. invisible on the standard (4:3) receiver. is performed by means of a quadrature
IFA The 16:9 receiver processes the addi- mirror filter (QMF). The frequency range
V
V "V" tional information in the "vertical helpers" to be divided into a lowpass and a high-
V
and, together with the information con- pass band is filtered so that the bands do
Dt
ventionally transmitted in the 432 lines not overlap and a flat frequency response
for the 4:3 receiver, it restores the initial is obtained when the subbands are com-
Fig 78 colour picture with 576 visible lines bined again in the receiver.
Principle of MACP coder. (Fig 79).
After interpolation (up sampling) by the
A motion detector is also required in the factor of three and decimation (down
MACP decoder as the signals in the sampling) by the factor of four, the verti-
receiver are not synchronized to the picture 16 cal low-frequency component of the
frame head and the picture contents are 9 576 active lines
432-line letterbox picture signal is
compared with the aid of corresponding obtained from the low-frequency compo-
fields [8]. nent of the vertical spectrum (referred to
transmission by
letterbox picture lines helper lines 576 active lines). The high-frequency
8.3 Compatible transmission with 72 component of the vertical spectrum is
16:9 aspect ratio 432 + transmitted in the vertical helper signal.
72
The change to the widescreen 16:9
aspect ratio is an important innovation. reproduction of Simpler linear phase filters may be used
432 lines 576 lines
The method used must however ensure 4 16
for vertical filtering of the colour differ-
full compatibility with the existing trans- ence signals since no vertical high-fre-
mission method employing the 4:3 3 9 quency components are obtained in this
aspect ratio. case. As already pointed out, the helper
on
4:3 screen 16:9 screen signals only contain the luminance infor-
Two methods can be used for reproduc- mation.
ing pictures with a 16:9 aspect ratio on a Fig 79
4:3 standard receiver: Transmission of 16:9 picture in letterbox format
with PALplus.
1. Side panel method
A 4:3 section of the 16:9 picture starting
from the picture center or variable (pan
and scan) is transmitted in line with the Fig 80 reproduction in
processing helper 4:3 16:9
existing standard. In addition to the pic- Line interpolation in
letterbox
lines lines receiver receiver
ture information of the side panels, which with letterbox
A A'= A A'= A A
remain invisible on the screen of the con- method.
ventional receiver, is transmitted using B B=-32B'--12 C
B'=-2 B+-13 C B'
special coding for the 16:9 receiver. 3

C C C
C'=-13 C+-2 D C'
2. Letterbox method D
3
D=-32 C'--12 C
The 16:9 picture is reproduced in full
width on the 4:3 receiver. The picture
height has to be reduced however, which
causes black bands to be displayed at the

32 Refresher Topics - Television Technology

With vertical filtering, a distinction is vertical
filter 3.1 MHz
made between frame filtering (film mode) Y144 Yhelper
144 helper
and field filtering (camera mode). Y
proc.
576
Y432 YL colour subcarrier
432 ∑
Y YH Y432
3.1 MHz
AM A Y D
U IFA CCVS
Y144 Yhelper
compr. camera/film
PALplus
C432
V D A
1.3 MHz U432
colour subcarrier 0˚ U432
432 IFA
Vout
PAL
576 mod.
Vin 1MHz f sc U, V V432
IFA V432
432
colour subcarrier
Fig 81 vertical filter 1.3 MHz
motion
Conversion of luminance signal of helper lines. det.

For the vertical helpers to remain invisi- status

cod.
ble, the ancillary information is modu-
lated after filtering and nonlinear precor- Fig 82
rection onto the colour subcarrier by ves- Simplified block diagram of PALplus encoder.
tigial sideband amplitude modulation
with carrier suppression and is then The low-frequency component YL is The status information is also transmit-
transmitted with reduced amplitude directly obtained from the letterbox pic- ted. Data signals inserted in the "active"
about the black level in the 72 lines at the ture signal. The high-frequency compo- part of line 23 by means of
upper and 72 lines at the lower screen nent YH is averaged over two fields (IFA), 14 bi-phase code elements of 1.2 µs dura-
edge (Fig 81). Compatibility checks have depending on the motion in the picture, tion each (Fig 83), inform the receiver in a
shown that the visibility of the vertical or not transmitted at all after a stepwise
helpers in the letterbox bands is suffi- reduction. A single switchover is shown – 1st group on the aspect ratio at
ciently low. in Fig 82 in a simplified form. As des- (4:3, 16:9, or others),
cribed above, the luminance signal from – 2nd group on the signal source and
Signal processing in the PALplus encoder the helper lines is modulated onto the image-improving measures (camera
is fully digital. This means that the com- colour subcarrier in the frequency band 1 mode, film mode; standard PAL, MACP;
ponent signals applied to the encoder MHz to 5 MHz and added to the letterbox helper),
input have to be in digital form. The inter- picture signal (YL and YH). – 3rd group on the type of subtitles (none
face to the input is the digital studio sig- or subtitles in the teletext, subtitles
nal which will be dealt with in the next The colour difference signals U and V are within or outside the active picture),
chapter. processed in parallel. The information – 4th group on miscellaneous
contained in the letterbox picture lines (e.g. surround sound).
The signal symbols of the PAL standard alone is transmitted −again as a function
are used in the simplified block diagram of the picture motion −as an average Since the transmission of this information
of the PALplus encoder (Fig 82). value of two consecutive lines (IFA) or must be highly immune to interference,
temporarily in the unfiltered original sig- the bi-phase code is used, and with this
The lines in the luminance signal path are nals. aspect ratio the 3-bit code word is addi-
first separated into the 432 letterbox pic- tionally protected by a parity bit. The 14
ture lines and 144 helper lines using the bi-phase code elements are preceded by
above-mentioned vertical filter (QMF). a run-in code and a start code consisting
of (29 +24) elements of the 5 MHz system
clock.

Refresher Topics - Television Technology 33

A reference burst of the colour subcarrier Signal processing in the receiver must be
with defined amplitude (Vpp = 300 mV) is identical but in inverse form. In a stand-
also transmitted in line 23. Together with 1.0
ard receiver configuration, two signal
status
the white reference bar (700 mV) in line V information processors (serial video processors) are
0.8
623 it guarantees correct voltage assign- used together with a field memory and a
ment in the helper signal for the letterbox reference frame memory. Simpler receiver concepts
burst
picture signal (see Fig 83) [7]. do without MACP decoding as some
0.3 interference suppression is already
Since lines 23 and 623 are occupied by achieved during processing of the
data and reference signals, only 430 of 0
PALplus CCVS in the transmitter.
the original 432 letterbox picture lines
remain for image reproduction. This how- Fig 83 No decision has as yet been made regard-
ever has no visible effect on the picture Status bit information and reference burst in line 23. ing the introduction of digital audio signal
displayed . transmission and the transmission of test
signals for echo correction. Further com-
ments on this subject will therefore not
be made.

34 Refresher Topics - Television Technology

9 Digital video studio signal
The deployment of digital techniques for Fig 84
Y
A
N
D
processing effects and characters and for Processing
Y Y
storing video signals on digital tape of digital 74 P
13.5 MHz ns
recorders paved the way for digital stu- serial video
dios. The introduction of PALplus made it studio
B−Y
A
N MUX N
rbit

D
necessary to apply a digital source signal signals. N = 8: 216 Mbit/s
N = 10: 270 Mbit/s
CB
to the coder to provide all the prerequi- 148 ns
6.75 MHz S
sites required for digital TV in the studio. C Y CR Y

The standardized digital video studio sig- R−Y

A
N
D 148 ns
nal DSC 270 Mbit/s can be applied to
CR
both the PALplus encoder and the 148 ns
6.75 MHz 27 MHz N · 27 MHz
MPEG2 encoder for further processing.

As early as in 1982, the digital studio The samples are encoded with 10 bits where the digital signal is available at
standard CCIR 601 valid worldwide for per sample at equally spaced quantiza- N = 10 parallel outputs for the sampling
525-line and 625-line systems was issued tion levels. period TS. This period is TS,Y = 74 ns for
in the form of a recommendation. At the the luminance signal Y and TS,B −Y
end of 1992, CCIR (Comité Consultatif The ITU-R BT.601 standard also allows for = TS,R−Y = 148 ns for the two colour dif-
International des Radiocommunications) 8 bit encoding which however causes vis- ference signals.
became the Radiocommunication Sector ible quantization noise in critical images.
(RS) of the International Telcommunica- The signals to be encoded are converted A multiplexer controlled by the 27 MHz
tion Union (ITU). Since then the digital to a uniform voltage range of Vpp = 1 V clock (T = 37 ns) combines the code
studio standard is available as Rec. ITU-R which corresponds to the picture signal words in the serial sequence CB-Y-CR-Y
BT.601 under the title Encoding Parame- level for white in the Y signal for which on N = 10 parallel lines. Subsequently a
ters of Digital Television for Studios. the following holds parallel-serial conversion of the multiplex
signal is performed with the 270 MHz bit
A distinction is made between the 4:2:2 Y = 0.30 x R + 0.59 x G + 0.11 x B (29) clock, yielding a bit rate of
and the 4:4:4 sampling schemes. The rbit = 270 Mbit/s for the digital serial
4:2:2 scheme is used in the following In the case of ±0.5 V colour difference sig- video components signal DSC 270
description. The 4:2:2 and 4:2:0 nals it corresponds to the maximum value (Fig 84).
schemes will be dealt with later on. of the standard colour bar with the new
colour difference signals reduced to the rbit = 13.5 x 106 1/s x 10 bit
The analog source signals Y, B −Y and following reference:
+ 2 x 6.75 x 106 1/s x 10 bit
R −Y, which are limited in a lowpass filter
to the 5.75 MHz (Y) and 2.75 MHz (B −Y, CB = 0.56 x (B−Y) = 270 Mbit/s (32)
R −Y) band, undergo analog/digital con- = −0.17 x R −0.33 x G + 0.50 x B (30)
version using the principle of pulse code The active video signal contained
modulation (PCM). In compliance with CR = 0.71 x (R−Y) therein with 576 active lines and 720 pix-
the sampling theorem by Shannon, the = +0.50 x R −0.42 x G −0.08 x B (31) els per active line has a bit rate of
sampling frequency fS must be at least
twice the highest signal frequency fsig max The quantization range of 0 to +1 V for the rbit = 720 x 576 x 10 bit x 1/40 ms
in the sampled signal. Y signal with 10 bit encoding and the
+ 2 x 360 x 576 x 10 bit x 1/40 ms
quantization range −0.5 V to +0.5 V are
The sampling frequency for the Y signal is assigned the quantization interval num- = 207.360 Mbit/s (33)
fS,Y = 13.5 MHz with 720 samples in the bers 64 to 940. Thus a small working mar-
active part of each line (53.33 µs digital gin remains. The code words for the With 8 bit encoding a value of
video signal), and fS,B−Y = fS,R−Y = quantization interval numbers 0 and 1023 165.888 Mbit/s would be obtained.
= 6.75 MHz for the colour difference sig- are excluded.
nals with 360 samples in the active part The bit rate for a digital serial HDTV stu-
of each line. Analog-digital conversion is performed dio signal, for which an international
separately for the Y, B −Y and R −Y sig- standard does not exist because no
nals. Fast parallel converters are used agreement could be reached between

Refresher Topics - Television Technology 35

the countries using 525 lines and those specifying the signal source and the The logic of the sampling frequency ratios
using 625 lines, is calculated with the aspect ratio. In contrast to the analog does not hold for the 4:2:0 scheme.
8 bit encoding employed in HDTV in the CCVS, the digital serial components sig- Starting with a reduction of the horizontal
case of nal DSC 270 also contains accompanying resolution of the colour difference signals
audio signals (Fig 85). in the 4:2:2 scheme, the vertical resolu-
– progressive scanning using 1152 tion is halved in a next step. This is per-
active lines, 50 frames/s, 1920 samples formed by averaging the superimposed
per active line for the Y signal and chrominance samples of two consecutive
Y DSC 270
960 samples for each colour difference A EAV SAV lines and assigning the averaged value to
B−Y digital
signal as video
623 624 22 the line pair symmetrically to four lumi-
R−Y D
nance samples. Fig 87 shows the constel-
rbit = 1920 x 1152 x 8 bit x 1/20 ms lation of the Y and CB−CR samples.
+ 2 x 960 x 1152 x 8 bit x 1/20 ms audio A digital
D audio Based on equation (33) and 8 bit encod-
= 1 769.472 Mbit/s (34)
ing, the bit rate of the active video signal
Fig 85 using the 4:2:0 sampling scheme and the
– interlaced scanning using 1152 active Sync signals (timing reference signals SAV, EAV) vertically reduced CB and CR sampling
lines, 50 fields/s, 1920 samples per and digital audio signals in digital video studio points is
active line for the Y signal and signal.
960 samples for each colour difference rbit = 720 x 576 x 8 bit x 1/40 ms
signal as The ITU-R Recommendations 601 and 656
+ 2 x 360 x 288 x 8 bit x 1/40 ms
refer to the 4:2:2 sampling scheme. It
should be pointed out in this context that = 124.416 Mbit/s
rbit = 1920 x 576 x 8 bit x 1/20 ms
the 4:4:4 scheme is used as a basis. The
+ 2 x 960 x 576 x 8 bit x 1/20 ms
values given here define the sampling in comparison with the 165.888 Mbit/s
= 884.736 Mbit/s. (35) frequencies of the R-G-B or obtained with the 4:2:2 scheme.
Y-CB-CR signals with 4 x 3.375 MHz =
The HDTV 1440 standard to be used for = 13.5 MHz each. The sampling ratio is
transmission using therefore 4:4:4. 1st

– 1152 active lines, 50 fields/s and 2nd

1140 samples per active line for the Y 1st

signal and 720 samples for each colour

1st
difference signal 2nd

yields a bit rate of 2nd

1st
field

rbit = 663.552 Mbit/s. (36) Y sample

CB, CR sample, calculated

2nd

for the active HDTV video signal. field

Y sample Fig 87
The digital standard signal components CB, CR sample 4:2:0 sampling scheme.
are combined to formthe digital serial
components signal DSC 270 Mbit/s as Fig 86 The 4:2:0 scheme is often used as a first
defined in Recommendation ITU-R Constellation of Y, CB and CR sampling points with step for the data compression described
BT.656 (Interfaces for Digital Compo- 4:2:2 sampling scheme. in the next chapter.
nents Video Signals in 525-Line and 625-
Line Television Systems operating at the Since the resolution of the colour differ-
4:2:2 Level of Recommendation ITU-R ence signals may be lower than that of
BT.601) the luminance signal, the number of sam-
ples of the colour difference signals CB
This recommendation also describes the and CR and therefore the sampling fre-
digital sync signals start of active video quency can be reduced to half the value
(SAV) and end of active video (EAV) (tim- of the luminance signal, i.e. to 6.75 MHz.
ing reference signals) as well as the dig- In this case the sampling ratio is 4:2:2.
ital blanking signals, which are mostly This yields the 4:2:2 picture sampling
replaced by ancillary data signals that scheme shown in Fig 86 with samples for
may also come in the form of digital the luminance signal Y and the colour dif-
accompanying audio signals, e.g. for ference signals CB and CR.

36 Refresher Topics - Television Technology

10 Data compression techniques
For the transmission of digital video stu- 10.1 Redundancy reduction for the the decoder in the receiver for restoring
dio signals in existing broadband chan- video signal the original PCM signals and to a decoder
nels to and between TV studios and par- Redundancy reduction can be effected in the transmitter for creating the predic-
ticularly for the distribution to viewers, for example by transmitting only the dif- tion signal. The difference value is of
the amount of data to be sent must con- ference of the successive picture pixel course smaller than the quantization
siderably be reduced. This data reduction values by way of differential pulse code range of the input signal because of the
may however not affect the picture qual- modulation (DPCM). statistical distribution of signal values. In
ity. the transmit and receive signal paths the
The signal value transmitted first serves 4 bit DPCM signal is first reconverted to
The bit rate of the digital video source sig- as a predictive value which is then cor- an 8 bit code word signal. To obtain the
nal is reduced by removing redundancy rected with the difference to the actual new prediction or to restore the new sig-
and irrelevant information from the sig- value. The corrected value becomes the nal value in the receiver, which is identi-
nal. new predictive value. Fig 88 gives an cal to the predicted value in the transmit-
example where the 8 bit PCM code words ter, the difference is added to the signal
Redundant information is transmitted in obtained with the sampling clock TS are value supplied last by the predictor.
reduced to 4 bit DPCM code words for the
– blanking intervals together with the difference transmitted. This halves the bit The data quantity to be transmitted
sync signal as well as in rate in the transmission channel. depends on the difference between the
– consecutive pixels and frames current signal value and the predicted
value. The better the prediction the
and can be removed without affecting smaller the difference and the lower the
the picture quality as perceived by the quantization
step number
PCM, 8 bit DPCM, 4 bit amount of data to be transmitted. When
human eye. 255 the difference is small, the restored
1011

1001

0011
10101110

10110001

10110010

10101111

receive signal will correspond better to

Redundancy reduction does not entail 178 +3
+1 the original signal. The aim of data reduc-
any loss in information and is therefore −3 tion is to obtain the best possible predic-
Vsig ∆Vsig
referred to as lossless coding. 174 tion.

Irrelevant information in 0
Video signals of moving pictures contain
Ts Ts a lot of redundancy as there is a signifi-
– details of the picture and in cant degree of commonality between
– the form of high chrominance resolu- Fig 88 successive frames. It seems obvious
tion Transition from pulse code modulation (PCM) therefore that only the difference
to differential pulse code modulation (DPCM). between successive frames is transmit-
can be removed by quantization and by ted. This is done by determining the dif-
using different resolution for the lumi- In practice, a PCM signal is used for gen- ference between two frame sections, i.e.
nance and chrominance signal compo- erating differential pulse code modula- between m × n pixel blocks.
nents without any quality degradation tion (Fig 89). The difference between the
noticeable to the human eye. current signal value and the predictive If the prediction obtained by shifting the
value obtained with the aid of the "pre- pixels of a block of frame n in the horizon-
Irrelevancy reduction involves a loss of dictor" is applied, after quantization with tal and vertical direction optimally
information and is therefore referred to as fewer steps than at the input signal, to matches the corresponding block in the
lossy coding.

The first step to redundancy reduction of PCM DPCM restored

PCM signal
the digital video studio signal is that only Q Q−1
signal signal
the active video signal is processed. Syn- difference quantized
chronization is transmitted by means of a signal
Q−1
difference signal
pred
few code words in the data stream. The
main part of redundancy reduction is per-
formed by utilizing the correlation Fig 89
between neighbouring pixels and partic- Practical implementation of differential
ularly between successive frames. pred pulse code modulation.

Refresher Topics - Television Technology 37

next frame n + 1, the difference between 10.2 Irrelevancy reduction for the The picture is divided into a number of
comparable blocks is very low or video signal blocks of m × n pixels. Each block is inter-
approaches zero. This is referred to as Irrelevant information is reduced by preted by means of a weighted superposi-
motion compensation. transformation coding where details of tion (summing) of basic pictures, which
the picture are transmitted with reduced are transmitted as coefficients in coded
resolution as a result of frequency- form. Starting from the DC component,
section of frame n section of frame n+1
responsive quantization. The video signal which corresponds to the mean bright-
is subjected to a linear, reversible trans- ness or luminance of the mxn pixel block,
formation. The coefficients of the trans- the structure of the basic pictures
formed signal thus obtained are quan- becomes finer the more the frequency
tized and transmitted at a relatively low increases in the vertical and horizontal
data rate. The favoured method of trans- direction. Fig 92 shows a section of the
formation coding of video signals is dis- normally used 8 × 8 basic patterns with
∆y
prediction from crete cosine transform (DCT). the limit value (black-white).
picture section n
∆x with ∆x, ∆y
The two-dimensional DCT of a pixel block
Fx
with subsequent quantization of trans-
Fy
Fig 90 DC form coefficients for irrelevancy reduc-
Example of motion compensation. tion is used as an example for describing
the first stage of another reduction tech-
Fig 90 gives an example of this type of nique (Fig 93).
compensation. The prediction in the block
is obtained by comparing (motion estima- Here coding is effected in an 8 × 8 pixel
tion) the picture content of a block of block the luminance distribution Y(x,y) of
frame n + 1 with the corresponding block which is coded with 8 bits per pixel (value
in frame n and successively shifting the 0 to 255). The luminance distribution in
viewed block of frame n by the values ∆x the block is specified by a weighted
and ∆y, referred to as motion vectors, superposition of the spectral coefficients
until the difference is minimized. F(x,y) onto the 8 × 8 DCT basic function.
Fig 92 In an intermediate step, the luminance
Section of DCT basic patterns. distribution is referenced to the mean
difference
input picture difference signal picture brightness (128). The DCT coefficients are
Q
PCM DPCM
signal signal

Predicted Q−1
picture

luminance distribution Y(x,y) after subtraction of mean value

44 53 62 ... −84 −75 −66 ...
pred pred Y(x,y) = 53 63 74 ... Y'(x,y) = −75 −65 −54 ...
n+1 n 62 74 87 ... −66 −54 −41 ...
notion ... ...
vectors ∆x, ∆y
n+1
∆x, ∆y
DCT coefficient matrix F(x,y) quantization table Q(x,y)
−2.25 −273.99 ... 16 11 10 ...
Fig 91 F(x,y) = −274.85 74.02 ... Q(x,y) = 12 12 14 ...
1.08 −1.21 ... 14 13 16 ...
Generation of DPCM (difference picture) signals
... ...
with motion compensation.

quantized DCT coefficients after rounding

In addition to the picture difference from −0.14 −24.90 ... 0 −25 0 ...
the m × n pixel block obtained by way of F(x,y)
= −22.90 6.16 ... F'(x,y) = −23 6 0 ...
Q(x,y)
DPCM, the two motion vectors ∆x and ∆y 0.07 0.09 ... 0 0 0 ...
... ...
for magnitude and direction are transmit-
ted in short code words to the receiver
(Fig 91), which uses the motion-compen- Fig 93 rounded DCT coefficients matrix F'(x,y)

Transformation of luminance distribution 0 −25 0 −2 ...

sated block for frame reconstruction. This −23 6 0 0 ...
technique considerably reduces the Y(x,y) of a block to the rounded DCT 0 0 0 0 ...
F'(x,y) =
amount of data to be transmitted in the coefficient matrix F’(x,y) and zigzag −2 0 0 0 ...
readout of coefficients. ... ...
signal.

38 Refresher Topics - Television Technology

 0 − 25 − 23 0 6 0 −2 0 0 −2 0 0 0 ... Fig 94 spacing above the spectrum of the ana-
Redundancy reduction log video signal, or as a complete signal
by run length coding in the frequency band 0 to 8.5 MHz with
DC (1,6) (1, − 2) (2 ,− 2 ) RL C (RLC) and variable length frequency modulation in an analog satel-
coding (VLC). lite channel.

Huffman tables VL C
The idea behind source coding with
MUSICAM (masking pattern universal
subband integrated coding and multi-
stored in a matrix. The coefficient in the Discrete cosine transform with quantiza- plexing) is that certain components of the
top lefthand corner represents the DC tion and rounding, zigzag readout of coef- sound level cannot and need not be per-
component. The horizontal frequency ficients as well as run length coding and ceived by the human ear in order to iden-
coefficient Fx increases from left to right, variable length coding result in a data tify the sound. Such components are not
the vertical frequency coefficient Fy from reduction of 10:1 without noticeable loss processed by the MUSICAM encoder.
top to bottom. The higher the frequency, in picture quality. Differential pulse code
the lower the eye’s sensitivity for fine modulation with motion compensation This source coding method involves both
structures. brings about a further reduction by a fac- redundancy and irrelevancy reduction.
tor of 4:1. The result is a hardly perceivable loss in
For this reason a frequency-responsive sound quality depending on the degree of
quantization of DCT coefficients is per- The hybrid DCT coder normally used for data reduction and the sensitivity of the
formed using a quantization matrix Q(x,y) data compression of video signals is a test person.
with the result that the quantized coeffi- combination of DPCM with motion com-
cients become very small as the distance pensation using 16× 16 pixel blocks and With MUSICAM, the digital source signal
to the DC component increases in the DCT using 8 × 8 pixel blocks. This is is generated with a data rate of 768 kbit/s
horizontal, vertical and diagonal direc- shown in a simplified block diagram in using a 48 kHz sampling frequency and
tions and are set to zero after rounding of the next chapter. 16 bit coding and then spread to 32 sub-
the quantized coefficients. Thus matrix bands of 750 Hz width giving an audio
F’(x,y) containing quantized and rounded 10.3 MUSICAM for the audio signal frequency bandwidth of 24 kHz. The sam-
DCT coefficients is obtained. Although the source data rate of a stereo pling frequency of each subband is only
audio signal of CD quality is only about 1.5 kHz. Twelve successive sampling val-
Starting from the DC component, the 1.4 Mbit/s and therefore of minor impor- ues of the subbands are combined to a
matrix F’(x,y) is read out in zigzag fashion. tance in comparison with the 100 times block of 8 ms duration, for which the max-
The resulting sequence of coefficients higher data rate of the video signal, the imum signal level is measured and a scale
may be used for further data reduction audio signal is also subjected to data factor with up to 64 steps is derived and
techniques. reduction. The reason is that in sound coded with 6 bits. With volume classes of
broadcasting digital audio signals are approx. 2 dB a dynamic range of about
The zigzag coefficient sequence is fol- transmitted in relatively narrowband 120 dB is obtained. Since the scale fac-
lowed by two-digit run length coding transmission channels. An exception is tors rarely change for short periods of
(RLC), the first number indicating the DAB (digital audio broadcasting) where time, variable length coding is an appro-
length of the continuous string of zeroes transmission is based on an OFDM signal priate technique to archieve further data
preceding the value equalling the second (see chapter 12). reduction.
number. Run length coding is completed
when only zeroes are left in the string, Digital audio signal transmission is widely As can be seen from Fig 95, an exact anal-
which occurs soon with most of the pic- used in the ADR system (ASTRA digital ysis of the entire spectrum by means of
tures. radio) where the digital audio signals are FFT (fast Fourier transform) is performed
broadcast on subcarriers with 180 kHz in parallel to band splitting to define the
The value pairs thus obtained are then
subjected to variable length coding
(VLC) with the aid of Huffman tables, digital
audio filter data
where frequently recurring value pairs signal bank reduction
are assigned a few bits only, and infre-
quent ones are transmitted with longer scale
factor
data
reduction
MUX
code words (Fig 94). Fig 95
Principle of masking dyn. bit
FFT threshold assignment
MUSICAM coding

Refresher Topics - Television Technology 39

masking threshold below which a tone is bines these substreams to generate an pression techniques form the basis for
no longer perceived by the human ear. output data stream with a bit rate of the following description of digital TV.
Quantization in the 32 subbands is con- 192 kbit/s, 128 kbit/s, 96 kbit/s or less per
trolled by a psycho-acoustic model simu- audio channel. Digital TV involves the emission and
lating the masking characteristics of the broadcasting of TV signals processed in
human ear. Other procedures will be described in the digital form as a bit sequence in the stu-
next chapter on audio coding to the dio and transmitted to the viewers with
Through dynamic bit assignment, the MPEG2 standard. the aid of suitable carrier modulation
constant output bit rate is split up into methods. The next chapter therefore
quantized signals in the subbands, scale The preceding chapters dealing with the describes processing of the data stream
factors, bit error correction and ancillary digital video studio signal and data com- applied to the transmission channel.
information, if any. The multiplexer com-

40 Refresher Topics - Television Technology

11 Video and audio coding to MPEG2 standard
In 1988, a working group was formed by The main features of the MPEG2 standard 11.1 Definition of profiles and levels in
the Joint Technical Committee (JTC) of relevant for digital TV are described MPEG2 video
the International Standards Organization below. In contrast to the digital studio To allow a systematic classification of the
(ISO) and the International Electrotechni- standard, the MPEG1 and MPEG2 stan- great variety of video coding methods to
cal Commission (IEC) under the name dards do not define parameter sets but MPEG2, criteria were defined for applica-
tool boxes as defined by MPEG2 in tions under profiles, and for quality
Moving Pictures Experts Group, under levels.
in short MPEG. Systems ISO/IEC 13818-1
Video ISO/IEC 13818-2 Definition of profiles:
The aim of this group was to present a Audio ISO/IEC 13818-3
standard for coding methods and algo- – Simple profile (SP) uses coding to the
rithms for video and audio signals. Stan- for creating optimized parameter sets for 4:2:0 sampling scheme and simple
dardization was to be implemented in different applications. This ensures opti- prediction without motion compensa-
several phases and has been concluded mum compatibility between the different tion.
by now. quality levels [10]. – Main profile (MP) also uses the 4:2:0
sampling scheme and is employed for
In the first phase, source coding for The MPEG data stream is characterized most applications.
storing video and audio data on CD- by two levels: – SNR scalable profile (SNRP) uses
ROM in multimedia systems was scalability of quantization in the blocks.
defined in the MPEG1 standard. The – The system level comprising bit organ- It permits the reproduction quality to
characteristic features of this standard ization, frame structure and other infor- be fixed depending on the bit error rate
are a maximum bit rate of 1.5 Mbit/s, mation required for demultiplexing the without a total failure of the picture.
frame processing and three quality video and audio data stream and for – Spatial scalable profile (SSP) uses
classes for mono and stereo audio sig- synchronizing video and audio signals spatial scalability with the picture reso-
nals. during reproduction. lution being changed as a function of
– The compression level containing the the transmission quality. SNR scalabil-
The MPEG2 standard is based on the compressed video and audio data ity is included.
MPEG1 standard but intended for stan- stream. – High profile (HP) uses coding to the
dard TV signals with data rates from 4:2:2 sampling scheme and complies
2 Mbit/s to 15 Mbit/s and HDTV signals with the most stringent demands but
from 16 Mbit/s to 40 Mbit/s. In a special represents the most complex solution.
case an upper limit of 20 Mbit/s or even
100 Mbit/s is possible. The MPEG2
standard processes video signals in the
Profiles
field and frame format. Audio coding was
Levels Maximum
extended to multitone with up to five Simple Main SNR Spatially High
bit rate
channels.
scalable scalable

The MPEG3 standard originally intended 80 Mbit/s HP@HL

High MP@HL
for HDTV signals does not exist because (100 Mbit/s)*) (x)*)
source coding of HDTV signals is covered
High 60 Mbit/s HP@H14L
by the MPEG2 standard. MP@H14L SSP@H14L
1440 (80 Mbit/s)*) (x)*)

The MPEG4 standard is being prepared 15 Mbit/s HP@ML

Main SP@ML MP@ML SNRP@ML
at present with a view to defining very (20 Mbit/s)*) (x)*)
low bit rate coding in the kbit/s range for
Low 4 Mbit/s MP@LL SNRP@LL
multimedia applications.

Profile/level combinations for video; *): higher bit rates with double chrominance resolution.

Refresher Topics - Television Technology 41

 video sequence The sequence layer represents the high-
group of pictures est level where the video signal sequence
can be accessed, mostly for inserting pro-
grams. It contains information on the
picture block
aspect ratio, the number of vertical and
slice macroblock
horizontal pixels (the number must be
divisible by 16), the number of pictures
displayed per second and the bit rate
8x8 pixel determining the size of the buffer mem-
ory required for decoding.
Fig 96
Layers of MPEG2 video data stream. The group of pictures layer describes a
group of frames taken from three differ-
Levels define maximum parameters for – SDTV (standard definition television): ent picture types:
picture resolution and refresh rate. The PAL quality of video signal, with
following levels are available: MP@ML, bit rate about 3 Mbit/s to – Intra-frame coded pictures (I frames)
6 Mbit/s. coded with DCT without prediction in
– Low level (LL) for coding TV pictures – EDTV (enhanced definition television): the frame or field.
with reduced resolution, similar to the approximately studio quality, with – Forward predicted pictures (P frames)
MPEG1 SIF (standard image format) HP@ML, bit rate about 6 Mbit/s to predicted with reference to a previous
with 352 × 288 pixels and 25 frames/s 8 Mbit/s. intra-frame-coded picture with motion
or 352 × 240 pixels and 30 frames/s for – HDTV (high definition television): high compensation.
luminance resolution. picture resolution, with HP@HL, bit – Bidirectional predicted pictures
– Main level (ML) for coding TV pictures rate about 20 Mbit/s to 30 Mbit/s [11]. (B frames) transmitted with reference
with a standard resolution (SDTV) of to a previous and subsequent frame.
720 × 576 pixels and 25 frames/s or 11.2 Layers of video data stream
720 × 480 pixels and 30 frames/s. The main objective in the definition of To ensure that decoding starts after
– High-1440 level (H14L) for coding coding algorithms is to keep the design of switch-on or after a transmission failure,
HDTV pictures with 1440 × 1152 pixels the decoder as simple as possible. In the I frames have to be transmitted regu-
and 25 frames/s or 1440 × 1080 pixels order to separate the basic elements of larly approximately every 0.5 ms (after
and 30 frames/s. the coding procedure in the data stream 12 frames).
– High level (HL) especially for coding and to allow suitable access mechanisms
HDTV signals using a 16:9 aspect ratio to become effective, a hierarchical struc- P and B frames are sent in a defined
with 1920× 1152 pixels and 25 frames/ ture in the form of six layers has been cre- sequence between the I frames (Fig 97).
s or 1920 × 1080 pixels and 30 frames/ ated for the video data stream. The Compared to the I frame, the quantity of
s. assignment is shown in Fig 96. data transmitted is about one third in the
P frame and about one ninth in the B
The following useful profile/level combi- – Sequence layer (sequence of video frame.
nations have been selected from 20 feasi- signals)
ble combinations. – Group of pictures layer (GOP) The sequence of the frames differs from
– Picture layer (frame or field) the constellation shown in Fig 97 because
The high profile chrominance resolution – Slice layer (slice) a P frame must be received after the I
can be doubled by way of the 4:2:2 sam- – Macroblock layer (macroblock) frame for the restoration of a B frame.
pling scheme (x). The values in brackets – Block layer (block)
apply for the maximum bit rate.

Designations frequently used for specify-

ing the quality levels of TV signals can be
assigned to the following profile/level
combinations: I B B P B B P I

1 2 3 4 5 6 7 ... 13
– LDTV (low definition television): simple
video and audio signal quality, with Fig 97
SP@ML, bit rate about 1.5 Mbit/s to Sequence of intra-frame-coded pictures (I), forward predicted pictures (P)
3 Mbit/s. and bidirectional predicted pictures (B) in a group of pictures.

42 Refresher Topics - Television Technology

The slice layer is responsible for picture signal in
difference signal in
space-time space-frequency
time space-time
synchronization. Its header transmits a domain
domain
entropy coding
Hybrid DCT
defined spot of the picture. A slice is a 16- Y zigzag
filter block 8x8 scan-
line frame section containing a minimum CB CR dec. sorting DCT
quant.
ning
VLC

of 16 pixels (i.e. a macroblock) or the full prediction

difference picture
frame width when the main profile is frame MUX
data data
pre- inv. inv. buffer
used. dictor
mem-
ory DCT Quant.
constant
restored picture bit rate
motion
In the macroblock layer reference is esti- difference of motion vectors VLC
mation
prediction mode
made to the frame section used for
motion compensation. A macroblock for header
gen.
the luminance signal contains four blocks Characteristics of input picture

and, depending on the sampling format

used for the two chrominance signals CB Fig 99
and CR, two blocks in 4:2:2 format or one Principle of MPEG2 video encoder.
block in 4:2:0 format (Fig 98). This constel-
lation has the advantage that motion esti-
mation is only required for the luminance only performed for the luminance signal. The MPEG2 standard Main Profile @
signal and that the motion vectors of the Macroblock or block sorting is required Main Level was selected for the trans-
luminance components can be used for because the necessary number of lines mission of data-compressed video signals
the chrominance signal. has to be present before a block or mac- to viewers in line with the European DVB
roblock can be formed. standard. Depending on the picture con-
tent, the resolution is selected to ensure
CB CR
Hybrid DCT has already been described in that at least PAL quality is achieved. With
Y
section 10.2. The data obtained from DCT, the resolution being limited to 5.1 MHz,
1 2 5 6 the motion vectors and the prediction this yields 544 pixels horizontally. The bit
mode (I, P or B frame) depending on the rate to be transmitted for obtaining PAL
3 4 7 8 picture content are forwarded to the data quality is between about 4 Mbit/s and 5.2
buffer memory via a multiplexer. Depend- Mbit/s [12].
16 x16 8x16 8x16
pixels pixels pixels ing on the occupancy level of the buffer, a
criterion is chosen for selecting the quan- 11.3 Layers of audio data stream
Y
tization used for the DCT. The buffer then Data compression in the audio signal is
CB CR
1 2 outputs a constant bit rate. based on MUSICAM described in sec-
5 6 tion 10.3. This technique has already
3 4 In the MPEG2 decoder essentially the been used by the MPEG1 standard for
8x8 8x8
pixels pixels same procedure is performed in the stereo audio signal coding.
16 x16
pixels reverse order. Motion estimation is not
required in this case as the transmitted Three layers have been defined for audio
Fig 98 motion vectors are used for motion com- data compression which differ in the
Block-to-macroblock assignment with 4:2:2 pensation. degree of compression and the complex-
sampling format (top) and 4:2:0 sampling ity of the circuitry involved.
format (bottom). The stages of data reduction for the video
signal are as follows: The data of a studio stereo audio signal
In the block layer, the blocks (8 × 8 pix- with a source data rate of 2 × 768 kbit/s
els) coded by means of DCT are transmit- DSC 270 Mbit/s >− 10 bit/pixel reduced to =
ted. With I frames the information from 8 bit/pixel >− 216 Mbit/s, 1536 kbit/s can be reduced in:
the original picture is transmitted, with 216 Mbit/s >− active video signal only
P and B frames the difference between >− 166 Mbit/s, Layer I to 384 kbit/s, minimum out-
frames only. 166 Mbit/s with 4:2:2 format >− 4:2:0 lay, used for digital compact
format with 125 Mbit/s, cassette (DCC)
The operating principle of an MPEG2 125 Mbit/s >− DPCM with motion compen-
video encoder is shown in Fig 99 [5]. The sation, approx. factor 4:1, Layer II to 256 kbit/s, medium out-
digital input signal (DSC 270 Mbit/s) is DCT with quantization, zigzag lay, used for digital audio
reduced to the coding format through fil- readout of coefficients and RLC, VLC, broadcasting (DAB), ASTRA
tering and decimation. Luminance and factor 12 >− up to 2.6 Mbit/s digital radio (ADR)... and for
chrominance signals are processed in without a significant loss in signal qual- MPEG2
parallel although motion estimation is ity.

Refresher Topics - Television Technology 43

 Layer III to 128 kbit/s, maximum out- for transmission in the MPEG2 data Since program stream packets are rela-
lay. stream [13,14]. tively long, they can only be transmitted
in a largely error-free environment. Exam-
The MPEG2 standard extends audio sig- 11.4 Packetized program and transport ples of such an environment are program
nal coding to multichannel audio trans- stream production in the studio or storage media.
missions, which is defined by Layer IImc. The video and the audio encodercgener- The longer the packet, the more difficult
ate an elementary stream each which, the resynchronization and restoration of
The following coding methods are thus after packetizing, is available as a pack- erroneous data.
available for the audio signal with etized elementary stream (PES).
MPEG2: A transport stream (TS) is formed for
A PES packet of the video signal contains data transmission over long distances,
– Single channel coding (mono signals) a data-compressed picture. PES packets e.g. for TV signals to be transmitted to the
– Dual channel coding (bilingual mono may be of variable length, the maximum viewer. Since transport streams of up to
signals) length being 216−1 = 65 535 bytes. 40 Mbit/s can be transmitted in a conven-
– Stereo coding (right, left signal) tional TV channel using channel coding in
– Joint stereo coding (stereo, high fre- Each packet starts with a header of fixed line with the digital video broadcasting
quencies mono only) length (6 bytes). The first three bytes sig- (DVB) standard, the transport stream con-
– Multichannel coding (multichannel nal the packet start. The next byte identi- tains several program streams. In order to
sound, 5 channels) fies whether video or audio information is fully utilize the permissible data rate, sev-
transmitted and the remaining two bytes eral transport streams of different pro-
For compatibility with stereo coding, the indicate the packet length. gram providers can be combined to form a
transport stream multiplex.
5 audio channels (L, C, R, LS, RS) A program stream (PS) is formed from
the packetized video and audio data Because of possible interference to be
are matrixed, i.e. linearly combined to form streams (video PES and audio PES) expected during transmission, the pack-
the two together with additional data, e.g. tele- ets should be as short as possible. The
text, and program-specific information transport stream therefore contains con-
stereo signals L0 and R0 (PSI) by adding a common time base, the secutive packets with a fixed length of
program clock reference (PCR). 188 bytes. A packet contains data of only
and the one signal component (video, audio or
Time stamps are inserted into the other data). The transport packet starts
three signals LSW, CW, RSW sequence of PES packets as the system with a 4 byte header leaving 184 bytes for
clock reference (SCR) required by the transmitting information data (payload).
decoder for synchronization (Fig 100).
A transport stream length of 4 bytes +184
bytes = 188 bytes was chosen because
an ATM cell in global communication net-
program 1 Fig 100 works using the asynchronous transfer
video video pack- video Packetized program (PS) mode (ATM) contains 47 bytes (188 bytes
encoder etizing PES
and transport (TS) data stream. = 4 × 47 bytes) and scrambling is based
audio pack- audio
audio
encoder etizing PES pro-
gram

data pack- MUX

(eg teletext) etizing PES
PS

time PCR
base SCR
trans- transport
port stream
TS
program n MUX

video pack- video

video
encoder etizing PES

PS
audio pack- audio
audio
encoder etizing PES pro-
gram

Data pack- MUX

(also PSI) etizing PES

time PCR
base SCR

44 Refresher Topics - Television Technology

on an 8 byte sequence (184 bytes = 23 × program clock reference (PCR) with compensated for by updating the time
8 bytes). reference to a specific program. stamps, i.e. re-stamping [15].
– Relative time stamps for the correct
For program reproduction in the receiver timing of decoded video and audio sig-
the individual elementary streams are nals. They are inserted in the pack-
first decoded and then synchronously etized elementary stream as decoding
output as video and audio signals. To time stamps (DTS) and presentation
allow synchronization, time stamps are time stamps (PTS). These stamps are
transmitted in the transport stream. indispensable because the decoding
There are two types of time stamps: and reproduction time is not identical
in the case of bidirectional prediction.
– Absolute time stamps derived from
the system time clock for synchroniz- The system clock of the receiver delayed
ing the system time base in the with respect to that of the transmitter due
receiver are inserted in the program to the transmission channel can be
stream as system clock reference restored from the transmitted system
(SCR) and in the transport stream as time clock. Delay time variations can be

Refresher Topics - Television Technology 45

12 Transmission of DVB signal
In digital TV, the terms MPEG2 and DVB 12.1 Error correction After energy dispersal,
are not always clearly distinguished. Due to the complexity of the transmitted
Processing of the digital transport stream data, one faulty bit in the transport channel encoding
as described above is in line with the stream may cause a complete picture fail-
ure. So, effective error correction meth- is performed in the form of concatenated
MPEG2 standard and defined worldwide ods are required in addition to channel- error protection. To optimize error protec-
by ISO and IEC in the standards matched carrier modulation. The aim of tion efficiency, outer coding based on
error correction is to reduce picture fail- the Reed-Solomon block code
– ISO/IEC 13818-1 (multiplex), ures to no more than one a day, which (RS204,188) plus inner coding by means
– ISO/IEC 13818-2 (video coding) corresponds to a permissible BER of max. of punctured convolutional coding are
and 10−11. used. The symbols obtained after outer
– ISO/IEC 13818-3 (audio coding) coding are reorganized in an interleaver
The data stream transmitted in the chan- (Fig 102) before inner coding is per-
With European digital video broadcast- nel should show an evenly distributed formed. Interleaving allows long burst
ing (DVB), the digital transport stream is power density spectrum. To ensure this, errors to be corrected in addition to single
transmitted in the TV broadcasting chan- periodically recurring bit patterns causing bit errors and short burst errors.
nels as stipulated by the guidelines of the discrete spectral lines should be avoided.
European Telecommunications Stand-
ards Institute (ETSI) in the standards The transport stream is therefore first
subjected to energy dispersal. The data outer inter-
leaver
inner
coding
coding
– ETS 300 421, DVB-S, satellite transmis- stream −without the sync byte at the
sion beginning of each 188 byte packet −is RS 204,188 punctured
convolutional coding
– ETS 300 429, DVB-C, cable transmis- linked bitwise to the data stream of a
sion pseudo-random binary sequence Fig 102
– ETS 300 744, DVB-T, terrestrial trans- (PRBS) generator using an exclusive OR Concatenated error protection with interleaving.
mission element (Fig 101).
The Reed-Solomon error correction code
In all three cases, the channel used for The PRBS generator is initialized after adds a sequence of 16 redundant bytes to
signal transmission is affected by inter- every eighth 188 byte packet with the aid each 188 byte transport packet so that
ference so that appropriate measures of a fixed bit pattern. The beginning of the packet protected by the RS204,188
have to be taken for the transport stream every eighth 188 byte packet in the data code has a total length of 204 bytes (Fig
to arrive at the MPEG2 decoder with a stream is marked by an inversion of the 103). This allows correction of up to 8
minimum of errors. Scrambling of data for currently present sync byte. This informs faulty bytes in the packet.
pay TV will not be dealt with in the follow- the receiver that the random data scram-
ing. bling following the known pseudo-ran-
dom bit pattern of the transmitter should SYNC info data redundancy
be removed. 1 byte 187 byte 16 byte
188
204

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Fig 103
Transport packet with Reed-Solomon error
protection.

=1
As specified by the DVB standard, Reed-
Solomon encoding is followed by inter-
leaving with a depth of l = 12. This
random data
& =1 means that the first bytes of 12 succes-
SYNC or
release
data SYNC not scrambled
sive transport packets are transmitted in
succession, then the second bytes and so
Fig 101
Scrambling of transport stream for energy dispersal.

46 Refresher Topics - Television Technology

 X1 X2 X3 X4 X1 X2 X3 X1 Y2
ter has a square-root raised cosine roll-off
convo- punc- re-
CI characteristic with a factor of r = 0.35.
1 2 3 4 lutional turing
encoder R = 3/ 4
sorting
CQ The Nyquist frequency is approx. 14 MHz.
Y1 Y2 Y3 Y4 Y1 Y2 Y3 Y1 X3
The 4PSK modulation product is pro-
Fig 104 cessed in an IF range at 70 MHz or
Convolutional coding with puncturing. 140 MHz. Fig 106 shows the block dia-
gram of the transmitting end of a satellite
channel in the 11 GHz or 12 GHz range.
on. Neighbouring bytes of the original bit assignment to the vector positions
data stream are thus spaced by at least because the absolute phase required for The symbol rate of 27.5 Mbaud transmit-
205 bytes in the output data stream. carrier recovery in the receiver can be set ted with 4PSK via a 33 MHz transponder
from the evaluation of the transmitted corresponds to a bit rate of 55 Mbit/s.
Convolutional coding is a powerful sync byte. The bit-vector mapping with Taking into account the Reed-Solomon
method of error correction but doubles reference to the in-phase component (I) code (RS204,188) and the punctured con-
the input data rate. A code rate of R =½ and the quadrature component (Q) of the volutional code with an assumed code
is thus obtained. The code rate can be carrier signal in line with the Gray code rate of R = 3/4, a useful net data rate of
increased by way of puncturing but this takes into account that only one bit of the RU = 55 Mbit/s x 3/4 x 188/204
reduces the error correction capabilities. two-bit combination is changed between = 38.015 Mbit/s is obtained for the trans-
Using a code rate of R = 3/4 as an exam- two vector positions (Fig 105). port stream.
ple, Fig 104 shows the conversion of the
serial input data stream to the parallel A carrier-to-noise ratio (C/N) ≥7 dB is
output streams required for further trans- I: 1 Q I: 0 needed for a BER of ≤10−11 required at the
Q: 0 Q: 0
mission by modulating a 0° and 90° car- demodulator input.
rier component.
12.3 Cable channel
Error correction in the receiver is per- Transmission in the cable channel is fairly
formed with the aid of a Viterby decoder. I undisturbed compared to that in the sat-
To analyze the redundant data streams ellite channel. This allows a higher carrier
with the aid of the Viterby algorithm, a modulation level to be selected for trans-
great variety of possible errors is calcu- mitting about the same data stream as in
I: 1 I: 0
lated with the aid of a Trellis diagram. Q: 1 Q: 1 the satellite channel at the narrower
The result with the greatest probability is bandwidth of the cable channel. Elabo-
taken as the decoded value. Fig 105 rate error correction as provided by punc-
Vector diagram for 4PSK with Gray coding. tured convolutional coding is not needed
A bit error rate of less than 2 x 10−4 is either.
required for error correction by means of The data signals cI and cQ obtained after
the Reed-Solomon block code RS204,188. convolutional coding and subsequent The data signal arriving as a DVB trans-
This means that the Viterby decoder has puncturing are pulse-shaped in the base- port stream with outer coding based on
to reduce the maximum permissible BER band filter which, in conjunction with the the RS204,188 code and byte interleaving
of approx. 10−2 to 2 x 10−4. After evalua- input and output bandpass filter in the or coming from a satellite receiver station
tion of the outer error correction code satellite transponder, limits the spectrum after evaluation and removal of the inner
(RS204,188), a quasi error-free data sig- of the 4PSK modulation product to a max- coding is subjected to byte-to-symbol
nal with a BER of approx. 10−11 can be imum permissible range. In most cases a mapping. With the 64-order quadrature
obtained. transponder bandwidth (−3 dB) of 33 MHz modulation (64QAM) used for the cable
is used as a reference. The baseband fil- channel, the bytes are mapped into 6-bit
12.2 Satellite channel
The transport stream containing error
correction information is transmitted in
the satellite channel with 4PSK modula- Data
base-
band energy inner to RF
dispersal outer coding satellite
tion which is optimal regarding interfer- inter- coding
convo-
lutional base- 4PSK channel,
face and punctured band modu-
ence and bandwidth. The DVB system RS
inter-
leaver convo- filtering lator transmitter
sync sync (204,188) lutional unit
uses absolute phase coding with direct Clock sepa- inversion coding
ration

Fig 106
Block diagram of transmitter for satellite channel.

Refresher Topics - Television Technology 47

symbols. The bits are grouped into two Fig 108
differentially coded, most significant bits, Superposition of original transmitter signal
and four least significant bits. Differential and multipath signals and
coding is used to simplify carrier recovery evaluation of receive signal echo signal 1
in the receiver. A reference carrier of cor- after guard interval.
rect frequency is obtained, the phase of echo signal 2
which has to be referenced to the quad-
rants. The vector ends within the quad- adjacent-channel
signal
rants are assigned to the four least signif-
icant bits so that the same values are
recovered on the I or Q axis when the received signal
axes are rotated by the absolute refer-
ence phase deviating from the transmit- without intersymbol
guard intervall
ted carrier by multiples of 90°. Fig 107 interference
TG ∆t = TS
shows a section of the 64QAM constella-
tion diagram.

ratio considerably higher than the 7 dB nents, which is the basis for hierarchical
value for 4PSK is required at the demodu- modulation (with 16QAM or 64QAM).
Q (IQ)1st quadrant: 00....
lator input. It can be assumed however This ensures that at least an acceptable
that the minimum C/N ratio of 28 dB nec- image is obtained under unfavourable
..1000 ..1001 ..1101 ..1100
essary for obtaining an error-free picture receiving conditions and the picture does
with 64QAM is well met by cable distribu- not fail completely.
..1010 ..1011 ..1111 ..1110 tion systems which must have a carrier-
to-noise ratio of 45 dB for analog TV chan- Multipath propagation is deliberately uti-
..0010 ..0011 ..0111 ..0110
nels. lized by inserting a guard interval into
the data stream for data evaluation to be
12.4 Terrestrial transmission channel started after a period that takes into
..0000 ..0001 ..0101 ..0100 When TV signals are distributed via ter- account the delay of the reflected signals
I restrial transmitters, the complete cover- and allows the receiver to settle (Fig 108).
age of the areas to be supplied is of fore- Thus local or large-area single frequency
most importance. This can be attained by networks (SFN) can also be imple-
Fig 107 setting up a sufficient number of trans- mented.
Section of 64QAM constellation diagram. mitter stations. Using directional instead
of omnidirectional receiving antennas With orthogonal frequency division
The 6-bit symbols are then assigned to improves the gain and reduces undesired multiplexing (OFDM), the data signal is
the I and Q components in the constella- reflections. Multipath reception produc- distributed to a large number of subcarri-
tion coordinates. Pulse shaping is per- ing ghost images and multiple image con- ers. A low data rate with a symbol period
formed by baseband filtering using a tours in the picture cannot always be of ∆t is transmitted on each subcarrier.
square-root raised cosine filter with a roll- avoided however. These echoing effects The spacing ∆f of adjacent carrier fre-
off factor of r = 0.15. The spectrum of the may be very disturbing in digital signal quencies is selected so that orthogonality
64QAM signals thus occupies a band of transmission. is obtained as defined by
7.935 MHz in an 8 MHz cable channel.
Therefore, a multicarrier transmission ∆f = 1/∆t (37)
A bit rate of 41.4 Mbit/s can therefore be technique, which has proved very suc-
transmitted with 64QAM in the 8 MHz cessful in digital audio broadcasting The number N of subcarriers depends on
cable channel. Taking into account the (DAB), is used for DVB via terrestrial the available channel bandwidth BWRF as
RS204,188 encoding, a net bit rate of transmitters. This OFDM (orthogonal fre- given by
38.015 Mbit/s is obtained. Thus transport quency division multiplex) method makes
BW
streams received from a satellite channel use of the 1705 or 6817 modulated carri- N= RF/∆f (38)
can also be transmitted in the 8 MHz ers intended for DVB. It permits 4PSK,
cable channel to the viewer. 16QAM or 64QAM modulation of the indi-
vidual carriers. According to the DVB-T
With 64-order quadrature amplitude standard the data stream can be split up
modulation (64QAM), a carrier-to-noise in high-priority and low-priority compo-

48 Refresher Topics - Television Technology

A frequency-time constellation of subcar- In the receiver, the time function is
riers is thus obtained within the available retransformed to the frequency domain
channel bandwidth BWRF (Fig 109). time domain frequency domain by means of discrete Fourier transform
u Im û (DFT) with assignment of data to the car-
Re
rier positions and reading out of the real
t f and imaginary components of the individ-
t ual positions.
frequency domain time domain

lm Re
Im uu DVB-T uses1705 of a total of 2048 carriers
∆t in the 2k mode and 6817 of 8192 carriers
ff t
in the 8k mode. 1512 (2k) or 6048 (8k) car-
riers contain information data, and the
∆f f
Fig 110 remaining carriers transmit reference sig-
BW RF Principle of Fourier transform (top) nals with information known to the
and inverse Fourier transform (bottom). receiver or are used for transmitting sig-
Fig 109 nalling or sync patterns.
Frequency-time constellation with OFDM. To generate the OFDM signal, the data
signal consisting of the parallel data The symbol period TS in the continuously
The maximum theoretical data rate is streams cI and cQ is successively assigned transmitted information data stream is
to N subcarriers by mapping of m bits at a extended by the guard interval TG. This
rbit,max = m x N x ∆f bit/s (39) time (2 bits with 4PSK, 4 bits with 16QAM, however reduces the transmitted infor-
6 bits with 64QAM), each subcarrier mation data rate. Depending on the maxi-
depending on 2m-level modulation. being described by a vector with real and mum permissible distance between the
imaginary component for the symbol transmitters of a single frequency net-
With the modulation methods used for period ∆t. Within the period ∆t, inverse work, the guard interval may have a
DVB-T (DVB, terrestrial transmission Fourier transform is performed to yield a length of 1/4, 1/8, 1/16 or 1/32 of the sym-
channel), the following values of m apply: continuous time function. This function bol period TS. A guard interval of 1/4 of a
corresponds to the complex envelope of symbol period is required for the largest
– 4-order modulation, 4PSK: m = 2 the OFDM signal in the baseband. The distance between transmitters. The total
– 16-order modulation, 16QAM: m = 4 OFDM signal is then converted to the symbol period
– 64-order modulation, 64QAM: m = 6 center frequency of an IF channel or
directly to the RF. Fig 111 gives a simpli- Ttotal = TS + TG (40)
The permissible number N of subcarriers fied illustration of this procedure using
is determined using the procedure 4PSK for the 1536 subcarriers commonly is thus extended to 1.25 times the symbol
employed for OFDM by way of discrete used in DAB, which are assigned data period and the effective transmitted data
inverse Fourier transform (DIFT) per- from a total of 2048 subcarriers (2k mode). rate reduced to 0.8 times the value deter-
formed on the basis of n x k memory loca- mined by the symbol period.
tions with digital signal processing. The
following two modes can be selected for
DVB-T: t
data signal

– 2k mode with 2 x 1024 = 2048 memory I +1 +1 −1 +1 −1 +1 −1 −1 +1

−1 −1 −1
locations Q +1 −1 +1 +1 +1 +1

– 8k mode with 8 x 1025 = 8192 memory

f
locations. mapping
to 0 1 2 1535 1536 2047 0 1 2 1535 1536
carrier No.

While in the case of Fourier transform a vector +1+j +1−j −1+j +1−j 0 0 −1+j +1−j −1−j −1+j 0
function is transformed from the time to g(f)

the frequency domain, as is shown in the

IFT
top part of Fig 110, inverse Fourier 2k ^
= 2048
transform performs transformation from time
domain
the frequency to the time domain (Fig 110, f(t)
bottom). complex OFDM
t envelope signal

center carrier

Fig 111
Mapping of I/Q data signals to vector positions of individual carriers in frequency domain and inverse
Fourier transform to complex envelope of time domain with OFDM.

Refresher Topics - Television Technology 49

Digital signal processing in OFDM signal Fig 112
energy
generation is based on a sampling clock Block diagram of data dispersal
outer inner
with the period T = 7/64 µs, which corre-
outer inner
transmitter for and error inter- error inter-
protection leaver protection leaver
sponds to a sampling frequency of DVB-T using
clock sync
inversion
approx. 9.14 MHz. Thus the useful symbol non-hierarchical
period is modulation.

TS = 2048 x 7/64 µs = 224 µs in the con-

insertion
2k mode and mapper
and frame of
D/A version to RF
con- to
TS = 8192 x 7/64 µs = 896 µs in the modu- matching OFDM guard
version carrier trans-
lator interval mitter
frequency
8k mode.

Based on orthogonality, the carrier spac- pilots,

TPS-, ref.
ing is therefore symbols

∆f = 1/224 µs = 4.464... kHz in the 2k mode

or
∆f= 1/896 µs =1.116... kHz in the 8k mode. As shown in the simplified block diagram The aim of the DVB system is a high trans-
of Fig 112, processing of the DVB-T data parency of data stream transmission from
A total symbol duration of Ttotal = 896 µs + signal is at first the same as with satellite the satellite channel to the cable channel
224 µs = 1120 µs is obtained in the transmission (DVB-S), i.e. energy disper- or to the terrestrial broadcasting channel.
8k mode to be introduced in Germany, sal, outer error protection based on Cable channel transparency require-
which uses a maximum transmitter dis- RS204,188 code, interleaver and error ments are fully met as the net information
tance of approx. 67 km and a guard inter- protection by punctured convolutional data stream of 38 Mbit/s from the satel-
val of TG = ¼ x 896 µs = 224 µs. The total coding. With DVB-T an inner interleaver lite can be fully fed into the cable chan-
bit rate of the transport stream to be using combined bit and symbol interleav- nel. This is not quite so with the terrestrial
transmitted on the 6048 carriers used for ing follows for distributing longer errors 8 MHz channel where, given the same
information data is therefore of individual or adjacent carriers across total data rate, the transmitted net data
the whole data stream so that errors are rate is lower than in the 8 MHz cable
rbit = 6048 x 4 bit/1120 µs = 21.6 Mbit/s already corrected by the bitwise inner channel as a result of inner error protec-
for 16QAM coding of the Viterby decoder [16]. The tion, guard interval and because only
and data stream of the I and Q channels is 6048 of the 6817 carriers are used. It can
rbit = 6048 x 6 bit/1120 µs = 32.4 Mbit/s mapped to the vector position of the carri- be assumed that a data rate of around 24
for 64QAM. ers transmitting the information signal. Mbit/s is transmitted in the terrestrial
Reference and sync signals are assigned 8 MHz channel.
The total RF bandwidth occupied by the to the pilot carriers, and the total number
OFDM signal up to the first null positions of carriers used is then converted into a This data rate is sufficient for the simulta-
outside the main spectrum as given by complex time function by means of neous transmission of 4 to 8 SDTV pro-
OFDM. Each symbol with the duration TS grams of PAL quality or 3 EDTV programs
1705 x 4.644... kHz = 7.611 MHz is extended by a guard interval TG. Finally of PALplus, i.e. of almost studio quality
for the 2k mode, digital/analog conversion and conversion [17], in comparison with a single analog
and by to the center carrier frequency are carried program.
6817 x 1.116 ...kHz = 7.608 MHz out.
for the 8k mode
The defined mixed distribution of data to
is approx. 7.61 MHz and thus lies within the individual carriers by means of inner
the 8 MHz TV channel. interleaving changes OFDM to COFDM
(coded orthogonal frequency division
multiplex).

50 Refresher Topics - Television Technology

13 References
[1] Theile, R.: Fernsehtechnik, Band 1: [8] Dobler, G.: Von PAL zu PALplus. Verlag [14] Grigat, R.R.; Ibenthal, A.: Audio- und
Grundlagen. Springer Verlag, Berlin - Technik, Berlin, 1996 Videodatenkompression mit MPEG-2.
Heidelberg - New York, 1973 [9] Hentschel, Chr.; Schönfelder, H.: Prob- Funkschau 67 (1995), No. 3, pp 26-33
[2] Mangold, H.: TV trials in Germany with leme des Formatwechsels bei [15] Ziemer, A. (publisher): Digitales Fernse-
vision/sound carrier power ratio of 10:1. zukünftigen verbesserten PAL- hen, 2nd edition. Hüthig Verlag,
News from Rohde&Schwarz(1975) No. Verfahren. Fernseh- und Kino- Heidelberg, 1997
69, p 38 Technik 45 (1991), No. 7, pp 347-355 [16] Reimers, U.: Digitale Fernsehtechnik:
[3] Schönfelder, H.: Fernsehtechnik, parts I [10] Teichner, D.: Der MPEG-2-Standard: Datenkompression und Übertragung für
and II. Lecture text. MPEG-1 and MPEG-2: Universelle Werk- DVB. 2nd edition. Springer Verlag, Berlin
Justus-von-Liebig-Verlag, Darmstadt, zeuge für Digitale Video- und Audio-App- - Heidelberg - New York, 1997
1972 and 1973 likationen (part 1) and Main Profile: Kern [17] Schäfer, R.: Der DVB-T-Standard.
[4] Mayer, N.: Technik des Farbfernsehens des MPEG-2-Video-Standards (part 2). Fernseh- und Kino-Technik 59 (1996),
in Theorie und Praxis. Verlag für Radio-, Fernseh- und Kino-Technik 48 (1994), No. No. 4, pp 153-158
Foto, Kinotechnik, 1967 4, pp 155-163 and No. 5, p 227-237
[5] Mäusl, R.: Fernsehtechnik - Übertra- [11] Freyer, U.: DVB - Digitales Fernsehen. Extracts from the following book were used for
gungsverfahren für Bild, Ton und Daten. Verlag Technik, Berlin, 1997 chapters 1 to 7 of this brochure:
Hüthig Verlag, Heidelberg, 1995 [12] Schäfer, R.: DVB bei den öffentlich-
[6] Kays, R.: Ein Verfahren zur verbesserten rechtlichen Rundfunkanstalten - [18] Mäusl, R.: Fernsehtechnik - Von der
PAL-Codierung und Decodierung. Eine Einführung. Fernseh- und Kino- Kamera zum Bildschirm. Hüthig Buch
Fernseh- und Kino-Technik 44 (1990), Technik 51 (1997), No. 10, pp 620-630 Verlag, Heidelberg, 1991
No. 11, p 595-602 [13] Schröder, E.F.; Spille, J.: Der MPEG-2-
[7] PALplus System Description. PALplus Standard: Audio-Codierung (part 4).
Consortium, 1995 Fernseh- und Kino-Technik 48 (1994),
No. 7-8, pp 364-373

Refresher Topics - Television Technology 51

 0704 (Bi as/ed)
Germany paper
Printedoninchlorine-free
PD 0756.8990.24 ⋅ ⋅Printed as/ed) in Germany
0704⋅ (BiPrinted

ROHDE&SCHWARZ GmbH & Co. KG ⋅ Mühldorfstraße 15 ⋅ 81671 München ⋅ Germany ⋅ P.O.B. 8014 69 ⋅ 81614 München ⋅ Germany
Telephone +49894129-0 ⋅ www.rohde-schwarz.com ⋅ CustomerSupport: Tel. +491805124242, Fax +4989 4129-13777, E-mail: [email protected]