Martin Bruchanov OK2MNM

Image Communication
on Short Waves

www.sstv-handbook.com
1

1
 Contents
1 Preface 7

2 Slow-scan television 9
2.1 The beginnings 9
2.2 Image transmission 10

3 Modern SSTV features 12

3.1 Signal modulation 13
3.1.1 Bandwidth 13
3.1.2 Modulation techniques of analog SSTV 14
3.2 Image resolution 15
3.3 Line speed 17
3.4 Black & white transmission 17
3.5 Colour transmission 20
3.5.1 Additive colour model 20
3.5.2 Composite colour model 21
3.6 Synchronization 23
3.6.1 Horizontal synchronization 23
3.6.2 Vertical synchronization — VIS code 24
3.6.3 Additional synchronization data 26

4 Formats of slow-scan TV transmission 28

4.1 Black and white SSTV systems 28
4.1.1 Modes for digital converters 28
4.1.2 BW transmission with computer software 29
4.2 Color SSTV modes 30
4.2.1 Wraase SC-1 32
4.2.2 Robot color system 32
4.2.3 The Martin synchronous system 35
4.2.4 Scottie 38
4.2.4.1 Scottie DX – special mode for long distance transfers 39
4.2.5 Amiga Video Transceiver 39
4.2.6 Wraase SC-2 42
4.3 High resolution transmission 43
4.3.1 FAX480 44
4.3.2 Pasokon TV 44
4.3.3 PD modes 45

2
4.4 Experimental modes 46
4.4.1 MSCAN TV 47
4.4.2 Kenwood FAST FM 47
4.4.3 Modes MP, MR, ML 48
4.4.4 Martin HQ 50

5 List of SSTV modes 52

6 SSTV Equipment 56
6.1 Transceiver 56
6.2 Station equipment for visual communication 56
6.3 Historical tidbits 57
6.3.1 SSTV monitor 57
6.3.2 Scanning devices 59
6.4 Early FSTV/SSTV converters 60
6.5 SUPERSCAN 2001 61
6.6 Tasco TSC-70P 63
6.7 Interactive Visual Communicator VC-H1 64

7 Computer operations 65
7.1 Hardware configuration 65
7.2 Using a sound card as a modem 66
7.2.1 Sound processing in PCs 66
7.2.1.1 Sampling 66
7.2.2 Analog-to-digital conversion 67
7.2.3 Interface between TRX and PC 69
7.2.4 PTT control 71
7.2.5 Eliminate supply noise 72
7.3 Timing oscillator configuration 72
7.3.1 Transmit timing offset 74
7.4 SSTV tuning 75
7.5 Video digitalization 76
7.6 Software for Windows 76
7.6.1 List of programs 76
7.6.1.1 SSTV software 76
7.6.1.2 Digital mode software with SSTV support 77
7.6.1.3 Software for dedicated interfaces 77

3
8 Ham radio image operations 78
8.1 The reporting system 81
8.2 SSTV not only for hams 81
8.3 Diplomas and QSL cards 82
8.3.1 IVCA DX Achievement Award DXAA 83
8.3.2 DANISH DX SSTV AWARD 83
8.3.3 Russian SSTV Award 83
8.4 Contests 84
8.4.1 DARC SSTV Contest 84
8.4.2 Russian SSTV Contest 84
8.4.3 NVCG SSTV Contest 85
8.4.4 Danish SSTV Contest 85
8.4.5 JASTA SSTV Activity 85
8.4.6 Ukrainian SSTV Contest 86
8.5 SSTV repeaters 86
8.5.1 HF and 50 MHz repeater list 86
8.6 Ham radio satellites and space broadcast 86
8.6.1 SSTV from Mir station 89
8.6.2 SuitSat 90
8.6.3 Amateur Radio on the International Space Station 90

9 Introduction to digital slow-scan TV 93

9.1 Digital communication basics 94
9.2 Error detection and correction 95
9.2.1 Cyclic redundancy check 97
9.2.2 Hamming code 97
9.2.3 Reed-Solomon code 98
9.3 Data compression 98
9.3.1 Information entropy 99
9.3.2 Huffman coding 100
9.3.3 Lossless data compression 102
9.3.3.1 Portable Network Graphics 102
9.3.4 Lossy compression 103
9.3.4.1 JPEG compression 103
9.3.4.2 JPEG2000 107
9.3.4.3 Lossy versus lossless image compression — conclusion 108

4
10 DSSTV transmission systems 111
10.1 Redundant Data File Transfer 111
10.1.1 RDFT operations 114
10.2 HamDRM system 116
10.2.1 Comparison of HamDRM and RDFT 119
10.2.2 Quadrature amplitude modulation — QAM 120
10.2.3 Orthogonal frequency-division multiplexing — OFDM 122
10.2.3.1 OFDM transfer 123
10.3 DSSTV software selection 125
10.4 Making QSO 125
10.5 Waterfall images 127

11 Facsimile — Radiofax 129

11.1 The history of image transmission 129
11.2 The fax mode 130
11.2.1 Image transmission 131
11.2.2 The reception 132
11.2.2.1 Facsimile transmission modes 134
11.3 Professional stations 134
11.4 Satellite imagery retransmission 138
11.4.1 Meteorologic satellites 139
11.4.2 Essential Services 141
11.5 Hamradio facsimile operations 141
11.5.1 EU – FAX – Diplom 142
11.5.2 The International HF – FAX – Contest by DARC 142
11.6 International facsimile standard recommedation 143

12 List of professional stations 146

12.1 Europe 146
12.1.1 Athens, Greece 146
12.1.2 Hamburg/Pinnenberg, Germany 146
12.1.3 Roma, Italy 146
12.1.4 Moscow, Russia 147
12.1.5 Murmansk, Russia 147
12.1.6 Northwood, The United Kingdom 147
12.2 Africa 148
12.2.1 Cape Naval, South Africa 148
12.3 Asia 148
12.3.1 Beijing, China 148
12.3.2 Beijing, China 148
12.3.3 Shanghai, China 149

5
12.3.4 New Delhi, India 149
12.3.5 Tokyo, Japan 149
12.3.6 Pevek, Chukotka peninsula 149
12.3.7 Taipei, China 150
12.3.8 Seoul, Republic of Korea 150
12.3.9 Bangkok, Thailand 150
12.3.10 Kyodo News Agency, Japan 150
12.3.11 Kyodo News Agency, Singapore 151
12.3.12 Northwood, Persian Gulf Base 151
12.4 South America 151
12.4.1 Rio de Janeiro, Brazil 151
12.4.2 Valparaiso Playa Ancha, Chile 151
12.5 North America 152
12.5.1 Halifax, Nova Scotia Canada 152
12.5.2 Iqaluit, NWT Canada 152
12.5.3 Resolute, NWT Canada 152
12.5.4 Sydney, Nova Scotia Kanada 152
12.5.5 Kodiak, Alaska USA 153
12.5.6 Pt. Reyes, California USA 153
12.5.7 New Orleans, Louisiana USA 153
12.5.8 Boston, Massachusetts USA 153
12.5.9 Inuvik, Canada 154
12.6 Australia and Oceania 154
12.6.1 Charleville, Australia 154
12.6.2 Wiluna, Australia 154
12.6.3 Wellington, New Zealand 155
12.6.4 Honolulu, Hawaii USA 155
12.7 List by frequency 156

13 Computer image processing 160

13.1 Image resizing 160
13.2 Color adjustment 161
13.3 Filters 164
13.3.1 Convolution matrix 164
13.3.2 Noise reduction 167
13.3.2.1 Spatial average filtering 167
13.3.2.2 Median filter 169
13.3.3 Sharpening 170

6
 1
Preface

There are various methods used in transferring messages through radio waves. With
miscellaneous communication modes that are suitable for use under different condi-
tions, varying in the speed of transmission, modulation or data protocol.
And many of them are used by radio amateurs to connect worldwide on short
waves, connections over satellite relays or message downloads from local packet-
radio BBS on very short waves. The book you are about to read is about image
transmission.
The most common method to transfer images is television broadcasting (Fast-Scan
Television). An analog FSTV broadcast can be produced in amateur conditions
too. Both picture and sound transmitted on amateur bands can be received via a
regular TV set or by a satellite tuner in the case of frequency modulation. These
connections are held only on ultrashort or microwave bands because the signal needs
a large bandwidth and thus the signal can be transported only through relatively
small distances.
However, the issue of this book is image transmission on short waves.
The most popular narrow-band mode for image transmission is SSTV – Slow-scan
Television. Unlike the classic TV it can broadcast only static images with lower
resolution.
An SSTV image is converted to an audio signal, which can then be transmitted
over a voice channel by a communication transceiver on shortwave bands. With
an expansion of radio broadcast digitalization, digital SSTV was also developed
and uses advanced technologies like data compression, error correction codes and
discrete multi-tone modulations for relatively fast narrow-band data transfer.
Another option for shortwave image transmission is radiofax, the predecessor of
what is now commonly known as office fax. Radiofax is mostly used by meteorolog-
ical stations for the broadcast of synoptic maps and satellite images. Or by press
agencies for the broadcast of news (and photographs in the past) on longwave and
shortwave bands. Synoptic maps should be transferred in high quality, so image
transmission takes about 10 minutes or more on average. Despite the pervasiveness
of Internet technology, this broadcast method is still widely used.

7
1 Preface

For a long time, the integral part of our hamshacks has been the personal com-
puter. An essential part of a PC; the sound card inputs signals into the PC. And
then specialized software converts the signal to data and vice versa. The data that
is of interest in this book will be that of transmitted images.
I hope this book will spur those who are interested in these fascinating modes of
communication to get immediately active in the field.

In Žďář nad Sázavou, November 17, 2019 Martin OK2MNM

8
 2
Slow-scan television

Slow-scan television (SSTV) is a mode of communication designed for image transfer.

Because SSTV is a narrowband mode, it can be transmitted via voice channels
with a standard SSB transceiver on all radio amateur frequency bands. Worldwide
communication is also possible during good conditions on high frequency bands.

2.1 The beginnings

In 1957 a student at the University of Kentucky, Copthorne “Cop” Macdonalds
(*1936 – †2011), WA2BCW (later VY2CM) found an article about a device devel-
oped by Bell laboratories for image transmission via telephone lines. The commu-
nication system fascinated the ham radio enthusiast because it needed a bandwidth
as narrow as that of voice broadcast and could be transferred by regular ham radio
transmitters.
Another image mode the radio fax (facsimile) was available then, but it required
a long duration (about 20 minutes) for a high-resolution image transfer. Such a
duration length could not provide an impression of time consistency during a QSO
and it also needed an intricate mechanical printer and electrosensitive paper. It was
necessary to invent something else.
There was an idea to transfer images coded into audio signals and display them
on long persistence displays (CRTs used in radars or slow-scan oscilloscopes).
Then Copthorne started to work on how to transfer images via radio waves with a
common ham radio transceiver. Within six months he carried out many experiments
with amplitude and frequency modulation, and it resulted in the design of slow-
scan television. During the next six months he created an SSTV image scanner, so
practical experiments could be done on the ham bands. The first television image
crossed the Atlantic on the 20th of December 1959.

9
2 Slow-scan television

The first image that crossed Copthorne

the Atlantic, received by Macdonald’s broadcast.
John Plowman G3AST.
Figure 2.1: Early slow-scan television images.

During the next ten years, Copthorne and a group of amateurs worked on SSTV
improvements and they created the basic standard for SSTV and developed a sam-
pling camera.
Their work was completed in 1968 when the FCC (Federal Communications Com-
mission) formally authorized SSTV operations.
A few months later ham radio magazines published the first articles about the
new communication mode. It led to a huge interest by ham operators and a real
SSTV boom.

2.2 Image transmission

The basic idea of SSTV is to transfer television images with the standard transceiver.
However, a television broadcast requires a large bandwidth. The reduction of the
television signal is achieved by lowering of horizontal (row) and vertical (image)
scans, which must be reduced to a minimal frequency. This means that a typical
3MHz signal of black and white television must be reduced to 3 kHz – the reduction of
bandwidth is around 1000 : 1. Nowadays the bandwidth reduction is bigger because
the color image needs approx. 6 MHz. Therefore, only static images with lower
resolution can be transferred due to the significant bandwidth reduction.
During experiments, it was found that an image was visible for about 8 seconds
on a long-persistent CRT with P7 phosphor. So after reception of the last scan-line,
the first is still visible, but in a while – the image slowly disappears. For the best
impression, it was necessary to view an SSTV monitor in a darkened room. Usually,
several same images were transmitted in a sequence. Each consequent image slowly

10
Slow-scan television 2

redrew the original which was still visible on the phosphor. So it was possible to
display images for a longer time or to record it on a tape for later playback.
It was found that the ideal time for the correct detection of line synchronization
pulses by electronic circuits is 5 ms and for image (vertical) synchronization it is
30 ms. Vertical synchronization initiates the automatic start-up of the image display
on the CRT.
The synchronization frequency for scan-lines and frames was derived from the
electric mains frequency. For horizontal scan 50 Hz divided by three – 16.6 Hz is
used. And for vertical scan 1/7.2 s = 0.1388 Hz is used, this is the mains frequency
divided by 360 (3× number of lines 120). The parameters are derived in the same
way for countries with 60 Hz mains.
The video signal band was chosen in the range from 1500 Hz for black up to
2300 Hz for white. Sync pulses have a frequency of 1200 Hz and because they are
“blacker than black” then they do not affect the image information.
All frequency components of SSTV are inside the low-frequency band and it is
possible to transfer them via voice channels.
Other SSTV modes came out from this original standard and in most cases; they
differ only in scan speed and in the addition of color transmission.

11
 3
Modern SSTV features

A milestone between methods of old and modern SSTV image transmission is with-
out doubt the usage of semiconductor memory chips. The creation of the first con-
verters between fast and slow-scan television signals, was credited to the existence
of permanent image storage in memory. Consequently, image transmission could
be improved because the usage of long persistence CRTs, which had been a ma-
jor constraining fact, was now eliminated. Due to it some new formats with longer
transmission time were developed. They brought more quality to black and white
transmission and helped to develop colour image transmission.
There was a trend in the design of new formats that created several modes in
each system. There were modes with faster transmission and lower resolution and
on the other hand, modes for the transfer of higher quality images but longer time
length. There is a possibility to change between them according to the actual band
condition.
The early phases of development were influenced by two companies – the American
Robot Research Inc. and the German Wraase Electronic led by radio amateur Volker
Wraase, DL2RZ. Each of them introduced an SSTV converter which used each
company’s own transmission system. The systems are different in the usage of
colour coding, scan line formats and synchronization methods. Their converters
provide several modes. Mode denotes a format of image transmission, its resolution
and transfer speed.
As often happens, the professional device did not fully satisfy ham radio users. So
new systems with more modes were implemented into the converter firmware. And
they were also re-implemented into other devices to ensure compatibility. Sometimes
a new genuine system was designed to overcome imperfections found in the classic
predecessors.
The number of those systems has grown unbelievably. Recently they were new
systems created for better utilization of modern computer potentials. Modern per-
sonal computers with the necessary equipment, are full successors of SSTV con-
verters. The advantage of computers is especially bigger memory and better image
resolution.

12
Modern SSTV features 3

If we were to count the number of all SSTV modes, we would find approximately
70! So it is possible to transfer SSTV images via seventy different modes, which
are mutually different in transmission time, resolution, colour coding, etc. The vast
majority of them are absolutely unique and incompatible…
You might be a little scared by the previous paragraphs, but let me reassure you
that only a few modes are actually used.
European amateurs widely used the SSTV mode called Martin M1, but in recent
times other modes; Martin M2 and Scottie S2 are also in use. A special mode used
Scottie DX; is characterized by very high image quality. And the mode Robot 36
Color is undertaken in space communication.
Fortunately, all modern converters and computer software are able to operate with
these popular modes, so the problem that two stations can not establish the QSO
should not occur.
A digital vertical synchronization for automatic mode selection will be described
shortly, because every mode uses a digital header for its identification. Thanks to this
any SSTV device can automatically switch to the correct mode and begin reception.
Computer software also supports mode detection by measuring the elapsed time
between two successive sync. impulses of image lines.
More details will be described in the following chapters.

3.1 Signal modulation

3.1.1 Bandwidth
Different communication channels, whether wired or wireless, have several charac-
teristics, which define their behaviour in the transfer of effective signals. These in-
clude for example attenuation. Attenuation defines how much the communication
channel reduces a transferred signal. Another important characteristic is the bias,
which refers to the various distortions that occurs due to imperfections within the
communication path.
There are several negative influences, that affect signal transfer within a commu-
nication path. Their effects are not negligible. The intensity of this effect depends
also on the frequency of signal. Generally, it is always possible to identify a range
of frequencies that a particular transmission path can transfer well and outside this
frequency range the transmission is too poor.
The signal bandwidth does not depend only on the frequency range used for
modulation, in our case 1,500 Hz to 2,300 Hz, but also on the signal spectrum.
Fourier analysis is used to determine the spectrum bandwidth. The analysis can
express any waveform in the form of the sum of a large number of sine waves –
harmonic components.

13
3 Modern SSTV features

Limited bandwidth has the effect that the harmonic components lying inside this
band will be transferred more or less without blemish and other harmonic compo-
nents pass with a huge distortion or not at all (more in chapter 7.2.1.1, page 66).
Bandwidth can be seen as a characteristic of the transmission path given by the
range of the signal spectrum.
The basic rule for the required bandwidth is called Nyquist rate. Its definition
is that optimal bandwidth equals a half of modulation speed. It is true that the
necessary bandwidth increases with the amount of transferred information per time
unit.

3.1.2 Modulation techniques of analog SSTV

An SSTV broadcast is usually carried out using single-sideband (SSB) amplitude
modulation with a common ham radio transceiver. Frequencies above 2,500 Hz are
strongly suppressed, so the frequency of white colour, the maximal level of an SSTV
signal, was chosen at 2,300 Hz.
SSTV signals are transmitted via frequency modulation of an audio signal. To
avoid any phase shift and drift (which both have negative impact on picture qual-
ity), the spectrum of video signal is modulated on the auxiliary carrier frequency
1900 Hz – sub-carrier. This modulation method is called Sub-carrier frequency mod-
ulation (SFCM).
The frequency of video signal varies from black by gray shades to white. The
bandwidth needed for SSTV transmission varies in the range of 1.0 to 3.2 kHz and
depends on the SSTV mode, transmission speed and also on image content, see
fig. 3.1.
Cheap modems (based on Hamcomm) do not use perfect continuous harmonic
signals, but also create the quantized signal. Step changes between quantization
levels require wider bandwidth, so some image details can get lost.
The emission classification code for the SSTV mode is J3F, which means:

⊳ J – Carrier modulation: Single-sideband with suppressed carrier.

⊳ 3 – Nature of modulating signal: One channel containing analogue information.
⊳ F – Detail of signal: television signals.

In the case of an SSTV transmission via a frequency modulated (FM) channel,

emission is classified as F3F and A3F for amplitude modulation (AM) with both
side bands.

14
Modern SSTV features 3

30
Resolution test
20 Photography
10
0
Amplitude [dB]

−10
−20
−30
−40
−50
−60
−70
0 1000 2000 3000 4000 5000
Frequency f [Hz]

Figure 3.1: The SSTV frequency spectre for two various images trans-
ferred in Martin M1 mode.

50×38 120×90
Figure 3.2: Image quality depends on resolution.

3.2 Image resolution

Resolution is a feature that tells what amount of details is possible to store in an
image, see fig. 3.2. The resolution has two parameters: horizontal resolution – the
number of image columns × number of image lines – the vertical resolution.
In television technology the more important parameter is the vertical resolution
(number of lines) and it is defined by the selection of the SSTV mode. To get the
horizontal resolution is more complicated.
As has been described in previous text, the image is broadcast through the SSB
channel on short waves and the maximal bandwidth is limited.

15
3 Modern SSTV features

The SSTV is an analog mode and cannot transfer images without loss. The image
is not exactly the same on the reception side as on the transmission side. Even
if the communication channel is without any interference or noise, the image is
still distorted due to transmission speed and limited bandwidth . The faster the
transmission speed is, the greater the distortion result. Therefore it is very difficult
to say what the horizontal resolution of an SSTV image is.
Most of the modes carry images with 240 lines and the image is displayed in
a 4 : 3 aspect ratio on a screen. We can then say that the number of columns is
240 × 4/3 = 320. This value then corresponds to a theoretical resolution, but not a
real image resolution.
The test chart (fig. 3.3) is used to qualify the horizontal resolution of images.
The resolution pattern contains alternating stripes of black and white in various
densities from very rough to fine. There is a comparison of this image with normal
photography in fig. 3.1.

Test pattern original

Martin M2(∼220 𝜇s)

Robot 36 Color(∼280 𝜇s)

Martin M1(∼450 𝜇s)

MP115(∼680 𝜇s)

Scottie DX(∼980 𝜇s)

Figure 3.3: Horizontal resolution
comparison for several SSTV modes.

All SSTV modes in figure 3.3 have 320 columns. But as we can see, not all can
transfer the image in actual quality. The note in brackets describes the approximate
time needed for the transfer of one pixel. While with the Martin M2, we can hardly
distinguish the second fine grid, the M1 mode with double transmission time can

16
Modern SSTV features 3

transfer it without problems, but its finest pattern is distorted. Compare it to the
real picture in fig. 3.4. The last two modes listed have longer times of transmission
and can transfer the finest details. Unfortunately, it is hardly compensated for by
the slow speed of transmission.

Martin M1 Martin M2
Figure 3.4: The comparison of two modes in real conditions of 14 MHz band.

3.3 Line speed

One of the most important parameters that is suitable for SSTV mode selection is
the total time required to transfer an image.
Due to present transmission speeds, SSTV is becoming similar to radio facsimile.
Therefore, the mode parameters are not defined by horizontal and vertical scan
rates, but in the number of lines transferred in one minute – lines per minute (lpm).
Line speed depends on the selected mode and varies in the range from 57 lpm
(Scottie DX), for high quality transmission of colour image (320×240) in nearly five
minutes, up to 1000 lpm for BW image (128×128) in just 8 seconds. SSTV modes
and their properties are described below.

3.4 Black & white transmission

For a black and white (BW) monochromatic image broadcast, only one signal is
needed. It represents brightness/luminance 𝑌 of each image element.
The frequency ranges from 1,500 Hz (black) to 2,300 Hz (white) transmit image
information. Each frequency in this range represents specific brightness – the level
of gray.

17
3 Modern SSTV features

Human vision can distinguish brightness in a wide range, but can only adapt to
the geometric mean value of actual brightness. Around this value about 100 to 110
gray scale levels can be differentiated.
Based on this fact; an ideal transmission could be regarded as 128 gray levels. At
this figure, the average observer would not normally see transitions between adjacent
grades.

2300 Hz

1500 Hz

1200 Hz

sync image scan-line

Figure 3.5: The scan line of BW image.

If we want to transmit images in 128 gray levels, this is the distance of signal levels
800 Hz / 128 = 6.25 Hz. The lowest frequency is for black and the highest is for white,
the remaining 126 gray levels lay in the linear range between these two frequencies.
An issue with the transfer of more gray levels, for example 256 levels, is that it
puts an increased demand on the demodulator. The demodulator must be able to
compensate for the frequency shift between the transmitter and the receiver. In this
case, the distance between the two levels of brightness is 3.125 Hz and it is necessary
to have a relatively large distance from the interference on the communication path,
to assure a pure transfer of all gray scale.
Normally, we can settle for a less bright resolution where it is possible to choose
the transfer of only 64 levels. This requires less of the demodulator because it only
needs to distinguish between 12.5 Hz steps.
True reproduction of colour images in gray scale is another issue. Human vision
cannot perceive the bright intensity of all three colour components at the same time.
When we watch three lights (red, green and blue) of the same intensity, the human

18
Modern SSTV features 3

perception considers the green light the brightest. Red and blue are not as bright
in our perception.
But a BW television camera only scans the level of intensity, and therefore the
resulting image would look like all the colours are the same. They will be charac-
terized by the same gray level depending on their intensity. Due to this fact, a valid
gray scale image 𝑌 created from basic colour components 𝑅, 𝐺 and 𝐵 (red, green
and blue) is defined as:
𝑌 = 0.30𝑅 + 0.59𝐺 + 0.11𝐵
Note that the biggest factor 0.59 is just for the green, so nearly 60 % of colours that
we can see depends on the green component and only 40 % is of the remaining
colour components! This is used for simplicity in colour scan converters for BW
images. In past years, BW images were not transmitted as true grayscale images,
but the brightness signal was derived from the green component of the image. The
difference in brightness between a true BW image and the green component of the
same image is insignificant in most cases.

Red component Green component Blue compoment

True colour image True grayscale Intensity

Figure 3.6: Decomposition of colour image to basic components.

19
3 Modern SSTV features

3.5 Colour transmission

3.5.1 Additive colour model

Every colour can be decomposed into three primary colours – red, green and blue.
The additive colour model produces other colours by combining these three primary
colours.
During image transmission, the image is decomposed into these independent
colour components on the transmitting side. Then they are gradually transferred,
and on the receiving side the components are re-composed into a colour image.

ta
en
w

en

ck
te

an
ll o

ag

ue
hi

d
e

a
ye

cy

re
gr

bl

bl
w

2,3

R
1,5
f [kHz]

2,3

G
1,5
2,3

B
1,5

Figure 3.7: Decomposition of

colour image into RGB signals.

If it is possible to detect about 64 frequency levels in the 800 Hz video channel, then
each colour component contains 64 brightness levels. And the resulting colour image
then contains 64 × 64 × 64 = 256144 colours. If a demodulator can distinguish

20
Modern SSTV features 3

256 levels, it is possible to transfer over 16 millions = 2563 colours. Colour SSTV
transmission can meet the most demanding requirements of colour depth.
Some colour SSTV systems also use a property of human vision, which is a different
sensitivity to the primary colour components. In this case; the image scan-lines
are not divided into three equal parts for each colour component. Because the eye
is most sensitive to green, the largest part of the line takes just this part and the
remainder are filled with red and blue parts. For example, the ratio is 4 : 2 : 2 for
𝐺 : 𝑅 : 𝐵.
The additive colour model is a method of transmission that take more time to
transmit, but it provides a transfer of true colours.

3.5.2 Composite colour model

The second type of colour transmission is called YCrCb. In fact it is a similar system
as is used in colour fast-scan television, where each colour component 𝑅, 𝐺 and 𝐵 ,
are transformed to luminance and chrominance (colour information) signals. Unlike
RGB, the transmission time of an image is shorter. This colour coding is used for
BW and colour compatibility in television broadcasts. In which colour broadcasts
can also be received by a BW television.
The image scan-line contains colours transformed into two components – lumi-
nance and chrominance. The chrominance signal is composed of two differential
colour signals 𝑅 − 𝑌 and 𝐵 − 𝑌. Signal 𝑌 is called luminance and contains the signal
corresponding to brightness produced by the equation 𝑌 = 0.30𝑅 + 0.59𝐺 + 0.10𝐵.
The 𝑌 is for chrominance signals subtracted from the red and blue components.
On the receiving side, the individual colour components are restored: 𝑅 = (𝑅 −
𝑌 ) + 𝑌 and 𝐵 = (𝐵 − 𝑌 ) + 𝑌.
We need a third green component the 𝐺, which is derived from 𝑅 − 𝑌 and 𝐵 − 𝑌
from the expression 𝐺 = 𝑌 − 0.51(𝑅 − 𝑌 ) − 0.19(𝐵 − 𝑌 ). Hereby we get complete
colour signals.
There are two formats of YCrCb colour transmission used in SSTV. The first for-
mat 4 : 2 : 2 transmits both chrominance signals (within half the time in comparison
with 𝑌) in one line. The second format 4 : 2 : 0 contains only one chroma signal. Odd
scan-lines could include for instance 𝑅 − 𝑌, and the even scan-lines could be 𝐵 − 𝑌.
The chrominance signal is then given by the average of two consequent scan-lines of
the original image.
The advantage of this type of transfer to RGB is significantly shorter transmission
time. In comparison to RGB transmission, YCrCb takes approximately half the
time yet guarantees almost the same image quality.
Its disadvantage compared with the RGB model, is a loss of image information
which is higher when the 4 : 2 : 0 format is used. Also, precise transceiver tuning
is needed, otherwise the colour information will be distorted. This is the reason

21
3 Modern SSTV features

ta
en
w

en
te

k
an
lo

ag

ue

ac
hi

d
e
l
ye

cy

re
gr

bl

bl
w

2,3
Y

1,5
2,3
R–Y
1,9
f [kHz]

1,5
2,3
G–Y
1,9

1,5

2,3
B–Y
1,9

1,5

Figure 3.8: Decomposition of

colour image into YCrCb signals.

why the YCrCb encoding is used less frequently. According to the positive or neg-
ative deviation from the carrier, the image is strongly hued to pink or green, see
figure 3.9.
The transmission for colour FSTV uses YCrCb and also uses special methods and
modulation (in PAL, SECAM) to eliminate this colour distortion, which can occur

22
 Modern SSTV features 3

+200 Hz
+100 Hz
+50 Hz
0 Hz
–50 Hz
–100 Hz
–200 Hz
Figure 3.9: Color distortion of YCrCb
when the station is improperly tuned.

on the transmission path. Unfortunately, this feature does not exist in SSTV and
so the result of selective fading1 can cause colour ghosts in image.
SSTV systems using YCrCb transmission are less resistant to interference than
their RGB counterparts, see fig. 3.10.

+200 Hz
+100 Hz
+50 Hz
0 Hz
–50 Hz
–100 Hz
–200 Hz
Figure 3.10: Color distortion of RGB
when the station is improperly tuned.

The RGB model is distorted by a low contrast or increased brightness when there is
significant deviation ±200 Hz from the transmitter carrier and thus provides better
colours than YCrCb.

3.6 Synchronization

3.6.1 Horizontal synchronization

There are two types of synchronization – synchronous and asynchronous.

1 Selective fading is a phenomenon, where the signal comes from two paths, in which one signal path
is the variable and causes instability of the ionosphere layers. It can be often seen in the 80 m band
in the morning and evening.

23
3 Modern SSTV features

Older SSTV systems use asynchronous transmission. This means that each infor-
mation frame, in our case a scan-line, will be received after the detection of hori-
zontal sync.
This system detects vertical (image) and horizontal (scan-line) syncs and only
after proper detection will it display the received lines. Asynchronous transmission
has a huge disadvantage. When interference happens close to the 1200 Hz frequency,
an SSTV device can lose several scan-lines if interference remains.
In this respect, all new SSTV systems are improved and use synchronous trans-
mission. These systems use free-run scan. It is not necessary to receive vertical
sync and it is possible to begin reception from any scan-line. After initial synchro-
nization, it is not required to detect horizontal sync. Thanks to this, synchronous
systems are much more resistant to interference. Scan-line sync are still transmitted
and then reception could start any time during transmission.
The disadvantage of free-run scan is in complying to the very precise line speed
of the corresponding parties. The line speed must be absolutely same. If the
values are different, there is an unpleasant effect on the picture – slant. For more
information on this subject see section 7.3.

3.6.2 Vertical synchronization — VIS code

Vertical synchronization is used to detect the start of transmission. The receiving
device can automatically begin the image scan after vertical sync.
The Robot Research company developed a new form of vertical synchronization
called Vertical Interval Signaling – VIS. All modern SSTV systems adopted the VIS
and use these longer syncs and digital headers for automatic SSTV mode recognition.
The VIS contains digital code, the first and last bits are the start and stop bits with
1200 Hz frequency. The remaining 8 bits provide mode identification and contain
one parity bit. Each bit is transmitted in order from the least significant bit.

1900 Hz 1900 Hz

parity
1st image scan-line

1300 Hz 0 0 0 0
start bit bit bit bit bit bit bit bit stop sync
1200 Hz bit 0 1 2 3 4 5 6 7 bit

1100 Hz 1 1 1 1
300 ms 10 300 ms 10 × 30 ms

Figure 3.11: Structure of VIS with value 42.

24
Modern SSTV features 3

Parity is used for simple error checking. SSTV use even parity. This means, that
the number of logical ones must be even in the whole 8bit code. If the number of
ones in 7 bits is odd, then the parity bit is set to one. If the number is even, the
parity bit is zero. Since the information part of code has 7 bits, it takes 128 values.
Each bit is 30 ms long, so the modulation speed is 33.3 bauds. The frequency
1300 Hz means the state of logical zero and 1100 Hz logical one. The first half of
code (least significant bits, LSB) specifies the type of mode (BW/colour, resolution).
The second half (most significant bits, MSB) contains information about the system
(Robot, Martin, AVT,…). The last bit is reserved for parity error checking.

MSB LSB Meaning

P 6 5 4 3 2 1 0
0 0 Color composite video

0 1 BW, red component

1 0 BW, green component

1 1 BW, blue component

0 Horiz. resolution 128 / 160 pixels

1 Horiz. resolution 256 / 320 pixels

0 Vertical resolution 128 / 120 lines

1 Vertical resolution 256 / 240 lines

0 0 0 Robot

0 0 1 Wraase SC-1

0 1 0 Scottie, Wraase SC-2

0 1 1 Scottie, Wraase SC-2

1 0 0 AVT, Scottie DX

1 0 1 AVT, PD

1 1 0 PD

1 1 1 Pasokon TV

X Parity bit

Table 3.1: The meaning of bits in VIS code.

The meaning of bits table 3.1 is valid for a system based on Robot Research
standard. As the number of new modes have expanded, the bit combination has no
additional meaning.

25
3 Modern SSTV features

Mode decimal hexa. binary

Martin M1 44 0x2C 0101100

Martin M2 40 0x28 0101000

Robot 36 colour 8 0x08 0001000

Robot 72 colour 12 0x0C 0001100

Scottie S1 60 0x3C 0111100

Scottie S2 56 0x38 0111000

Scottie DX 76 0x4C 1001100

Wraase SC-2 180 55 0x37 0110111

Table 3.2: The VIS

codes of popular modes.

The comprehensive table of all VIS code is on page 52.

2300
Frequency f [Hz]

1900

1500

1300
1200
1100

0 100 200 300 400 500 600 700 800 900

Time t [ms]
Figure 3.12: The vertical synchronization of Martin M1, the VIS
code value is 44.

There is a vertical synchronization in fig. 3.12 with a value of 101011002 (44 dec-
imal). The parity bit is 1, and first three bits 010 distinguish the Martin system.
The vertical and horizontal resolution can be determined from the value of the sec-
ond nibble – 1 256 lines and 1 320 columns, the last two bits with value 00 mean
colour transmission.

3.6.3 Additional synchronization data

Some SSTV software append a signal with additional data to the synchronization,
e.g. call sign identification, which can then be decoded and used as an input for an

26
Modern SSTV features 3

electronic station log. Unfortunately these additional signals have no standardized

format and they are not compatible with other SSTV programs.
Some of them append the data transmission to the first scan-line image (ChromaPix)
or even prior to broadcast of VIS (WinPix, MMSSTV).
Some newly developed SSTV systems do not use standard VIS code with 8 bits
and send 16 bits (MP, MR, ML modes) or use odd parity for error checking. This
is because of 27 , from which 128 possible combinations of the VIS code is almost
exhausted. Differences in these and other systems will be described in further chap-
ters.

27
 4
Formats of slow-scan
TV transmission
4.1 Black and white SSTV systems
The earlier modes of SSTV transmission were displayed on long persistent monitors
with radar CRT. The duration of transmission for each image frame took 7.2 to 8
seconds, and when the last line was received the first line was still visible. It was
possible to see the whole picture in a darkened room.
Both 7.2s and 8s modes were used in the same period. The 7.2 s frame speed
mode, was used in Europe while the 8 s were used in America. The synchronization
of signals is derived from the electrical mains – 50 or 60 Hz. If an image was syn-
chronized at 60 Hz and received on 50 Hz equipment, it was still readable, but the
image was a little distorted. For long distance QSOs, it was possible to change the
oscillator to achieve European or American synchronization.
The disadvantages of 8s SSTV are low image resolution and a loss of synchroniza-
tion due to signal interference. The loss of synchronization could lead to the loss of
a few lines or the whole image.
The differences between modern SSTV modes and this old system are many, but
one parameter remains the same. Almost all new systems use 1200 Hz frequency
for sync pulses and the frequency band from 1500 Hz (black) to 2300 Hz (white) for
video signals. Also the old 8s mode is still supported by many SSTV programs for
transmission. It is important to note that 8s mode has the shortest transfer time
and should be used in special conditions.

4.1.1 Modes for digital converters

There are many modes for B&W image transmission which differ in transfer time
and resolution. Wrase and Robot modes are implemented in modern converters;
the transfer is extended to 256 lines and the transmission time is also prolonged to
achieve better horizontal resolution.
Historically common modes are: the 16 second mode with 128 lines,the 32 second
mode with 256 lines and the 64 second mode with 256 lines;which provides maximal

28
Formats of slow-scan TV transmission 4

image quality. All these modes are related to the original 8s mode and also have an
image aspect ratio 1:1. The number of lines, columns or both were simply multiplied
twice. This design was used in Wraase’s B&W converters.
While Wraase’s modes were derived from the European 7.2s mode, Robot Re-
search developed an original system for their converters. Robot’s modes aren’t sim-
ply derived by “doubling” parameters but are derived from line speed. While the
American 60 Hz/8s standard has a transfer speed of 900.0 lpm, Robot’s line speeds
for new modes were set at 600.0 lpm so that 120 lines were transferred in 12s and
240 lines in 24s. The mode with the best resolution has a line speed of 400,0 lpm
and a total transmission time of 36 seconds.
The Robot SSTV system reserves the first 16 or 8 lines (for a 240 or 120 line
image) for gradation gray scale. The scale can be used to tune the signal more
precisely.
Although Robot Research cooperated with Copthorn MacDonald, they ignored
the trend in amateur construction of digital converters with doubled modes. Never-
theless, the Robot converter Robot 300 became quite popular despite the high price
tag of over $800 in the mid-seventies.
During the 70’s and 80’s, the ham radio market was not the only outlet for SSTV
converters, but companies found opportunities in the telecommunication market and
sold SSTV monitors and cameras as devices for image transmission over telephone
lines.

4.1.2 BW transmission with computer software

An example of B&W mode implemented with computers is the AVT 125 BW mode
of the Amiga Video Transceiver system and it is suitable for good quality image
transfer in circa 2 minutes. The mode has a vertical resolution of 200 lines because
the Amiga computer resolution was 320×200. The AVT system is different from the
previous B&W modes because it has no line sync like WEFAX. The transmission is
based on a fully synchronous communication and the exact timing of corresponding
stations. This special feature is described in more detail in chapter 4.2.5 about
color AVT modes.
There is also the FAX480 mode for high resolution transmission, with 512×480
image resolution described further in chapter 4.3.1.
Early B&W modes Wraase and Robot, need to be synchronized with both line
and vertical synchronization. The line speed describes the free-run speed, but in
reality it can be deviated up to ± 5 %.
Modern modes like FAX480 and AVT 125 BW need accurate precision of line
speed, because just a little deviation of values in tenths causes image slant and
distortion.

29
4 Formats of slow-scan TV transmission

Mode Resolution Aspect Sync. Scan line Line speed

ratio (ms) (ms) (lpm)
7.2s (50 Hz) 120×120 1:1 5.0 55.0 1000.0
8s (60 Hz) 120×120 1:1 5.0 60.0 900.0
Wraase SC-1 8 128×128 1:1 5.0 55.0 1000.0
Wraase SC-1 16 256×128 1:1 5.0 115.0 500.0
Wraase SC-1 16 Q 128×256 1:1 5.0 55.0 1000.0
Wraase SC-1 32 256×256 1:1 5.0 115.0 500.0
64 s mode 256×256 1:1 5.0 115.0 250.0
Robot B&W 8 160×120 4:3 10.0 56.0 900.0
Robot B&W 12 160×120 4:3 7.0 93.0 600.0
Robot B&W 24 320×240 4:3 12.0 93.0 600.0
Robot B&W 36 320×240 4:3 12.0 138.0 400.0
AVT 125 320×400 4:3 — 312.5 192.000
FAX 480 512×480 1:1 5.12 262.144 224.497
SP-17 BW 128×256 4:3 5.0 62.0 895.520

Table 4.1: Parameters of black and white SSTV modes.

The advantage of longer transmission is improved image quality. The disadvan-

tage, is that a lot of time is needed for the transfer, which could be better utilized
for the transmission of color images.

4.2 Color SSTV modes

You might find it incredible that the first color transfer was made before the era
of digital converters using long persistence monitors. Each color image channel was
obtained using color filters, which were subsequently held in front of the camera. A
sample result could be that the first channel transferred was blue, then green and
the last red. Slightly more difficult was the processing on the receiver side. This
was because each color channel had to be photographed from the monitor screen
and then the resultant color picture was combined from all three components. It
was a very laborious process, but it was put into practice a few times!
Further experiments with color SSTV transmission were based on frame sequential
transfer. Three complete images were transferred in 8s mode and each contained one
color channel, together they formed one color image. During broadcast, a color

30
Formats of slow-scan TV transmission 4

Robot B&W 8 Robot B&W 12

Robot B&W 24 Robot B&W 36

Figure 4.1: Comparison of Robot system’s BW modes.

original was progressively scanned with a BW camera through each of the color
filters. Received images had to be stored in a digital converter in three different
memories. When simultaneously displayed on a color monitor they created a full
color image. This is the reason why BW modes of Robot and Wraase families have
three different VIS codes for BW transfer. The codes are sent for adjustment of color
components for frame-sequential transmission. Individual images were usually sent
in the order of red – green – blue. But the order of the channels could be changed
under the agreement of corresponding stations, or some images could be broadcast
repeatedly. With such a method, it is possible to transfer only static scenes. If
an object moves during manual scanning of an image, the color components do not
correspond and the result image has colored ghosts.
The transfer was not always reliable; due to interference and fade-outs, the im-
age component had to be sent several times. And in practice, it was sometimes
problematic to complete all color channels. To improve color transmission the line
sequential transfer was developed. The principle is that it transmits a single image

31
4 Formats of slow-scan TV transmission

and each scan-line carries all three color components. A receiving equipment can al-
ready display color images during transmission. This method where the color image
is transferred in one frame is referred to as SFC – Single Frame color.
More properties of SSTV systems will be introduced in following sections, with
all their pros and cons and details of mode formats described in detail.

4.2.1 Wraase SC-1

This line sequential system was first among newly developed SFC systems. Wraase
SC-1 comes from the workshop of famous SSTV engineer Volker Wraase, DL2RZ.
The system was most likely created by modifying existing equipment to operate in
8s mode or for frame-sequential transmission.
Each scan-line begins with 6.0ms sync, then a green component follows and then
the blue and red components. A separate sync of 6.0 ms length precedes each colour
component.
Wraase SC-1 has a major deficiency. If the receiver loses sync during interference,
then the display system loses the ability to synchronize colors. Because all lines are
sent in the same way, the color components cannot be recognized and the probability
that the system reverts back to correct color sync is equal to one-third. In practice
the system works, but when the noise level is too high, the received image contains
few color bands as the converter loses and restores synchronization. For this reason,
an additional sync pulse was added to subsequent productions of the SC-1 converter.
It consists of a truncated 5ms sync before the red line,which is immediately followed
by a short pulse of 2300 Hz frequency lasting 1-2 ms. It allows the converter to
regain synchronization after the noise subsides. Additional synchronization occurs
as a thin red stripe in the left edge of the image.
All SC-1 modes have an image aspect ratio of 1 : 1. The original SC-1 mode is the
24s mode with 128 lines, so the image quality is not better than the 8s mode, but
the colors improve the picture.
The system was soon upgraded for modes with longer transmission. First, the
number of lines was doubled to 256 and the transfer extended to 48 seconds. The
last SC-1 96s mode has better horizontal resolution for good image quality.
The professional converter Wraase Electronics SC-1 was most popular in Germany,
but its market share was lower in comparison to the Robot converters produced in
the same period.

4.2.2 Robot color system

The Robot modes are named according to the converters in which they were im-
plemented. They are scan-converters Robot 400C, 450C and 1200C. They were
produced in San Diego by Robot Research Inc.

32
Formats of slow-scan TV transmission 4

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync G B R (lpm)
Wraase SC-1 24 Color 24 s 128×128 G–B–R 6.0 54.0 54.0 54.0 333.3

Wraase SC-1 48Q Color 48 s 256×128 G–B–R 6.0 108.0 108.0 108.0 175.4

Wraase SC-1 48 Color 48 s 128×256 G–B–R 6.0 54.0 54.0 54.0 333.3

Wraase SC-1 96 Color 96 s 256×256 G–B–R 6.0 108.0 108.0 108.0 175.4

Table 4.2: The Wraase SC-1 scan-line timing.

They do not use RGB color coding as SC-1, but YCrCb. Scan-lines consist of a
luminance signal 𝑌 followed by differential chrominance signals 𝑅 − 𝑌 and 𝐵 − 𝑌.
Due to this, the color modes are compatible with their B&W variants. So a 12s
color mode can be displayed by 8s monitors, etc.
From a total of 8 modes 4 are intended for color transmission. Half of the color
modes use YCrCb in a 4:2:0 format. The scan-line contains only one chrominance
signal, and colors are obtained from the average of two adjacent lines in the original
image. The other two modes use the 4:2:2 format and send all color information in
one scan-line.
The original Robot system uses asynchronous transfer. To receive the image, it
is needed to detect the vertical sync (VIS code). And for proper reception of the
image, the sync pulse must be detected. This process is a major disadvantage.

2300
Frequency f [Hz]

1900

1500

1200

0 50 100 150 200 250 300

Time t [ms]
Figure 4.2: Two scan-lines of Robot
36 Color when color bars are sent.

The scan-line is composed of the starting sync, followed by a short 3.0ms gap of
1500 Hz and then the image part with luminance and chrominance. The chrominance
differential signals begin with additional sync pulses. The 1500Hz sync is before

33
4 Formats of slow-scan TV transmission

𝑅 − 𝑌 and the second 2300Hz is before 𝐵 − 𝑌. Due to the additional sync with
a different sync frequency, it is possible to re-synchronize 4 : 2 : 0 formats after an
interruption. The chrominance syncs are separated from the scan-line with 1500Hz
gap that lasts 1.5 ms.

2300
Frequency f [Hz]

1900

1500

1200

0 50 100 150 200 250 300

Time t [ms]
Figure 4.3: The scan-line of Robot
72 Color when the color bars are sent.

The main disadvantage of the Robot modes lies in color coding, because the receiver
must be perfectly tuned to the SSTV signal. Otherwise the image hue is distorted
when the deviation is greater than ±50 Hz. For this reason, Robot Research intro-
duced the transmission of gray gradation scale at the beginning of image transfer
and the receiver device can auto-tune for video signal.
The whole frame has 256 or 128 lines, of which the first 16 or 8 lines are reserved
just for gradation scale. Some converters and PC software add some basic station
info, such as call sign, and this part of the frame is called “header”.
The memory storage capacity of the Robot 1200C converter allows it to store an
image with a resolution 256×240 pixels or four images with 128×120, and they are
displayed in a 4 : 3 aspect ratio.
The fastest mode of the Robot family and the fastest color SSTV mode is the 12s
mode. It contains 120 lines transmitted in the 4:2:0 format. Another mode is 24s
with a 256×120 resolution and 4:2:2 color format. The other two modes allow the
transfer of images in 256×240 resolution, either in less quality for 36 seconds in a
4:2:0 format or in better quality in 4 : 2 : 2 format for 72 seconds.
Although the Robot modes were pushed away by modern synchronous modes that
are more resistant to interference, the 24s and 36s modes are faster than modes
with RGB color coding, and have better resolution than RGB modes with the same
transmission time. You can find their benefits on VHF with FM transmission,
because it eliminates the need for precise tuning.

34
 Formats of slow-scan TV transmission 4

Robot 12 Color Robot 24 Color

Robot 36 Color Robot 72 Color

Figure 4.4: Comparison of Robot color modes.

Robot 36 Color was used in MAREX2, SAREX3 and ARISS4 programes for SSTV
transmission from orbital stations Mir, ISS and space shuttle missions. It is a pretty
good compromise between image quality and transfer time, because space stations
on low earth orbit can be received within just 10 minutes during their orbit.

4.2.3 The Martin synchronous system

The creator of this popular system is Martin Emmerson, G3OQD. He originally
named it “New Modes”, but to avoid confusion between other newly emerging SSTV
modes, the community universally named modes after their creators. The Martin
was created to overcome SFC problems in systems like SC-1 due to two main changes.
The first change was that instead of three separate syncs before each color com-
ponent, there is just a single sync sent before each scan-line. The horizontal sync
lasts 4.862 ms. After the horizontal sync, the green component is sent, then blue

2 Mir Amateur Radio Experiment

3 Shuttle Amateur Radio Experiment
4 Amateur Radio on the International Space Station

35
4 Formats of slow-scan TV transmission

Mode Transfer Resolution Color Compatible

name time format B&W mode
Robot 12 Color 12 s 160×120 4 : 2 :0 Robot B&W 8
Robot 24 Color 24 s 320×120 4 : 2 :2 Robot B&W 12
Robot 36 Color 36 s 320×240 4 : 2 :0 Robot B&W 24
Robot 72 Color 72 s 320×240 4 : 2 :2 Robot B&W 36

Mode Color Sync pulses of Scan-line Speed

name sequence line color color Y R−Y B−Y (lpm)
Robot 12 Color YCrCb 7.0 3.0 — 60 30 600.0
Robot 24 Color YCrCb 12.0 6.0 6.0 88 44 44 300.0
Robot 36 Color YCrCb 10.5 4.5 — 90 45 400.0
Robot 72 Color YCrCb 12.0 6.0 6.0 138 69 69 200.0

Table 4.3: The Robot parameters and scan-line timing.

and last is the red component. Between each color components, there are short gaps
of 1500 Hz lasting 0.572 ms. Just like in the SC-1, the sequence green – blue – red
was chosen. Regardless of the order in which components are sent, the image qual-
ity will not change. But it is important that the receiving device identifies which
component it is currently receiving.

2300
Frequency f [Hz]

1900

1500

1200

0 50 100 150 200 250 300 350 400 450

Time t [ms]
Figure 4.5: Scan-line of Martin M1 when color bars are sent.

An important feature of using only one sync before beginning the color scan-line
sequence, is that a converter will not replace the individual color components and

36
Formats of slow-scan TV transmission 4

degrade the color information. In time intervals where the line sync is not transmit-
ted, the gaps are filled with a reference level of black at 1500 Hz for 0.572 ms.
The second improvement has a substantial effect on image reception. Unlike the
Robot or SC-1, the detection of horizontal syncs is not necessary during reception.
And the broadcast between stations is fully synchronized. The results of the use
of such a system are sharper images and more contrasted edges. Although the
transmission conditions on the lower HF bands often do not allow the transfer of the
image in 100% quality, old systems relying on line sync usually lose synchronization
in such conditions.
The Martin system was originally implemented as a modification of the Robot
1200C converter and it preserves the transmission of the header gradation scale.

Robot 36s Color Martin M1

Figure 4.6: A comparison of systems
in real conditions on the 3.7MHz band.

Line syncs and inner scan-line gaps have a similar duration at all four speeds, but
the number of lines and horizontal resolution for each mode is different. Although the
syncs aren’t necessary for transmission, they are still transmitted at the beginning of
each scan-line in order to synchronize the converter at any time during the reception.
It is important because it consumes a lot of broadcast time and the station does not
have to wait for the start of a new frame, but a receiver can get synchronization at
any time during transmission.
The Martin system allows us to work with four different modes/speeds. The most
popular version is the Martin M1 with 256 lines per frame in two minutes. Other
modes of the Martin system have either half the line or half the horizontal resolution
of the best quality M1. The mode M4 has the lowest quality and 128 lines. Modes
Martin M1 and M2 are often used between European stations.

37
4 Formats of slow-scan TV transmission

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync G B R (lpm)
Martin M1 114 s 320×256 G–B–R 4.862 146.432 146.432 146.432 134.3947532

Martin M2 58 s 160×256 G–B–R 4.862 73.216 73.216 73.216 264.5525975

Martin M3 57 s 320×128 G–B–R 4.862 146.432 146.432 146.432 134.3947532

Martin M4 29 s 160×128 G–B–R 4.862 73.216 73.216 73.216 264.5525975

Table 4.4: The Martin scan-line timing.

4.2.4 Scottie
These modes were created by Eddie T. J. Murphy, GM3SBC. He modified the orig-
inal firmware of Wraase SC-1. Martin Emmerson also implemented Scottie modes
to Robot 1200C later on.
Scottie has the same improvements as the Martin system does, but its scan-line
composition and scan timing are different.
After vertical synchronization, the sequence of scan-lines is; a 1.5ms short gap of
1500 Hz, then a green component, a 1.5ms short gap again, a blue component, then a
horizontal sync, another gap and lastly, a red component. This unusual order is the
result of the system adaptation to SC-1, where the additional sync was used right
before the red component. Syncs are permanently sent for any time synchronization
during reception.
The Scottie relies on exact timing like the Martin, although the original version for
SC-1 was not fully synchronous and syncs were still processed by the converter. But
in newer systems the modes are implemented for free-run reception, so the system
is equivalent to the Martin.
The implementation of Scottie in Robot 1200C slightly differs, because the first
scan-line includes an additional 9.0ms sync at the beginning of the scan-line right
after vertical synchronization. All other modes implemented in Robot 1200C have
sync at the beginning of the scan-line but the Scottie has the sync in the middle of the
scan-line which then caused color distortion. Perhaps some other implementation
of Scottie has this difference too.
The Scottie system also has four conventional modes (and a special one described
later). Two with 256 lines per frame and two with 128 lines. The difference in timing
is not the same as in the Martin, where the line speed of the faster mode is exactly
twice the speed of the slower mode, so the speed of the faster mode is lower than
twice that of the slower mode.
Image quality in the Scottie and Martin modes is the same. Theoretically, a
slightly better quality can be achieved in Martin M1 than in Scottie S1 due to
longer transmission, but the difference is imperceptible.

38
Formats of slow-scan TV transmission 4

The Scottie S1 and S2 are quite popular for North American stations and can
often be heard on high frequency bands.

4.2.4.1 Scottie DX – special mode for long distance transfers

This mode of the Scottie family achieves the best possible results in the transmission
of slow-scan television images. There is one simple reason for this; the transmission
takes about 2.5 times longer than Scottie S1.
There is an extended duration of the scan-line, but the duration of sync and gaps
between color components remained the same. This improvement is best seen on
the receiving side. The longer transmission time supports better image quality.
The improvement relies on the fact that; each pixel can be read more times during
signal sampling and that the loss of a few samples does not affect overall quality.
It means that each pixel takes a long time and this gives better noise and phase
distortion immunity. But these qualities are compensated by a very long image
transmission time of about 4.5 minutes. During this time, two images with the
same resolution can be sent with other RGB modes.
The Scottie DX mode offers high quality images, but sometimes the optimal con-
ditions for DX connections do not last long enough for the transfer of a whole picture.

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync G B R (lpm)
Scottie S1 110 s 320×256 G–B–R 9.0 138.240 138.240 138.240 140.1148942

Scottie S2 71 s 160×256 G–B–R 9.0 88.064 88.064 88.064 216.0667214

Scottie S3 55 s 320×128 G–B–R 9.0 138.240 138.240 138.240 140.1148942

Scottie S4 36 s 160×128 G–B–R 9.0 88.064 88.064 88.064 216.0667214

Scottie DX 269 s 320×256 G–B–R 9.0 345.600 345.600 345.600 57.12653528

Table 4.5: The Scottie scan-line timing.

4.2.5 Amiga Video Transceiver

AVT modes were originally intended for SSTV operations with Amiga computers.
AVT author Ben B. Williams, AA7AS developed a dedicated interface and software
which was produced by AEA (Advanced Electronic Applications Inc.). Although
the creator claimed that this system was a revolution in SSTV transmission, these
modes did not gain popularity like other modes. The AVT modes are practically
not in use today.
A reason for this could be the fact that the manufacturer wanted to keep the
image scan parameters of the system, secret. However, by intercepting signals and

39
4 Formats of slow-scan TV transmission

reverse engineering, the parameters of the AVT modes were implemented in other
devices by the SSTV community. This was done without the additional software
tools that made the AVT unique.
The AVT system contains four line sequential RGB modes and one B&W. The
scan-lines have no gaps between color components and a really unusual thing is that;
the modes do not use any horizontal sync. Another unusual feature is the mandatory
function of vertical synchronization, that is sent as a digital header before the image
transfer begins.
The AVT family contains 5 modes and each of them has the following four options:

1. Default variant is the same as conventional SSTV modes, but does not have any
line syncs.
2. Narrowband variant uses shorter band for video signals from 1700 Hz for black
to 2100 Hz for white.
3. QRM variant, that uses picture interlacing just like in analog television.
4. The combination of the QRM and narrowband variant.

The fastest mode is the AVT 24 with 120 lines and it is transferred for 31 seconds.
The next mode is AVT 90 with a resolution of 256×240 and an image quality slightly
worse than in the Martin M1. ATV 90 sends each color component in 125.0 ms, thus
the speed is 2048 pixels per second (in binary notation this gives a nice rounded
number). The other two modes have somewhat atypical resolutions in comparison
with other SSTV modes, but these resolutions are normal system resolutions on
Amiga computers. It is AVT 94 with 320×200 and AVT 188 with the same line
speed, but twice the scan-lines – 320×400. The image is displayed in an aspect ratio
of 4 : 3 in both cases.
For some SSTV systems/scan-converters, the detection of vertical sync is a must.
So, the VIS code is repeated three times for accurate reception. VIS is necessary for
image reception when no line sync is sent and later synchronization is not possible.
The original AVT software however, does not need to receive VIS, but relies more
on the digital header.
After a series of VIS code, there is a digital header (see fig. 4.7), which contains
synchronization data. It is a sequence of 32 frames of 16 bits. Each frame contains
only 8 bits of information, but it is sent twice – first in normal form and second
inverted. Normal and inverted parts can be compared for error detection. Each
frame starts with a 1900Hz pulse while data modulation uses 1600 Hz for the rep-
resentation of logical zeros and 2200 Hz for logical ones. Narrow-band variants use
1700 Hz for zeros and 2100 Hz for ones. Both variants use a modulation speed of
exactly 2048/20 = 102.4 Bd, so the data pulse has a length of 9.766 ms.

40
Formats of slow-scan TV transmission 4

2300
Frequency f [Hz]

1900

1500

1200

3400 3600 3800 4000 4200 4400

Time t [ms]
Figure 4.7: The digital synchronization header of AVT 90 mode (VIS
68, normal variant).

The first three bit of each 8bit word identifies the mode:

⊳ 010 – AVT 24,

⊳ 011 – AVT 94, AVT 188, AVT 125 BW,
⊳ 101 – AVT 90.

The last five bits are used as a count down before image transmission. Actually
these five bits are important for an accurate set of image initiation and synchroniza-
tion. They vary between all 32 binary combinations during transmission. At least
one binary code must be properly detected. At the beginning, all bits are in 0 states
with 1 in inverted parts. When the countdown starts, all five-bit sequences run (e.g.
for AVT 24):

010 00000 101 11111

010 00001 101 11110
010 00010 101 11101
…
010 11101 101 00010
010 11110 101 00001
010 11111 101 00000

When the count down gets to zero, the image scan-lines are sent. AVT recep-
tion depends on the first eight seconds of synchronization,for some implementations
without the ability to synchronize later. Although the AVT modes are quite reli-
able, noise could cause a loss of the whole image. Sometimes it is not possible to
receive a digital header due to interference, even if the interference later disappears.
However, the original AVT software was capable of image reconstruction in this case.

41
4 Formats of slow-scan TV transmission

Because the image data is completely synchronous, the data simply has to be shifted
in memory until the RGB data is aligned correctly, and then the image comes out
perfectly. Again, the AVT system provided means to hot reconfigure the data after
reception. So reception without/after sync header worked fine.
The earlier listed options for each mode can improve its performance. The first
is the narrow-band transmission which uses a 400 Hz band from 1700 Hz (black) to
2100 Hz (white). With an appropriate filter, the resistance to interference can be
improved with minimal loss of image quality. For instance; the 400 Hz wide CW
filter can be used with a variable IF shift.
The second option is the “QRM mode”, where an entire image is sent interlaced.
Within the first half of image transmission time, half of the scan lines (every odd
one) is sent. Then the scan loops back to the beginning and sends the remaining
half lines (even lines). The fact that some of the disturbed lines of the first field is
interlaced with fine lines received from the second will definitely improve the overall
subjective impression of image quality. The original AVT software also contains
tools for handy image improvement – it is possible to select distorted lines and
the program will reconstruct them by averaging neighborhood lines. It is also is
possible to shift the second field horizontally independently of the first field. This
allows you to compensate if there is a significant multi-path delay in regard to the
two fields.
In ATV implementations, the system can work well without this interactive tools.
But in practice, especially on shortwaves where conditions change quickly; the second
field could be phase-shifted and this causes the notable “toothy” edge of the picture.
The QRM option can be combined with the narrow-band mode.

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync R G B (lpm)
AVT 24 31 s 128×120 R–G–B — 62.5 62.5 62.5 960.000

AVT 90 98 s 256×240 R–G–B — 125.0 125.0 125.0 480.000

AVT 94 102 s 320×200 R–G–B — 156.25 156.25 156.25 384.000

AVT 188 196 s 320×400 R–G–B — 156.25 156.25 156.25 384.000

AVT 125 BW 133 s 320×400 Y — 312.5 192.000

Table 4.6: The AVT scan-line timing.

4.2.6 Wraase SC-2

A later version of Wraase modes was first built in the newer converter SC-2 from
Wraase Electronics. Again, it provides another variant of line sequential systems.
The author dropped the sequence of colors used in the earlier SC-1 converter, so the

42
Formats of slow-scan TV transmission 4

colors are now sent in the order: red – green – blue. Additionally, there is only one
horizontal sync at the beginning of each line, just as in the Scottie and Martin.
Unlike other systems, the RGB system in the SC-2 has one characteristic that
distinguishes it from other conventional modes. Image transfer is achieved when the
transmission time for the green component is equal to the sum of the transmission
time of the red and blue components, i.e. the ratio 2 : 4 : 2 of R : G : B components.
Between color components short gaps are not sent.
As we already know that the human eye is most sensitive to green by more than
50 %. The remaining 50 % in SC-2 is split evenly between the red and blue com-
ponents. Red and blue components are not processed for a differential signal. This
color reduction is not visible on common pictures, but it may happen that some im-
ages (e.g. B&W mosaic) may lose color information. The system is less precise for
color interpretation in comparison with YCrCb modes, but better in tuning resis-
tance. One disadvantage of color reduction is found, when green shadows appears
on the image in stations without precise clock timing.
This mode is preferable in comparison to YCrCb because bad tuning will only
reduce the contrast or saturation, but the hue is not distorted. Occasional green
shadows remain as a tax for reduced transmission time.
The Wraase SC-2 family just like all other systems, also has four different modes.
The SC-2 180 offers best quality for three minute transmission, and unlike the pre-
vious modes does not use the RGB ratio 2 : 4 : 2 and is therefore a faster alternative
to the Scottie DX mode. The two-minute SC-2 120 uses the RGB format 2 : 4 : 2.
The remaining two modes, SC-2 30 with 128 and SC-2 60 with 256 lines have about
half of the resolution found in SC-2 120.

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync R G B (lpm)
Wraase SC-2 30 30 256×128 R–G–B 5.0 58.0 117.0 58.0 249.600

Wraase SC-2 60 60 256×256 R–G–B 5.0 58.0 117.0 58.0 249.600

Wraase SC-2 120 120 320×256 R–G–B 5.0 117.0 235.0 117.0 126.175

Wraase SC-2 180 180 512×256 R–G–B 5.0 235.0 235.0 235.0 84.383

Table 4.7: Wraase SC-2 scan-line timing.

4.3 High resolution transmission

High quality images consume a lot of memory but memory was very expensive in
early computer systems. High resolution images were a real luxury, but over the
years memory has gotten cheaper, therefore modern SSTV systems now have modes
for high resolution broadcast too.

43
 4 Formats of slow-scan TV transmission

4.3.1 FAX480
Synchronous mode was the first high resolution mode. It was first implemented in
the ViewPort VGA interface and software for IBM PCs in 1993. The old VGA cards
with 256 kB of memory can hold an image with a resolution 640×480 with only 16
colors. This provides only gray scale images, so this mode is used for only B&W
transmission.
The image resolution of FAX480 is 512×480 and the transmission time is 138 sec-
onds. In the early days of high resolution transmission, the only way to transmit hi-
res images was facsimile (see chapter 11). So the creator Ralph Taggart, WB8DQT
called it FAX480, but compared with classic facsimile there are not many similarities.
The synchronization of the FAX480 is derived from the reference frequency of 4.0
MHz, and a time unit is 4 MHz/2048 = 1953.125 Hz.
Vertical sync is resolved as follows. In the first five seconds a rectangular frequency
modulation of 244 Hz between the black (1500 Hz) and white (2300 Hz) levels is
transmitted. This creates the APT5 signal.
The tone 1500 Hz is transmitted for 4 time units (4 × [1/1953.125] = 2.048 ms)
and 2300 Hz for 2.048 ms too. This gives a frequency of an ATP tone also 244 Hz
(1/[2.048 + 2.048] = 244 Hz). This sequence is then repeated exactly 1,220×.
Originally the system did not use the VIS code, but the code 85 was later added.
Originally, vertical sync is followed by a phasing interval of 20 white lines. Each
begins with 5.12 ms sync 1200 Hz (10 time units), but this interval is omitted in
some implementations.
Now it’s time to transfer the image itself. It is composed of 480 lines. Each
line begins, unlike the facsimile, with 1200Hz sync with a length of 5.12 ms and
then continues a scan-line with 512 pixels. The duration of the scan-line is 512 ×
(1/1953.125) = 262.144 ms.
According to the creator, the horizontal resolution of 512 points was selected just
because the FAX480 operating software had a control menu to the left of the screen.

4.3.2 Pasokon TV
These synchronous modes were released with Pasokon TV interface from John
Langer, WB5OSZ. These modes retain essential SSTV parameters. They also used
color coding to transmit the individual color components in the order of red – green –
blue with the format 1 : 1 : 1.
There are three modes in the Pasokon system. They have different transmission
times: 3, 5 or 7 minutes, so the image quality differs.
Each mode has a default timing for scan-lines:

5 Automatic Picture Transmission signal, see section 11.2.1 .

44
Formats of slow-scan TV transmission
4

⊳ Pasokon P3 …4800 Hz
⊳ Pasokon P5 …3200 Hz
⊳ Pasokon P7 …2400 Hz

The scan-line starts with a sync of 20 time units, then there is a 5 unit black gap
followed by the red component. It has 640 units, so there is one unit for each pixel.
There are 5 unit black gaps between color components and at the end of the scan-
line before the sync of the next line. These gaps should help improve the detection
of syncs.
Pasokon P7 has the best image quality and longest tramsmission time which takes
nearly seven minutes. If we split such an image into four equal parts, the image
quality of one of them would correspond to that produced by the Martin M1 or
Scottie S1 modes. The upper 16 lines are used for gray scale, and the remaining 480
for your own image.
There are also two other modes with 480+16 lines. The P5 has a transfer time of
almost 5 minutes with lower image quality and the P3 runs fastest at three minutes
with a horizontal resolution about half of a P7.
A potential disadvantage of these modes is actually quite a long transfer time,
which makes it difficult to use on highly variable short-waves. For those who do not
mind the long transmission times, it can be used for exchanging pictures on VHF.

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync R G B (lpm)
Pasokon P3 203 s 320×496 R–G–B 5.208 133.333 133.333 133.333 146.56488550

Pasokon P5 305 s 640×496 R–G–B 7.813 200.000 200.000 200.000 97.70992366

Pasokon P7 406 s 640×496 R–G–B 10.417 266.667 266.667 266.667 73.28244275

Table 4.8: The Pasokon TV scan-line timing.

4.3.3 PD modes
PD modes are the result of a cooperation between Paul Turner G4IJE and Don
Rotier K0HEO. The mode was first introduced in May 1996 and it was developed
to improve image quality and especially to reduce transfer times in comparison with
Pasokon TV.
For speeding up transmission, YCrCb color coding is used in the 4 : 2 : 0 format.
If you divide the total time between two syncs by four, the result is the actual
time for each color component. The scan-line begins with 20.0ms sync, then there
is a 2,080 ms gap of black, and the first luminance signal 𝑌1 . It is followed by
chrominance signals 𝑅 − 𝑌 and 𝐵 − 𝑌 without any gap. Then there is a second
luminance 𝑌2 . The exact timing of modes is:

45
4 Formats of slow-scan TV transmission

⊳ PD-50 – 286 𝜇s/pixel ⊳ PD-180 – 286 𝜇s/pixel

⊳ PD-90 – 532 𝜇s/pixel ⊳ PD-240 – 382 𝜇s/pixel
⊳ PD-120 – 190 𝜇s/pixel ⊳ PD-290 – 286 𝜇s/pixel
⊳ PD-160 – 382 𝜇s/pixel

YCrCb color coding needs accurate signal tuning to prevent color distortion.
Thanks to a wide horizontal sync, it is possible to detect frequency deviation and
compensate color distortion. There is also gray scale on the top of image for tuning
detection.
The main advantage is reduced transmission time compared with RGB modes.
The PD-290 mode supports a resolution of 800×600 and its transfer time is nearly
five minutes, although at the cost of little color loss. Some modes have resolutions
of 640×480, while PD-160 has 512×384. The fastest two-minute PD-120 has a worse
image quality, but in many cases it is still sufficient. Beside the five modes with high
resolution, the system includes two with standard resolution. PD-90 uses 320×240
and has a better image quality than Martin M1 or Scottie S1, because it is based on
a longer transmission time per pixel. The last mode is the very fast PD-50, which
provides a similar resolution as Scottie S2.

Mode Transfer Color Scan line (ms) Speed

Resolution
name time sequence Sync 𝑌1,2 𝑅 − 𝑌 𝐵 − 𝑌 (lpm)

PD-50 50 s 320×240 Y-C 20.0 91.520 91.520 91.520 309.150866

PD-90 90 s 320×240 Y-C 20.0 170.240 170.240 170.240 170.687301

PD-120 126 s 640×480 Y-C 20.0 121.600 121.600 121.600 235.997483

PD-160 161 s 512×384 Y-C 20.0 195.854 195.854 195.854 149.176545

PD-180 187 s 640×480 Y-C 20.0 183.040 183.040 183.040 159.100552

PD-240 248 s 640×480 Y-C 20.0 244.480 244.480 244.480 120.000000

PD-290 289 s 800×600 Y-C 20.0 228.800 228.800 228.800 128.030044

Table 4.9: The PD modes scan-line timing.

4.4 Experimental modes

During the years of the SSTV boom many modes were created, but never gained
popularity. Many of them are totally forgotten, like WinPix GVA, Proscan J-120,
WA7WOD system or ScanMate, although some of them have a few interesting fea-
tures which we are about to delve into.

46
Formats of slow-scan TV transmission 4

4.4.1 MSCAN TV
The modes TV-1 and TV-2 were one of many experiments in the SSTV transmission
field. An interesting feature is the use of interlaced transmission. They do not use
the same half-frame interlacing like normal television does. But the whole image,
is divided into four quarter-frames. These frames are transmitted gradually in the
direction from top to bottom, so you can get a first preview of the image after the
first quarter of transmission time, but only in low resolution. Thanks to interlacing
the resolution increases gradually during transmission up to 320×256.
It is possible to receive these modes with conventional equipment without inter-
lacing support, because their line speed are the same as for Wraase SC-180 (TV-1)
and Martin M1 (TV-2) modes. But in this condition, the image will contain four
bars with all quarter-frames.

1. 2. 3. 4.
Figure 4.8: MSCAN TV image interlacing.

4.4.2 Kenwood FAST FM

This mode is built in the mobile SSTV converter Visual Comunicator VC-H1 from
Kenwood (see section 6.7). This unit support some normal modes and the “FAST
FM” mode.
The FAST FM mode sends video signals in the 2800 Hz (black) to 4400 Hz (white)
band. The vertical synchronization and VIS code format is similar to Robot’s stan-
dard, it has a value of 90, but uses odd parity (the number of logical ones must be
odd). After the VIS code there is a digital header and then an image with a resolu-
tion of 320×240.
The duration of one scan-line is 53.6 ms, so the total transmission time for an
image is 13.5 seconds. The mode uses YCrCb color coding in the 4 : 2 : 0 format.
The brightness signal occupies 35.4 ms of scan-line, and than there is a pulse of
3600 Hz that lasts 0.41 ms and then color signals are sent. Each even scan-line
contains 𝑅 − 𝑌 and odd line 𝑅 − 𝑌. The scan-line is ended by 0.41 ms pulse again.
The transmission of a whole image is ended by one second pulse of 1900 Hz.
Due to fast transmission, the used bandwidth of FAST FM is in 1.0 to 6.2 kHz
range, so it cannot be used in the SSB voice channel, but only in FM channels on
VHF. The image quality is comparable to the Robot 36 Color mode.

47
4 Formats of slow-scan TV transmission

4.4.3 Modes MP, MR, ML

These modes were created by Makoto Mori, JE3HHT, the author of MMSSTV
software. Some of these modes became quite popular, because of the success of
MMSSTV. The author created modes with both standard and high resolutions.
They use YCrCb colors and extended VIS code. Some modes use a narrower band
for syncs and video signals.

Martin M1 MP115
Figure 4.9: The comparison of modes
in real conditions on the 3.7MHz band.

The change he made to the traditional VIS specification extends the code by 8 extra
bits, so a 16-bit code is sent instead. The first 8 bits (LSB) are the same for each
mode with a value of 35 (0x23) that identifies the system. While the remaining bits
(MSB) distinguish a particular mode. Odd parity is used as a simple check.

2300
Frequency f [Hz]

1900

1500

1300
1200
1100

0 200 400 600 800 1000 1200

Time t [ms]
Figure 4.10: The 16-bit VIS code
of MP115 mode with a 0x2923 value.

48
Formats of slow-scan TV transmission 4

VIS used in narrowband modes has a very little in common with the original stan-
dard. Initially, during vertical synchronization N-VIS pulses of 1900Hz and 2300Hz
in 100 ms are sent, followed by a start bit of 1900 Hz (see fig. ??).
All code bits have a duration of 22 ms (modulation speed is 45.45 Bd). Logic one
has 1900 Hz and logic zero 2100 Hz. The code word length is 24 bits and it is divided
into four groups of 6 bits, bits are sent in the following order:
Each group has the following meaning:
⊳ Group 0 (5–0) = 101101
⊳ Group 1 (15–10) = 010101
⊳ Group 2 (25–20) = N-VIS
⊳ Group 3 (35–30) = 010101 xor N-VIS
For example, MP73-N has N-VIS = 000010 (0x02) and the whole code word is:
101101 010101 000010 010111.

2300

2100
2044
Frequency f [Hz]

1900

1500

1200

0 500 1000 1500 2000

Time t [ms]

Figure 4.11: Vertical synchronization and scan-line of the MP110-N

narrowband mode.
MP modes use the same principle as PD modes. The sync takes 9.0 ms followed by
a short 1500 Hz gap of 1.0 ms, then odd scan-line 𝑌 brightness is sent followed by
the 𝑅 − 𝑌 and 𝐵 − 𝑌 chrominance signals. The chrominance signals are the average
of two neighborhood scan-lines. The scan-line is ended by the even 𝑌 luminance
signal. These sequence is repeated 128×.
MP modes also have narrowband variants (MPxx-N) and their video signals oc-
cupy frequencies from 2044 to 2300 Hz.
The MR and ML modes use YCrCb color coding in 4 : 2 : 2 format, same as the
Robot 72 Color mode. Horizontal syncs are same as in MP modes. The scan-line
begins with luminance 𝑌, then 0.1 ms gap is sent followed by 𝑅 − 𝑌, a gap, and
𝐵 − 𝑌, the line then ends with a 0.1ms gap. These gaps should have the same

49
4 Formats of slow-scan TV transmission

frequency as the last pixel of the previous color component. The MLxx group has a
high resolution of 640×496.
The MC-N modes are narrowband, but they use RGB color coding. Horizontal
pulses last 8.0 ms and are followed by a 0.5 ms gap of 2044 Hz. The order of color
components is red – green – blue.

Mode Transfer VIS Color Scan-line (ms) Speed

Resolution
name time 16-bit sequence Sync Y R–Y B–Y (lpm)
MP115 115 s 320×256 0x2923 YCrCb 9.0 223.0 223.0 223.0 133.037694

MP140 140 s 320×256 0x2a23 YCrCb 9.0 270.0 270.0 270.0 110.091743

MP175 175 s 320×256 0x2c23 YCrCb 9.0 340.0 340.0 340.0 87.591241

MR73 73 s 320×256 0x4523 YCrCb 9.0 138.0 69.0 69.0 419.140761

MR90 90 s 320×256 0x4623 YCrCb 9.0 171.0 85.5 85.5 340.618791

MR115 115 s 320×256 0x4923 YCrCb 9.0 220.0 110.0 110.0 266.489007

MR140 140 s 320×256 0x4a23 YCrCb 9.0 269.0 134.5 134.5 218.858289

MR175 175 s 320×256 0x4c23 YCrCb 9.0 337.0 168.5 168.5 175.361683

ML180 180 s 640×496 0x8523 YCrCb 9.0 176.5 88.25 88.25 330.305533

ML240 240 s 640×496 0x8623 YCrCb 9.0 236.5 118.25 118.25 248.292986

ML280 280 s 640×496 0x8923 YCrCb 9.0 277.5 138.75 138.75 212.276667

ML320 320 s 640×496 0x8a23 YCrCb 9.0 317.5 158.75 158.75 185.960019

Narrowband modes:
Mode Transfer Color Scan-line (ms) Speed
Resolution N-VIS
name time sequence Sync Y R–Y B–Y (lpm)
MP73-N 73 s 320×256 0x02 YCrCb 9.0 140.0 140.0 140.0 210.526316

MP110-N 115 s 320×256 0x04 YCrCb 9.0 212.0 212.0 212.0 139.860140

MP140-N 140 s 320×256 0x05 YCrCb 9.0 270.0 270.0 270.0 110.091743

Sync R G B

MC110-N 110 s 320×256 0x14 R–G–B 8.0 143.0 143.0 143.0 137.142857

MC140-N 140 s 320×256 0x15 R–G–B 8.0 180.0 180.0 180.0 109.389243

MC180-N 180 s 320×256 0x16 R–G–B 8.0 232.0 232.0 232.0 85.166785

Table 4.10: The parameters of MMSSTV modes.

4.4.4 Martin HQ
The Martin HQ system from Martin Emmerson’s workshop was released at the end of
1996. These modes were developed for Robot 1200C, SUPERSCAN 2001 and other
compatible converters with the EPROM version 4.6, or 1.6. Unlike previous Martin

50
Formats of slow-scan TV transmission 4

modes, they use YCrCb color coding. The transmission time of chrominance signals
is half of luminance (format 4 : 2 : 2). There are 6 signals sent between two doubled
syncs. The first three signals create an odd scan-line: luminance 𝑌, 𝑅 − 𝑌, 𝐵 − 𝑌.
And the next three signals contain even scan-line: luminance 𝑌, 𝑌 − 𝑅, 𝑌 − 𝐵.
The opposite “polarity” of chrominace compensates for possible color distortion
when signals are not tuned precisely. The HQ1 mode has 90 seconds for image
transmission and HQ2 has 112 seconds.
Unfortunately, the author refused to disclose the exact specification of the system,
so this improved system is not commonly found.

2300
Frequency f [Hz]

1900

1500

1200

0 100 200 300 400 500 600 700 800 900

Time t [ms]
Figure 4.12: Two scan-lines of Martin
HQ modes, when color bars are sent.

51
 5
List of SSTV modes

System Mode Colors VIS code Duration [s] Lines Columns lpm
𝑎
AVT 24 RGB 64,65,66,67 31 128 128 960.000
AVT 90 RGB 68,69,70,71𝑎 98 240 256 480.000
Amiga Video 𝑎
AVT 94 RGB 72,73,74,75 102 200 320 384.000
Transceiver
AVT 188 RGB 76 ,77,78,79
𝑏 𝑎 196 400 320 384.000
𝑎
AVT 125 BW BW 80,81,82,83 133 400 320 192.000
FAX480 BW 85 138 480 512 224.497
FAST FM YCrCb 𝑓
90 13 240 320 1118.881
M1 RGB 44 114 256 320 134.395
M2 RGB 40 58 256 320 264.553
Martin
M3 RGB 36 57 128 320 134.395
M4 RGB 32 29 128 320 264.553
HQ1 YCrCb 41 90 240 320 85.055
Martin HQ
HQ2 YCrCb 42 112 240 320 68.680

52
List of SSTV modes 5

System Mode Colors VIS code Duration [s] Lines Columns lpm
𝑒
MC110-N RGB 0x14 110 256 320 137.143
MC140-N RGB 0x15𝑒 140 256 320 109.389
𝑒
MC180-N RGB 0x16 180 256 320 85.167
MP73 YCrCb 0x2523 𝑑 73 256 320 210.526
MP115 YCrCb 0x2923 𝑑 115 256 320 133.038
MP140 YCrCb 0x2a23 𝑑 140 256 320 110.092
MP175 YCrCb 0x2c23𝑑 175 256 320 87.591
𝑒
MP73-N YCrCb 0x02 73 256 320 210.526
𝑒
MP110-N YCrCb 0x04 115 256 320 139.860
𝑒
MMSSTV MP140-N YCrCb 0x05 140 256 320 110.092
MR73 YCrCb 0x4523 𝑑 73 256 320 419.141
MR90 YCrCb 0x4623 𝑑 90 256 320 340.619
MR115 YCrCb 0x4923 𝑑 115 256 320 266.489
MR140 YCrCb 0x4a23 𝑑 140 256 320 218.858
MR175 YCrCb 0x4c23 𝑑 175 256 320 175.362
MR180 YCrCb 0x8523 𝑑 180 496 640 330.306
MR240 YCrCb 0x8623𝑑 240 496 640 248.293
MR280 YCrCb 0x8923𝑑 280 496 640 212.277
MR320 YCrCb 0x8a23𝑑 320 496 640 185.960
TV-1 RGB 104 320 256 320 84.383
MSCAN
TV-2 RGB 105 320 256 320 134.530
P3 RGB 113 203 496 640 146.565
Pasokon P5 RGB 114 305 496 640 97.710
P7 RGB 115 406 496 640 73.282
PD 50 YCrCb 93 50 256 320 309.151
PD 90 YCrCb 99 90 256 320 170.687
PD 120 YCrCb 95 126 496 640 235.997
PD PD 160 YCrCb 98 161 400 512 149.177
PD 180 YCrCb 96 187 496 640 159.101
PD 240 YCrCb 97 248 496 640 120.000
PD 290 YCrCb 94 289 616 800 128.030
Proskan J120 RGB 100 120 240 320 128.046

53
5 List of SSTV modes

System Mode Colors VIS code Duration [s] Lines Columns lpm
Color 12 YCrCb 0 12 120 160 600.000
Color 24 YCrCb 4 24 120 160 300.000
Color 36 YCrCb 8 36 240 320 400.000
Color 72 YCrCb 12 72 240 320 200.000
Robot 𝑐
B&W 8 BW 1, 2, 3 8 160 120 900.000
B&W 12 BW 5, 6, 7𝑐 12 320 240 600.000
𝑐
B&W 24 BW 9, 10, 11 24 320 240 300.000
𝑐
B&W 36 BW 13, 14, 15 36 320 240 200.000
S1 RGB 60 110 256 320 140.115
S2 RGB 56 71 256 320 216.067
Scottie S3 RGB 52 55 128 320 140.115
S4 RGB 48 36 128 320 216.067
DX RGB 76 𝑏 269 256 320 57.127
Scottie DX2 RGB 80 136 256 320 112.905
SP-17 BW BW 125 17 256 128 895.520
Vester RGB 86 414 480 512 74.832
Color FAX
8 BW 17, 18, 19𝑐 8 120 128 1000.000
𝑐
16 BW 21, 22, 23 16 120 256 500.000
24 BW 25, 26, 27𝑐 24 256 128 930.520
𝑐
32 BW 29, 30, 31 32 240 256 500.000
Wraase SC1
24 RGB 16 24 128 128 900.000
48 RGB 20 48 128 256 489.102
48Q RGB 24 48 256 128 900.000
96 RGB 28 96 256 256 500.000
30 RGB 51 30 128 320 249.595
60 RGB 59 60 256 320 249.600
Wraase SC2
120 RGB 63 120 256 320 126.175
180 RGB 55 180 256 320 84.383

Notes:

⊳ 𝑎 – VIS code order: Normal, Narrow mode, QRM mode, Narrow+QRM;

⊳ 𝑏 – Scottie DX and AVT 188 were created at the same time and accidentally share the same
VIS code;
⊳ 𝑐 – VIS codes for each color component (red, green, blue). The green component is comonly
used for BW images;
⊳ 𝑑 – Uses 16 bit VIS (see sec. 4.4.3);

54
List of SSTV modes 5

⊳ 𝑒 – Uses another digital code N-VIS (see sec. 4.4.3);

⊳ 𝑓 – Uses odd parity;

55
 6
SSTV Equipment

6.1 Transceiver
SSTV images can be received via a standard communication transceiver (or receiver)
that covers the HF amateur bands and supports SSB modulation or a VHF trans-
ceiver with FM. There is no need for further modifications, although the IF filter
should not be narrower than 2.5 kHz, a width of 3 kHz is recommended. The SSTV
signal is taken from the audio output or headphone jack that is plugged into the
SSTV decoding device.
For transmission, a common HF or VHF transceiver with an SSTV signal con-
nected to the microphone jack should be used. The band in which you are determines
the usage of LSB or USB, which is the same as in voice transmission.
No transmitter modifications are necessary. But when operating SSB, it is im-
portant to realize that the broadcast of very loud speech can be achieved at approx-
imately 3 /4 of output power, so the load of the output amplifier is fairly low and
there is no danger of overheating. But SSTV signals transmitted via a voice channel
create a 100% load due to the presence of an auxiliary carrier.

Keep the limits set by the manufacturer for SSB operations. The usual max-
imal keying with full load is about 20 minutes for professional equipment.
Modern transceivers switch on cooling during heavy load and it is not suitable
to switch off TRX immediately after the end of QSO, but wait a few minutes
for the equipment to cool down.

SSTV transmission is not dangerous, but it is advised to adhere to some safety

measures.

6.2 Station equipment for visual communication

1. Computer system – a sound card or other special hardware interface and software.

56
SSTV Equipment 6

video SSTV
digitizer modem

Figure 6.1: SSTV station equipped for computer operations.

Currently, the most common SSTV device is a personal computer with a sound
card. There are a number of programs for personal computers with Windows,
Mac, GNU/Linux and DOS. There are also special modems (MFJ, Roy1, AOR
TDF370,…) or the very simple Hamcomm modem. Hamcomm is based on a
simple comparator circuit and connected to the RS232 serial port. But it is
only usable for old DOS based software. The most varied software options are
for Windows and a sound card.
Additional equipment can be used such as a web camera or a television card
with analog video input.
2. Digital scan-converter is a stand-alone device that digitizes received signals and
stores them in memory. The decoder converts memory content to analog signals
(PAL or NTSC) for display on a normal TV set or monitor. The converter can be
connected to any color or monochrome camera, which then transmits live images.
Due to digital data processing, most converters are equipped with a computer
interface. This allows for the addition of texts to images and the upload and
storage of images to/from computer. Tape recorders were historically often used
to record SSTV and for storing QSO images.
3. Long persistence CRT monitor and circuits for signal filtering and vertical and
horizontal drives, etc. The usage of these monitors is long over. Electro-
mechanical scanners or sampling cameras were used as SSTV signal sources in
these days.

6.3 Historical tidbits

6.3.1 SSTV monitor

Long persistence monitors were the most important SSTV equipment in the seven-
ties. There were other commercial products available, but most homemade monitors
were built by SSTV enthusiasts.

57
6 SSTV Equipment

SSTV
converter

Figure 6.2: An SSTV station equipped with a stand-alone SSTV converter.

Figure 6.3: Monitor Robot Model 70 and camera Model 80 from

Robot Research Inc.

Products from Robot Research Inc., Wraase Electronics and Venus were very
popular. These products were not produced for just the ham radio market, but
were also found in the image communication over telephone lines markets.
A typical monitor consists of several basic parts, see fig. 6.4: input and limiter
circuits, video and sync detectors, scanning circuits, cathode ray tube drivers and
power supply. The long persistence CRTs are made with special photoluminescence
phosphor. In simple terms, phosphorescence is a process in which the energy ab-
sorbed by a substance is slowly released in the form of light. These CRTs were most
used in radar displays or oscilloscopes for the monitoring of slow processes.
The path of frequency modulated signals that contains video and syncs goes
through the limiter, where the signal is limited to constant amplitude and then
flows into image discriminator. There are video detecting circuits for syncs and

58
SSTV Equipment 6

video separation here. Then signals from the discriminator are amplified and drive
both the vertical and horizontal scans. The output voltage of these circuits is the
saw-tooth voltage and drives deflection plates of long persistent CRT.

SSTV image image

limiter discriminator
input amplifier

sync
separator detector

sync
amplifier
low pass

sync
rectifier

vertical horizontal
scan scan

Figure 6.4: Block scheme of SSTV monitor.

After the separation of sync, the image signal goes to the image intensifier and de-
tector. After filtration, the signal is fed to the cathode ray tube grids and modulates
the electron beam. Subsequently the image is displayed on the screen. The disad-
vantage of this process is that the image is seen clearest during the reception and
then the brightness fades. To be able to view the image after the 8 seconds transfer
a well darkened room was necessary.

6.3.2 Scanning devices

The image scanning methods used in early SSTV transmission can be classified into
electronic and electro-mechanical methods. Purely electronic systems used cameras
with a sensor element like vidicon, plumbicon or other camera tubes.
In SSTV cameras, the vertical scan frequency was adjusted from 50 Hz to 16.6 Hz
(i.e. horizontal scanning frequency for 7.2s SSTV or 15 Hz for 60Hz standard). Then
a whole camera or just a deflection unit only was rotated 90∘ . The TV camera scans

59
6 SSTV Equipment

the image line by line, providing the sampling circuit with input. The sampling
circuits reads short samples from each line during each camera scanning beam cycle.
All the samples from scan-lines of the FSTV camera create one scan-line of slow-
scan TV picture. In the next scanning beam cycle, the sampling position moves to
the left and creates the next scan-line. The cycle is repeated until the whole picture
is sampled.
The next type of image scanner often used was a scanner with a photomulti-
plier for the scanning of transparent or non-transparent originals (FSS – Flying
Spot Scanner). The light through transparent originals falls on the photomultiplier,
whose output is a voltage that is proportional to the transparency of the original.
This creates an amplitude modulated video signal, which can be converted to the
frequency modulated signal of SSTV.
The electromechanical scanner was used for non-transparent originals, which were
scanned from a rotating roller. The mechanical part was assembled from a roller with
the mounted original, a screw-thread for sliding and a drive unit with a synchronous
motor. The second part consisted of a lens, a light bulb, a photo transistor and
sensor circuit for the generation of the SSTV signal.

6.4 Early FSTV/SSTV converters

SSTV/FSTV converters usually sample and digitize incoming SSTV signals and
store them in memory. Simultaneously, the memory content is read and converted
to an analog signal, which controls the fast-scan TV modulator.
The received SSTV signal is limited to the constant amplitude in input circuits
and then continues into an analog/digital converter. Digital data is processed by
the converter’s microprocessor firmware. Its task is to digitize every scan-line of the
image and store it in memory. The memory capacity is equivalent to the resolution
and number of colors.
The memory is continuously read in the FSTV scanning frequency and data goes
into a digital/analog converter. The output analog signal is displayed on normal
television. The SSTV image is stored in memory until it is overwritten by the
newly received image. The reverse process of digitization of an FSTV image and its
transmission by SSTV is similar.
One of the first SSTV converters was the Robot 300 model. This converter con-
tains 69 transistors, 41 integrated circuits, 41 diodes and its heart is a silicon memory
tube. The function of this tube was the same as a cathode ray tube or vidicon. The
electron beam electromagnetically diffracted and focus was directed into the scan-
ning electrode, which consisted of a dielectric memory layer used for analog image
recording.

60
SSTV Equipment 6

The modern concept of converters began with Robot 400C and its successors 450C
and especially 1200C, which became available in 1986. In these years everyone, who
was serious about SSTV had to own one! A camera and a monitor were necessary.

RED RGB D/A

memory and controller

composite
NTSC / PAL video for
RGB A/D GREEN
encoder TV monitor
multiplexer converter memory

NTSC / PAL
video BLUE
to RGB RF RF output
input memory
converter modulator for TV
receiver

system display
clock controller

SSTV FM signal SSTV

input zero detector COMPUTER FM modulator output

interface for
front panel
computer and
control
printer

Figure 6.5: Block scheme of stand-alone SSTV/FSTV converter.

Robot 1200C was modified with optional EPROMs and a timing circuit, so it can
be used for operations in various SSTV modes. Its production ended in 1992, but
during the nineties there were clones available on the market – FH-21P in Germany,
SUPERSCAN 2001 in the United Kingdom, LM-9000C and Ribbit 1200C. These
machines can still be used for SSTV operations and many old-school operators own
them. They are also sometimes available in Ebay auctions.

6.5 SUPERSCAN 2001

The production of SUPERSCAN started shortly after the end of Robot 1200C pro-
duction. Its designer Jad Bashour had worked with Martin Emmerson. SUPER-
SCAN is actually a much improved 1200C and includes improved modifications to
the original 1200C. The price of this unit was about £ 750 and with additional mod-
ules the price could exceed £ 1000.

61
6 SSTV Equipment

Figure 6.6: The front panel of SUPERSCAN 2001

Main features of SUPERSCAN 2001:

⊳ Total compatibility with all classic SSTV systems.

⊳ System upgrade is simply achieved by an EPROM upgrade. Its last version 1.6
supports these modes:
⊳ Color modes:
⊳ Scottie S1, S2, S3, S4, DX;
⊳ Wraase SC-1: 24, 48Q, 48, 96;
⊳ Robot Color 12, 24, 36, 72;
⊳ Wraase SC-2: 30, 60, 120, 180;
⊳ AVT 24, 90, 94, 188, plus QRM, Narrow variants.
⊳ B&W modes:
⊳ Robot 8, 12, 24, 36;
⊳ Wraase SC-1 8, 16, 16Q, 32;
⊳ AVT BW 125.
⊳ Radio fax reception:
⊳ 60, 90, 120, 240 lpm.
⊳ Contains four memory banks and stores images in a resolution 256×240 with
18bit color depth (262,144 colors).
⊳ TV PAL decoder with delay lines for perfect image digitization.
⊳ High speed parallel interface for computer connection.
⊳ RGB video output.
⊳ Control by computer mouse available (firmware 1.3).
⊳ Text addition.
⊳ Backup of CMOS memories for texts and configuration.
⊳ High stability oscillator for free-run reception.

62
SSTV Equipment 6

6.6 Tasco TSC-70P

A modern type of converter is the TSC-70P (TSC-70N works with NTSC norm).
This converter includes a DSP for better reception of weak signals. It supports all
conveniences such as the automatic detection of VIS code and free-run reception.

Figure 6.7: Japan converter Tasco TSC-70P.

Supported modes:

⊳ Martin M1, M2;

⊳ Scottie S1, S2;
⊳ Robot Color 36, 72;
⊳ AVT 90, 94 (Narrow regime, only in TSC-70N available).

Image processing is done in a real-time digitizer, and images are stored in memory
with a resolution of 416×256 with 2 million colors. With an optional EM-70 module,
the video memory capacity can be doubled.
Tasco TSC-70P working with PAL video signal, you need to have a television with
video inputs or a color TV monitor. Control equipment is made via an infrared
remote control (WR-70) or via a RS232 serial interface. Using EB-232VP software,
images can be moved at 115 kbit/s speed between the converter and the computer.
The desktop PC can be equipped with an optional ISA card EB-70P that triples
the speed of data exchange. For greater convenience the converter can be controlled
by a computer program i.e. HIRES-70P or WINTSC.
The converter weighs 450 g (60 g remote control) with dimensions of 140 mm
(width) × 140 mm (length) × 25 mm (height) and it is powered by DC 11–15 V
with a consumption smaller than 250 mA. It is specifically designed for mobile or
portable operations and can be used with miniature television, such as EV-5xx from
CASIO, with small LCD display 7 cm, it weighs about 195 g.

63
6 SSTV Equipment

6.7 Interactive Visual Communicator VC-H1

The VC-H1 was produced by Kenwood. It is a device intended for mobile SSTV
operations. The dimensions are similar to hand-held transceivers – 7×3.5×17 cm.
The unit has a built-in CCD Camera, 1.8" LCD color monitor and a microphone.
Its memory allows the storage of an uncompressed image in the image buffer and
10 JPEG compressed images. The JPEG memory has a battery backup, so it is
possible store images when the unit is switched off.
The converter has an RS232 interface for computer connection (115 kbit/s). The
input and output for external video signals is only NTSC.
VC-H1 is powered by four AA batteries or external DC supply with 6.0 V. The
maximum consumption is 650 mA when digitizing an image, otherwise the consump-
tion is 450 mA when the LCD is on or 100 mA in stand-by mode with the LCD
off.

Figure 6.8: Mobile

SSTV converter VC-H1.

Supported modes:

⊳ Martin M1, M2;

⊳ Scottie S1, S2;
⊳ Robot Color 36, 72;
⊳ AVT 90, 94;
⊳ FAST FM.

64
 7
Computer operations

A fully equipped hamradio operator can set up SSTV reception in an hour. Start
with creating a connection between a transceiver and a sound card using a compati-
ble cable. Then,download relevant SSTV related software and tune into 14.230 MHz
using upper-sideband (USB). There is round the clock SSTV activity around this
frequency.
The available software for SSTV operations are very similar and they all provide
the same basic functions, of course with different comfort levels. Some of them are
intuitive and well-arranged, while others require studying a manual. There is the
possibility to choose from more variants and see what works best for you. Every
operator must know how to carry out these basic tasks:

⊳ configuring an accurate sampling frequency for reception and transmission,

⊳ setting proper sound card volume levels,
⊳ using the tuning indicator,
⊳ manually change the SSTV mode,
⊳ loading and saving images in common graphic formats,
⊳ creating an image gallery for transmission and
⊳ adding text into transmitted images.

7.1 Hardware configuration

Take your time in selecting a suitable configuration for the computer in your ham-
shack. The unpleasant fact is that, operating systems and distributed software
constantly require better CPUs and more memory. For SSTV operations it is still
possible to use obsolete PCs with old 386 processors. The Hamcomm modem and
some other popular programs from the nineties run on DOS. In this case, 4 MB of
RAM, a hundred megabytes hard disk and SVGA graphics card with 256 colors are
enough. But graphic cards with 32 or 64 thousand colors are more suitable.
To use a sound card, a good PC is necessary. The PC should be configured
with at least a 150MHz Pentium, 64 MB RAM and few gigabytes hard disk and a

65
7 Computer operations

graphics card that supports at least a 1024×768 resolution in 64 thousands or 16.7

million color mode. The minimal operating system is Windows 95 OSR 2, but some
new programs may not run in Win95. I recommend having better hardware but you
don’t need the latest mega-hyper model.
There are also software products for Mac OS X and for GNU/Linux, but the largest
selection is for Microsoft Windows.

7.2 Using a sound card as a modem

A sound card is standard PC equipment these days. It can be used as a music
player for multimedia, games and recording. The main application of a sound card
in hamshacks is as a MODEM. The modem – MOdulator/DEModulator allows
computer information to be transmitted and received over physical media like radio
waves or telephone lines. The modem translates analog signals to digital data and
vice versa.

7.2.1 Sound processing in PCs

To allow the computer to work with sound signals, the signal must be converted
into a format suitable for data processing (digital or discrete signal).

7.2.1.1 Sampling
The digital conversion process begins with sampling. Sampling is an activity which
periodically scans the current value of the analog signal. For example, this happens
11,025 times per second or depends on a user defined sample rate supported by
the sound card. The sampling frequency of sound cards ranges from 8 kHz suitable
for internet telephony up to 96 kHz designated for more exacting requirements of
recording studios.
The answer to the question of what sampling frequency should be used is given by
Shannon’s theorem (also knows as Nyquist-Kotělnik-Shannon theorem). It defines
that a signal continuous in time, containing spectral components with the highest
frequency 𝑓𝑚𝑎𝑥 , can be clearly reconstructed from a sequence of evenly spaced sam-
ples with a sampling frequency 𝑓𝑠 greater than double 𝑓𝑚𝑎𝑥 :
𝑓𝑠 > 2𝑓𝑚𝑎𝑥

You can see the importance of Shannon’s theorem in the example below. The signal
in figure 7.2a expresses the dependency between time 𝑡 and amplitude 𝐴. Using
Fourier’s analysis, we can find (see chapter 3.1.1), that the signal contains two
harmonic components, showed in 7.2b.

66
 Computer operations 7

A A N7
6
5
4
3
2
1
0
t n n
100100010001011101110…
sample analog to
and digital
hold converter
analog input samples of analog digital data
signal

Figure 7.1: The conversion of an analog signal into numeric data.

A A

t t

a) Surveyed signal b) Spectral components

Figure 7.2: Example of signal.

By using Fourier’s transformation, the signal can also be expressed as a depen-

dency between amplitude 𝐴 and frequency 𝑓 – signal spectrum. Both frequency
components are evident 𝑓1 and 𝑓2 = 𝑓𝑚𝑎𝑥 in the signal in fig. 7.3.
For explicit signal reconstruction the condition 𝑓𝑠 > 2𝑓𝑚𝑎𝑥 must be satisfied, see
fig. 7.4. If a sampling frequency is lower than 2𝑓𝑚𝑎𝑥 then the higher frequency
components causes signal distortion. This error is called aliasing.
For the sampling of common narrow band signals like SSTV, RTTY, PSK31
or WEFAX, which are transferred via SSB channel with a bandwidth of about
2 500—3 000 Hz; a sample rate of 11,025 Hz is enough.

7.2.2 Analog-to-digital conversion

The analog signal sample continue into the analog-to-digital (A/D) converter. The
current value of the signal is converted into digital data in this device. Some A/D

67
7 Computer operations

A
f
A

f1 f2 f

Figure 7.3: The frequency spectrum of signal in fig. 7.2a

A A

t n

Figure 7.4: The signal sampled with a frequency higher than 2𝑓𝑚𝑎𝑥

converters work with a resolution 8 or 16 bits according to the type or settings of

a sound card. The resolution of an A/D converter indicates the accuracy of signal
amplitude scan in a defined range, for 8 bits it is 28 = 256 values and for 16 it is
65,536.
A constrained resolution of the A/D converter causes a quantization error. E.g.
for an 8 bit converter processing voltage range 0–5.0 V it’s the error 5.0/(28 −
1) ≐ 0.02 V. The 8bit converter cannon distinguishes between voltage levels lower
then 0.02 V. So for input voltage 3.111 V it could find a corresponding numeric
value 100111102 ≈ 3.098 V or 100111112 ≈ 3.118 V because less significant bits
are influenced by the quantization error. The size of the error can be decreased
by greater resolution of A/D converter. For our purposes the 16bit resolution is
acceptable.

68
Computer operations 7

A A

t n

Figure 7.5: The sample rate does not meet the Shannon’s theorem.

A modern sound card could be equipped with a digital signal processor (DSP),
which extends card functions e.g. for filtering or data compression during recording
so it can lighten the load of the main computer CPU. E.g. Sound Blaster Live!
contains a programmable DSP labeled EMU10K1.
The choice on which sound card to use, depends solely on the user’s preferences
and intentions. With the rise of cheap Software Defined Radio (SDR) solutions,
always choose a sound card with a higher sampling rate and analog anti-aliasing
circuits at the input. Many PCs come with an integrated sound card on the moth-
erboard.

7.2.3 Interface between TRX and PC

The basic interface is made of shielded cables and 3.5mm jack plugs. A reception
cable connects sound card input Line In and TRX headphones output or output for
external speakers. To use the sound card microphone input, use the TRX output
often labeled as AF OUT with an impedance of about 10 kΩ which gives a maximum
voltage output of 100 mV. This output could also be used for interfacing tape decks or
audio amplifiers.A sound card’s microphone input has an automatic gain controller
(AGC) for better recording and it is possible to connect dynamic microphones with
impedance ranges from 600 to 10.000 Ω.
For the transmission it is possible to use Line Out with impedance of about 600 Ω.
The Line Out can be connected to the microphone input of the TRX or a rear panel
connector like PATCH IN.
Some transceivers have a feature that the microphone input and rear panel input
are interconnected, so it is necessary to disconnect the microphone during AFSK
transmission, because noise in the hamshack could interfere with the sound card’s
signal! Check your TRX instruction manual for relevant information.
The last thing you need to set up, is the audio level for received and transmitted
signals. This can be done using operating system tools. The level of transmitted

69
7 Computer operations

Reception from TRX

Line In

Microphone
Transmission to TRX

Line Out

Speaker Out

Figure 7.6: Basic interface between transceiver and

sound card.

signal should be about 2/3 of maximum volume levels. The signal should not be
too attenuated or over-excited and distorted. You can detect it by monitoring the
outgoing signals. For the incoming signal you can set the proper level directly on
the TRX and check the input level in your SSTV software.

Figure 7.7: Software volume control.

After the audio mixer configuration, it is useful to save your sound card settings (
so you can easily restore it before SSTV sessions). A program such as QuickMix can
easily store your settings. This is important because other PC program can change
sound card levels.

70
Computer operations 7

http://quickmix.softpedia.com/

7.2.4 PTT control

The PTT (Push-To-Talk) button switches between reception and transmission. There
are several methods used for PTT control:

1. Manual PTT switching; this method is not very elegant, but can be used for
initial experiments.
2. Automatic switching can provide the TRX with a VOX feature, when the TRX is
automatically keyed by signal in the input. A disadvantage of this method may
be that its reaction is not immediate, so in case of digital modes the beginning
states of transmission or SSTV vertical synchronization can be lost. Keep in mind
that operating systems often produce malicious sounds that could accidentally
key the transmitter.
3. Computer controlled automatic PTT switching; all SSTV programs support PTT
control over a simple serial port (COM, RS-232) circuit. The circuit contains one
switching transistor or opto-isolator and few passive parts. See schematic 7.8
for details. The control signal is connected to RTS pin (7 at Cannon DB9 connec-
tor, 4 at DB25) or DTR (4 at DB9, 20 at DB25), the selected pin can be changed
by software configuration. The ground is on both serial ports wired on pin 5 at
DB9 and 7 at DB25.
A large number of handheld TRXs use the same connector for both microphone
input and PTT. In this case an audio signal should be galvanicaly separated
by the capacitor at about 100 nF and the PTT signal can be connected by a
resistor; whose resistance can be found in the TRX instructions or you can
connect a trimming resistor at about 15 kΩ and test the max. value when TRX
switching.
4. Some transceivers support control over the serial port. The CAT (Computer
Aided Transceiver) interface can provide PTT switching. Commands can also
be sent over the CAT interface e.g. for tuning, mode control, etc. This method
must be supported by the installed software, for example MixW can control some
TRXes so it is not needed to practically touch the TRX buttons.

What can you do if your computer is not equipped with a serial port? Some
motherboard manufacturers build only one serial port and notebook manufacturers
don’t provide any. If this happens, you can use VOX or buy a USB/RS232 interface.
Some programs also support similar switching circuits as described before but on
parallel port (LPT).

71
7 Computer operations

7.2.5 Eliminate supply noise

A computer and a TRX can have slightly different electrical potential and in this
case the direct connection causes annoying noise within the communication channel.
It is possible to remove the noise by the galvanic separation of both devices. The path
of the audio signal should go through the galvanic transformer and PTT control
switch with an opto-isolator, e.g. 4N25, 4N33, etc. You might need to change R2 to
a lower value if the opto-isolator is not switched on properly when serial port PTT’s
signal is active.

Figure 7.8: The galvanic separation of transceiver and

sound card.

7.3 Timing oscillator configuration

There is a description of synchronous (free-run) SSTV system in section 3.6.1.
Horizontal synchronization pulses (syncs) are detected only at the beginning of re-
ception and after synchronization, a reception device stops detect syncs and receive
with free-run scan. Due to this there are excessive requirements for accurate timing
of corresponding stations.
If the timing slightly differs then images are distorted – inaccurate timing causes
image slant. You can see image slant for 0.01 % timing difference in fig. 7.9. If a
transmitting station has higher timing (and reception lower) the image slants to the
right (7.9a) in opposite situation to the left (7.9b).

72
Computer operations 7

a) Opposite station b) Opposite staion

has higher timing has lower timing
Figure 7.9: Image slant distortion when inaccurate timing is
used for free-run modes.

However timing derived from sample rate is not used to be exactly 11,025.00 Hz, but
often can differ up to few tenths of percent for each peace of hardware. For speech
and music processing it doesn’t matter, but in free-run transmission of SSTV it
causes problems.
The configuration of accurate timing/sample rate for reception and transmis-
sion apart must be done to meet the strict requirements for synchronous SSTV
broadcast. Your signal must be acceptable for any SSTV device.
All SSTV program are equipped with a tool for the timing configuration.
It is possible to receive SSTV signal from a band and by an edge of received image
set the timing – program will automatically compute timing deviation. This way
has a disadvantage, because not all SSTV stations has proper transmit timing. This
is caused by offset between receiving and transmitting timing.
Much more precise way is use of shortwave broadcast of timing normal. Programs
are equipped with special reception option, which displays spectrum in a second
cycles. For timing setting just tune to the frequency of broadcast and leave to plot
received pulses for several minutes.
The usable transmitter is a Moscow station RWM operating on frequencies 4,996.0,
9,996.0, 14,996.0 kHz with 8 kilowatts power. So it can be nicely received in Eu-
rope/Asia region. Select CW mode and tune your receiver directly to one of station
frequencies. The unmodulated carrier is transmitted between 0. and 8. minute of
an hour, telegraphy identification goes from 9. minute and then the timing signal
will continue. Pulses in intervals 1/60 and 1 Hz goes between 10. and 20. minute
and 10Hz pulses goes between 20. and 30. minute. This is repeated every 30 min-
utes.

73
7 Computer operations

Figure 7.10: The configuration of accurate timing with RMW re-

ception in MMSSTV.

The reception of WWV station is the next possibility. This station broadcast tim-
ing pulses and announcement on frequencies 2,500.0, 5,000.0, 10,000.0, 15,000.0,
20,000.0 kHz and uses double sideband (DSB) modulation. You can receive it with
AM mode selected. The WWV operates from the North America, Fort Collins in
Colorado. The used power ranges from 2.5 to 10 kW.
There is yet another way with WEFAX station reception, because these stations
must have accurate timing too due to synchronous transfer.
The deviance error you should measure use to be expressed like absolute value of
actual frequency, e.g. 𝑓 = 11024,45 Hz or like deviance from 𝑓𝑠 the Δ𝑓 = −0,55 Hz.
Some program this measure in parts per million (ppm) unit. The ppm deviance can
be computed:

Δ𝑓
Δ= ⋅ 106 .
𝑓𝑠
For the 𝑓 = 11024,45 Hz the deviance in ppm is:

Δ𝑓 −0,55
Δ= ⋅ 106 = ⋅ 106 ppm ≐ −50 ppm.
𝑓𝑠 11025,00

7.3.1 Transmit timing offset

There is necessity to configure transmit timing TX offset after the precise config-
uration of reception sample rate, when received SSTV images are not slanted. It
is important for your own transmission, because inaccurate transmit timing causes
image slant on reception side.

74
Computer operations 7

Some programs makes possible to monitor outgoing SSTV signals, so with this
feedback it is practicable to check the TX offset – deviance between reception and
transmission sample rate. The feedback can be internal or external. Extenal feedback
needs to connect Line Out and Line In with cable and it requires a sound card with
full-duplex mode enabled. By this way you can set TX offset precisely on your own.
The internal feedback doing almost the same, but no external cables is needed. But
some sound cards support only software feedback, so you will find zero deviance, but
it is not real fact! Then the TX offset setting must be done with external feedback or
with opposite station help. You need to disable any automatic corrections of received
signals in this way.
Anyway you need to make “dry run” QSO before your first CQ. This helps you
to uncover possible problems with TX offset, supply noise, audio level, etc.
The TX offset issue is often pretty messy. You can notice that some software
running concurrent with your SSTV program can influence sound cart output and
then the change of sample rate occurs. Even the simple Volume Control tool can
do this. So it is useful to stop unnecessary program running in the background.
Especially programs that can influence sound card output or decreases stability of
Microsoft Windows.
You may notice a strange behaviour if you are user of modern sound card with
full duplex mode enabled with several output channels with a support of different
sample rates for each channel. This is for example SB Live! Value. I have noticed
that my TX offset randomly changes! I have this experiences with SB Live! Value
and I found that another radio amateurs has same. You can try to set other sample
rate than 11,025.0 Hz in this case, if this doesn’t load your computer too much. For
example try 48,000.0 Hz, this value is fixed sample rate (see your card user’s guide)
and best results you can achieve with using of this value or its half or quarter –
24,000.0 kHz or 12,000.0 kHz. When you change this value you need to recalibrate
your accurate timing again.
To avoid these problems you can constantly monitor the outgoing signal through
the external feedback with the TRX monitor enabled and before your today first
transmission you will check that everything is fine. It’s unpleasant that problems
often occurs during QSO and then you will stunned by counterpart replay images.

7.4 SSTV tuning

First of all, we need to find SSTV stations by listening near calling frequencies.
Thanks to typical SSTV sound and clattering of syncs it is not a problem to distin-
guish between SSTV and other communication modes.
Every SSTV program is equipped with precise tuning indicators – spectroscopes,
see fig. 7.11. The spectroscope shows frequency band from 1000 Hz to 2500 Hz with

75
7 Computer operations

marks for critical frequncies – 1200 Hz for syncs, 1500 Hz and 2300 Hz for the video
signal.
Is is possible to simple detect band of video signal and syncs during clear reception.
Rotate the tuning knob to achieve that all important frequencies are aligned in
spectroscope display.

Chroma Pix MMSSTV JVComm32

MSCAN QSSTV MixW

Figure 7.11: Spectroscopes in common SSTV programs.

7.5 Video digitalization

The video digitizer should be additional equipment of SSTV station. The device
can convert output signal from camera into computer form. There is a great choice
of many different devices with varying capabilities, parameters and price. You can
choose some webcams, frame grabbers, TV cards or digital cameras. Then your
broadcast will not be limited only to pre-prepared images and you will have a lot
more fun with live transmission.
The cheapest option are web cameras, they are equipped with a low-resolution
CCD and low-cost optics, but provided quality is suitable for SSTV.
An another option is a TV card with video input. This possibility is more ex-
pensive because you must connect an external camera where the choices ranges
from cheap CCTV black and white or color CCD cameras up to professional studio
equipment.

7.6 Software for Windows

7.6.1 List of programs

7.6.1.1 SSTV software

76
Computer operations 7

⊳ Chroma Pix – http://www.barberdsp.com/

⊳ JVComm32 – http://www.jvcomm.de/
⊳ MMSSTV – http://mmhamsoft.ham-radio.ch/
⊳ MSCAN SSTV, Meteo – http://www.mscan.com/
⊳ Winskan, SSTV32 – http://webpages.charter.net/jamie_5/
⊳ W95SSTV – http://www.barberdsp.com/w95sstv/w95sstv.htm

7.6.1.2 Digital mode software with SSTV support

⊳ MULTIPSK – http://f6cte.free.fr/index_anglais.htm
⊳ MixW – http://www.mixw.net/

7.6.1.3 Software for dedicated interfaces

This programs don’t use sound card as modes but special interfaces.

⊳ Bonito Radiocom – http://www.computer-int.com/rc.htm

⊳ Roy 1 – http://www.roy1.com/dvb_ham/dvb_1.htm
⊳ Wraase SC-4 – http://www.wesacom.de/sstv/

77
 8
Ham radio image
operations
SSTV and radiofax broadcasting has their own rules, as well as other communication
modes. If you already have a station equipped with some SSTV device and you are
familiar with your SSTV software and have TRX interfaced with computer, you can
start your image operations.
The ideal beginning is to ask experienced operator for a first test QSO. During
this first QSO you or your partner can reveal some problems like bad settings of
synchronization rates or noise affecting your computer signal or another problem
you cannot reveal yourself. You can also contact your local SSTV party for help.
International Amateur Radio Union (IARU) recommends usage of amateur bands
and recommends specified frequencies for voice, digital and image operations. Latest
band plans dates of March 2009. For I.T.U. region I (Europe, Africa and Middle
East and Northern Asia) there are recommended frequencies 8.1in table.

Frequency Recommendation
3,735 kHz center of activity
7,165 kHz center of activity (previously 7 030 –7 040 kHz)
14,230 kHz center of activity
21,340 kHz center of activity
28,680 kHz center of activity
144,500 kHz calling frequency for SSTV
432,500 kHz narrowband SSTV
433,400 kHz SSTV (FM/AFSK)
Table 8.1: Band plan recommendations
for image communication.

78
Ham radio image operations 8

The usage of side band is same like for voice operations, on bands below 10 MHz it
is LSB and for above bands it is USB.
Before you start calling CQ make sure that chosen frequency and its neighborhood
are free. SSTV operations are recommended in same sections of bands as voice
communication and other modes so it is very unpleasant to interference each other.
So before you start transmission ask on chosen frequency: “Is this frequency free for
SSTV?” and again listen if the frequency is really free. There is unpleasant feature of
some bands, e. g. 20 meters (14 MHz), that closer stations we can’t detect, although
only a noise sounds from speakers doesn’t mean that no connection is make on the
frequency.
There are centre of activities recommended on all bands, so for stations we should
look around these frequencies. They are also can be used as calling frequency and
after the station calling CQ on the frequency makes contact, the both station should
tune to another free frequency (QSY) within the SSB segment.
Unfortunately, reality does not comply with this, so situation on very crowded
band 20 meters is such, that stations are glued to each other around 14,230 kHz,
they are interfering each other and weaker long distance stations are noised by
undisciplined nearer stations. If you find that there is activity tune to another
frequency at least ±3 kHz. Good practice is chose frequency near centra of activity
in 3kHz steps, e.g. on 15 meters:
… 21,334 21,337 21,340 21,343 21,346 …
There is great probability to find stations calling CQ or your own CQ will be
heard by the other side. You can call CQ by sending the image in the desired
SSTV mode. The image must contain code CQ. It is good practice to place CQ
test to the bottom of a picture, so a station that tunes to the frequency later finds
what is going on. If you call CQ on calling frequencu add code QSY (Change to
transmission on another frequency [(or on         kHz]) and specify frequency where
you can continue the QSO. ̅
An answering to CQ call is possible in two ways. The first one is that you answer
by sending the image always in the same mode as called station. Of course, listen
first that they are not another answering station. Add the call signs, e. g. OK1AAA
de OK2BBB and report RSV into your picture. The second method, which is less
used, you can contact the station by voice and than send your image.
Beyond that it depends on your choice what style you will prefer when you make
the SSTV connection. It is possible to communicate only in SSTV, when all infor-
mations are transmitted in pictures or use SSTV as addition for voice operations,
when both stations during QSO change few images. First way prefers mainly Eu-
ropean stations, but in northern America is preferred second one for making QSOs.
SSTV operation is closely linked with voice operation and although “one picture
can say thousand words”, it is sometimes more effective to use the microphone for

79
8 Ham radio image operations

communication. In particular, if we enter into an ongoing QSO or calling party of

more stations it is preferred to call firstly by voice before you send an image.
In earlier times operators before sending image said the mode they are use, but
now the SSTV software can automatically detect modes during transmission and
the announcing mode is obsolete.

Figure 8.1: Sample received SSTV images for your inspiration.

Broadcast images can contain almost everything, do not forget to send your own
image or photograph of hamshack, your equipment and QTH. You can also add
descriptive texts. It is suitable to get a video digitizer or small webcam and broadcast
live images.
I do not recommend send images which may be distorted during transmission
like fractals, stereograms and contain a lot of details, which is lost due to analog
transmission. Keep in mind that some picture topics might embarrass your QSO
partner.
Texts in picture should by written by some nice readable font. Choose a color
that has enough contrast with the image background. It is good practice to add
contrast edge to letters. Note, that conditions on the opposite side is not always
perfect and it should be hard to decipher small letters in noisy image.

80
Ham radio image operations 8

8.1 The reporting system

The message about report contains info about – readability, signal strength and
picture quality report. The report message is transerred as a three-letter code RSV
(Readability, Strength, Video), see table 8.2.
Readability – shows quality of signal reception in 5 degrees.
Signal strength – describes strength of received signal in 9 degrees.
Tool to determine the strength is measuring instrument
called S-meter, which is part of most receivers. Ab-
solute deviation of his pointer is not very good indica-
tion, because by the S-meter we can only compare the
signal with anothers in current conditions on a band.
Video – the quality of received image you can consider visually in 5 de-
grees, see fig. 8.2. The same method is used in ATV picture qual-
ity reporting system.
R – Readability S – Sig. strength V – Video
1 Unreadable Faint signal, barely perceptible Barely perceptible
2 Barely readable, occasional words Very weak Poor
distinguishable
3 Readable with considerable difficulty Weak Fair
4 Readable with practically no difficulty Fair Good
5 Perfectly readable Fairly good Excellent
6 Good
7 Moderately strong
8 Strong
9 Very strong signals

Table 8.2: Signal reporting with RSV code.

The report can expand information on interference (QRM, QRN) or if the image
is slanted add entry SLANT. The example report is RSV 595 when the reception is
awesome.

8.2 SSTV not only for hams

The SSTV found application in other sectors in the past, mainly due to commercial
production equipment destined for amateurs, and image transfer by phone.

81
8 Ham radio image operations

Barely perceptible Poor Fair

Good Excellent
Figure 8.2: Picture quality reporting system.

Perhaps the most notable applications is the use of SSTV to monitor active Vol-
cano [20]. U.S. Geological Survey installed in September 1987 sensing camera and
FM radio transceiver for observation of volcanic activity on Mount St. Helena.
SSTV was used for remote medical applications during eighties, e. g. transfer of
radiologic pictures over phone lines and over satellite narrow band communication
channels.
Often repeated mistake is that NASA has used the same system as amateurs
for the SSTV transmissions from space in Apollo mission and that the first images
from the Moon were transmitted in same way as amateur slow-scan TV. An NASA
system for video transmission is different, but their engineers also named it slow-
scan TV, but it transmitted images at rate 10 frames per second with 320 lines. The
conversion to a normal television broadcast was made by optical path, the television
camera panned SSTV monitor for broadcast to millions of households.
The Amateur SSTV fly to space later, when SSTV images were sent from space
shuttle in SAREX missions or from orbital stations Mir and ISS. These broadcasts
were received by many hams over the World.

8.3 Diplomas and QSL cards

A tangible confirmation of ham radio contact is a QSL card, although after finished
QSO you may have few saved pictures on your hard drive as a memory of the
contact. But QSL cards still remains as traditional contact confirmation. Also, list

82
Ham radio image operations 8

of confirmed QSOs (those you’ve obtained QSL cards) must be also accompanied
for obtaining a number of ham radio diplomas. So the picture of QSL has not same
value as real QSL card.
In addition to diplomas issued directly only for SSTV contacts, the diplomas like
WAS, WAC, ADXA, CQ DX Award and many more can be obtained with indication
of used communication mode and some additional variants like QRP.
There are few diplomas for exclusive two-way QSO using slow-scan television
listed bellow.

8.3.1 IVCA DX Achievement Award DXAA

The award is issued to amateurs and listeners for two-way SSTV contact/reception
with 50 countries from DXCC list. Additional stickers are for every more 25 coun-
tries.

8.3.2 DANISH DX SSTV AWARD

The award can be obtained by radio amateurs and listeners for confirmed contacts
with 50 different ARRL’s DXCC countries. A QSL list must contain: call, time ,
date, band, mode (SSTV) and country and confimed by two licensed amauteurs and
signed by applicant. All bands legally used in applicant’s country are accepted, but
no contacts via repeaters. It is possible to get additional stickers for contacts with
100 countries and 1 OZ station (silver), 150+2 (gold) and 200 (diamant).
The fee for basic award is $8 or €10.
Award manager: S.K.Mogensen OZ6SM, Rundhøjvej 8, DK 7970 Redsted, Dan-
mark, email: oz6sm@nypost.dk.

http://www.ddxg.dk/awards/sstvaward.htm

8.3.3 Russian SSTV Award

This award is sponsored by the CRC of Krenkelya and the Moscow section SSTV to
popularize the use of SSTV in Russia and other countries of the CIS. Earn 75 points
by contacting Russian and CIS stations on the SSTV mode on or after 1st March
1998. Contacts with members of the Moscow SSTV is valuated by 3 points and one
point for contacts with CIS countries or stations not members of the Moscow SSTV
1.

83
8 Ham radio image operations

Send GCR list and fee of $5, 10 IRC or equivalent to: Verigin Dmitriy Andree-
vich, Lyubertsy, Moscovskaya oblast, Novaya street, 10-48, 140002, Russia, email:
ra3ahq@online.ru

http://olympia.fortunecity.com/bruno/544/award/msstvs/rusaward.html

8.4 Contests
Contest Date
DARC SSTV Contest 3rd weekend in March
Russian SSTV Contest 2nd Saturday in April
NVCG SSTV Contest 2nd week of July
Danish SSTV Contest 1st weekend in May
DARC HF–FAX–Contest 3rd weekend in August
JASTA SSTV Activity from 1st to 31st August
Ukrainian SSTV Contest 1st Saturday in December

Table 8.3: none

8.4.1 DARC SSTV Contest

It takes place the 3rd weekend in March, begins at 12:00 UTC on Saturday and
ends at 12:00 UTC on Sunday. Competing has categories operator and SWL, on
the bands 3.5 to 28 MHz. Transmit RSV and the number of QSOs beginning with
001, each QSO is valued by a point. Multipliers are WAE/DXCC countries, W, VE
and JA districts. Logs should be sent within 4 weeks after the contest at: Werner
Ludwig DF5BX, Post Box 1270, D-49110 Georgsmarienshuette, Germany, email:
df5bx@darc.de.

8.4.2 Russian SSTV Contest

Takes place in 2nd Saturday in April from 00:00 MSK to 24:00 MSK (UTC = MSK – 3).
Bands: 80, 40, 20, 15, 10, 6, 2 m. There are several categories: A. Multi-operators,
all bands; B. One operator, all bands (legal in Russia); C. One operator, all bands;
D. One operator, one band; E. listener. A CQ image should contains CQ RUSTEST.

84
Ham radio image operations 8

Transmit RSV and the number of QSOs beginning with 001, Russian stations trans-
mit RSV, zone and number of QSOs. Scoring: 6 points for each QSO, plus 2 points
for QSO with MsstvS member, additional points for QSO with station from MsstvS
scoring list. Final score: total sum of points plus additional points. Send log sepa-
rately for each band, must include band, date, time in UTC, call sign messages sent
and received. Stations with multiple operators must deliver list of names and sig-
natures for all operators. The log send to 24th April. Organizer: Russian SSTV
Contest Manager Krenkel of CRC Russia, P.O. Box 88, Moscow, Russia.

8.4.3 NVCG SSTV Contest

The contest is organize Nishi Nippon Visual Communication Group and takes up 9
days in July. Score is 2 points for two-way SSTV QSO with NVCG member (they
used “M” letter in report, e. g. 595M), and 1 point for another QSO. Only one QSO
with same station is counted regardless of band. The multiplier is the total number
of different prefixes. Send log to Susumu Tokuyasu JA6GN, 3-1-6 Jyousei Sawaraku,
Fukuoka 814-0003, Japan or email: sstv-contest-nvcg@wak.bbiq.jp.

8.4.4 Danish SSTV Contest

The contest is organized by Danish SSTV Group. Takes place first weekend in May,
from Saturday 00:00 UTC to Sunday 24:00 UTC. Bands: 80, 40, 20, 15, 10, 6, 2 m.
Score: 2 points for every DXCC country, 1 point for contact and bonus 1 point
for contact with Danish station. It is possible to make QSO with same station on
different band. The stations on 1st to 5th place will receive certificates. Mail logs
to: Carl Emkjer, Soborghus Park 8, DK 2860 Soborg, Denmark.

8.4.5 JASTA SSTV Activity

Takes place in August from 1st 00:00 UTC to 31st 24:00 UTC on 3.5MHz band
and all upper bands. There are two categories: “J” – Japanese stations; “S” – all
stations operating outside of Japan. Exchange the usual RSV and number of QSO
starting with 001. Regardless of bands used a station may only be contacted only
once each UTC day. The points for QSOs depend on band: 1 point 3.5–28 MHz,
2 points 50–430 MHz and 3 point for 1200 MHz and upper bands. Multipliers are
districts JA1 to JA0, DXCC countries and working days (max. 10). Prefixes 7K to
7N are all JA1 districts. Contest manager: Yoshikazu Tanabe JA3WZT/1, 905-8,
Shimotaniganuki, IRUMA, SAITAMA, 358 Japan, send logs in TXT format to
ja3wzt@mue.biglobe.ne.jp.

85
8 Ham radio image operations

http://homepage3.nifty.com/jasta/

8.4.6 Ukrainian SSTV Contest

Takes place in first Saturday in December, from 12:00 UTC to Sunday 12:00 UTC.
Bands: 80, 40, 20, 15 and 10 m. There are contest categories: A. One operator;
B. One operator, one band; C. Multiple operators; F. listeners. The CQ picture
should contain text CQ UKR Contest. Exchange report RSV and number from 001.
Ukrainian stations send two-letter region code too. The score is 1 point for QSO
with same country station, 2 points for same continent QSO, 3 points for QSO with
other continent, QSO with Ukrainian station is for 10 points. Multipliers are DXCC
and WAE countries and Ukrainian district for each band. The score is counted
separately for Ukrainian and foreign stations. Logs send in usual form for each
band to organizer: UKR SSTV CONTEST, P. O. Box 10, Kerch, 98300 Ukraine.

8.5 SSTV repeaters

An SSTV repeater is radio station for relaying of SSTV signals. A typical repeater
is equipped with HF or VHF transceiver and computer with sound card. A software
must have an option to work as repeater.
SSTV repeaters are used by amateur radio operators for exchanging pictures. If
two stations can not copy each other, they can still communicate through a repeater.
To activate repeater send the activation tone of frequency 1750 Hz, when the re-
peater is activated, it’s send –.– (K) in morse code. The station must start sending
a picture in approximately 10 seconds. After reception the received image is trans-
mitted on the repeater’s operation frequency.
Repeaters should operate in common SSTV modes, but it depends on the software
used (MMSSTV, JVComm32, MSCAN). Some repeater are not activated by audio
tone, but instead by the SSTV vertical synchronization signal.
Some repeaters works also as beacon and sends periodically random images with
identification and timestamp.

8.5.1 HF and 50 MHz repeater list

8.6 Ham radio satellites and space broadcast

For SSTV operations can be used a linear relay installed on some of the amateur
radio satellites. Amateur satellites orbiting the Earth for elliptical orbits. Linear

86
Ham radio image operations 8

Freq. Call QTH Activation Power Note

3,720 F5ZFJ Haute Saône, JN27UR image linked with repeater on
144,525 MHz
14,236 VK3DNH Rochester Active 24 hours
14,239 VK2ISP Coogee, New South Aktive 24 hours
Wales
21,349 VK6ET Brackenridge, approx. 50–100 W QRV 22:00-08:00 UTC
20 km north from
Brisbane
28,660 GI4GTY Lisburn image
28,688 HB9AC Eighental, Lucerne 1750 Hz Linked with repeater on
JN47CA 144,825 MHz FM.
28,690 K3ASI North Carolina 1750 Hz 45 W Aktive 24 hours, beacon
every 15-20 minut.
28,700 ON4VRB Heist o/d Berg 1750 Hz Linked with repeater on
433,925 MHz.
28,750 ON0DTG Doornik
28,900 EA8EE
50,500 F6IKY Haute Savois (700 m) USB
50,510 OZ6STV Copenhagen, JO65ER 1750 Hz 60 W Beacon every 30
minutes.

Table 8.4: none

relays (transponder) performs retransmission of the wider frequency range, typically

50 to 250 kHz. So the satellite then transmits all the signals (CW, SSB, …) received
on the band (not like the FM ground FM repeater to allow operations to only one
user). If you have station equipped for satellite communication you can try also
SSTV.
You can also monitor experimental SSTV transmission from International Space
Station and receive SSTV signals with your 2m FM transceiver, see section 8.6.3.
Space communications provides few problems. The first of these is Doppler effect,
named after the famous Austrian physicist, which reflects changes in wavelength
of the signal between the observer and the signal source on a moving object. In
practice this means that if the satellite is closer to your position the signal appears
to have a shorter wavelength and the receiver must tune to higher frequencies, when
the satellite is moving away it’s exactly the opposite.

87
8 Ham radio image operations

Other problem is variance of signal quality due to satellite rotation, that causes
a leakage signal. The antenna with circular polarization should be used for these
purposes.

Frequency band Designator

21–30 MHz H
144–146 MHz V
435–438 MHz U
1.26–1.27 GHz L
2.40–2.45 GHz S
5.6 GHz C
10.4, GHz X
24 GHz K
47 GHz R

Table 8.5: Uplink

and downlink bands.

Frequency bands of linear relays are shown in table 8.5. These frequencies describe
satellite relay operating modes. It is fixed by satellite design on chosen by control
center. The designator like U/V describes uplink 435–438 MHz (U) and downlink
144–146 MHz (V). E. g. Fuji-OSCAR 29 (FO-29) operates in mode V/U, the up-
link is in the range of 146,000 to 145,900 kHz CW/LSB and downlink 435.800 to
435.900 kHz CW/USB. Note that relay inverts signal frequency (LSB to USB). Other
satellites carry on board the single-channel FM transmitter, such as the popular
AO-27 with uplink 145,850 kHz FM and downlink 436,795 kHz FM.
It is required to monitor own SSTV signals on downlink, when working on linear
relays. Some sound cards support full duplex operation, so the computer can also
send and receive in same time. Then the operator changes the transmit frequency
so that receiving frequency appears to be same, just follow the position of the syncs
on spectroscope. This way compensates the Doppler effect.
During years there were many satellites used for SSTV operations, like FO-29,
VUSat OSCAR 52 (beacon 145,936 kHz), AO-51, SO-50, etc. But satellite lifetime
is limited, in time board batteries getting weaker and ground control center switching
off transponder and waits for their recharge from solar panels. You can find actual
informations and satellite statuses on website of Amateur Satellite Corporation, i. e.
AMSAT.

88
Ham radio image operations 8

http://www.amsat.org

8.6.1 SSTV from Mir station

Days of the orbital station Mir are already numbered, but as a reminder there is
description of the experiences with SSTV transmission, which took place in Manned
Amateur Radio Experiment (MAREX) in late 1998 and 1999.
The project anticipated broadcast on the frequency 437.975 MHz, but due to
some problems with antenna systems we have to make do with only the occasional
broadcast on the two meters band.
Transmit frequency was 145,985 MHz FM ± Dopplers’s frequency shift. The
packet radio AX.25 BBS R0MIR-1 was normally operating on this channel.
Station at low orbit passed 5 times a day over Europe at approximately 1.5 hours
intervals.
The Robot 36 Color mode was chosen for SSTV transmission. The pictures were
sent in 2 minutes interval, so during one orbit you could copy transmission for 10
minutes and receive about 5 pictures. Each picture was introduced by morse code
–.././/.–./– – – – –/– –/../.–. DE R0MIR and then transmitted.

Figure 8.3: SSTV picture form station R0MIR.

I found, that there was about ±5 kHz frequency change caused by Doppler’s effect.
So during orbit it is good to tune receiver, it’s ideal to use transceiver with continuous

89
8 Ham radio image operations

FM tuning (I use FT-767). When Mir approached the horizont and it was coming
near the tunning frequency was 145,990 MHz and when it was fly away the frequency
is lower, i.ė. 145,980 MHz. Some transceivers measure discrimination of FM signal,
so it is very easy to tune on carrier frequency.
Designers of SSTV station chose Robot 36 Color mode, it’s not resistant to noise,
so when there is great shift from center carrier frequency the signal used to be noisy
and picture quality is distorted. The frequency of AFSK signal transmitted throught
FM channel doesn’t change, so the color distortion, known from SSB transmission,
doesn’t appear.
The antenna of my station for Mir monitoring was 3 element yagi with vertical
polarization (normally used for ground repeaters). I directed it to azimuth, where
Mir was nearest to my site. Later I tried to direct rotator during orbit, the azimuth
and time I had computed and it was possible also to direct yagi by signal strength
displayed on transceiver S-metr.

8.6.2 SuitSat
In early 2006 (originally planned to release about 3 months earlier) were from the
International Space Station (ISS) launched the satellite in an unusual project ARISS
(Amateur Radio on the International Space Station). The satellite was named Suit-
Sat (the code name is the AMSAT-OSCAR-54 [AO-54]). And its name describes
the full implementation of the satellite, because on-board equipment was built into
expired Russian space suit (type Orlan).
The transnsceiver Kenwood TH-K2 was tuned to frequency 145,990 MHz and its
power source was realized from the batteries, so its lifetime was limited to a few
weeks.
The satellite was programmed to broadcast a voice message, prepared SSTV image
(in Robot 36 Color) and telemetry data that contained information such as mea-
sured temperature and radiation. The entire broadcast session lasts approximately
9 minutes.
After few hours after SuitSat’s release, ham radio operators reported only weak
signals and was monitored only couple of days. Probably due to low temperature
the on-board batteries lost capacity.
The SuitSat AO-54 should starts the series of similar experiments, like project
Arissat-1.

8.6.3 Amateur Radio on the International Space Station

The successor of popular ham radio projects on Mir is the project ARISS. The targets
are to build SSTV beacon/repeater, packet radio BBS and world-wide propagation
of ham radio hobby.

90
Ham radio image operations 8

Figure 8.4: SuitSat built into the spacesuit (NASA, source cat.
no. ISS012E15666).

The SSTV equipment on ISS consist of SpaceCam 1 software (from ChromaPix

authors). It runs on normal PC, support video digitizer and it can works as repeater
or transmits slide-show pictures from station cameras. The SpaceCam transmit

91
8 Ham radio image operations

pictures every 120 seconds in Robot Color 36 mode and every picture is started
with morse identification (R0ISS, NA1SS).
The following frequencies are currently used:

⊳ Voice and Packet Downlink: 145.800 MHz (Worldwide)

⊳ Voice Uplink: 144.490 MHz for Regions 2 and 3 (The Americas, and the Pacific)
⊳ Voice Uplink: 145.200 MHz for Region 1 (Europe, Central Asia and Africa)
⊳ Packet Uplink: 145.990 MHz (Worldwide)
⊳ Crossband FM repeater downlink: 145.800 MHz (Worldwide)
⊳ Crossband FM repeater uplink: 437.800 MHz (Worldwide)
⊳ Worldwide SSTV downlink: 145.800 MHz

For latest ARISS news and status check the website:

http://www.ariss.org

92
 9
Introduction to
digital slow-scan TV
The development of computers and new opportunities, which gives us the use of pow-
erful processors and sound cards modems have resulted in the design of new com-
munication modes. One of these modes is the digital slow-scan television (DSSTV),
which allows transmission of images without any loss of quality.
We have two choices for digital image transmission. The first system is RDFT –
Redundant Data File Transfer, which is the result of several years of creative efforts
of Barry Sanderson, KB9VAK and a group of ham radio enthusiasts.
The second system is called Digital Radio Mondiale (DRM). It’s open standard
for digital broadcast on short waves. It was developed by DRM Consorcium and
was standardized by organizations ITU, IEC and ETSI. The DRM system is used
mainly by short-wave broadcast stations and its modification for ham radio purposes
is called HamDRM.
It is possible to use these modes also for transfer of any types of files (text, sounds,
software,…) instead of images.
The difference between analog and digital SSTV is huge. There are used entirely
different modulation principles and essentially some files in JPEG, JPEG200, PNG,
etc. formats are sent. Also the error correction and detection is implemented using
Reed-Solomon code.
The result is, that image is transfered without any distortion (means transfer
distortion, the images should be distorted due to loss compression techniques, but
you can control the degree of it). There is no constrain in image resolution, it is
given by image file parameters. Only constrain is the bandwidth of SSB channel
and the resulting maximum data rate and the time required for transmission.
There are also bigger requirements for stations equipment in comparison with
classic SSTV. You need better computer, at least 1GHz CPU and 256 MB RAM. In
your ham-shack you can use without any problems an oldish computer (e.g. Pentium
150 MHz) for digital modes like RTTY or PSK, packet radio and analog SSTV,
but algorithms of signal processing for DSSTV is so complex that a slow 150MHz
Pentium would not be able to process signals in real-time.

93
9 Introduction to digital slow-scan TV

Also, there are increased requirements on the transceiver used. Used modulation
techniques use the maximal width of the communication channel and SSB transceiver
without the linear range of SSB channel isn’t usable for DSSTV. Of course, there
must be switched off any additional signal and modulation filters (speech processor,
equalizer,…).
Sound card interfacing is same as for classic SSTV, the connection between TRX
and sound card is enough. Output level of sound card should be about 1/3 of max-
imum (switch off software AGC). When signal is overexcited, the intermodulation
causes disproportionately large signal and distortion and the signal cannot be de-
coded. A peak power with 100W transceiver is about 20–25 W.

9.1 Digital communication basics

Before I describe the transmission systems we look at some important concepts
of data communication. What interests us most is the speed which is possible
to transmit information – we distinguish between the speed of transmission and
modulation:

Symbol rate 𝑣𝑚 – express the number of changes 𝑎 of carrier signal per

second. It is measured in unit Baud (Bd) or Symbols
per second (S/s). Symbol rate does not say anything
about how much information transmitted on signal
carrier.
𝑣𝑚 = 1/𝑎 [Bd]
Bit rate 𝑣𝑝 – indicates the amount of information transferred per sec-
ond. It is expressed in bits per second (bps). Bit rate says
nothing about how fast the signal carrier changes.
𝑣𝑝 = 𝑣𝑚 ⋅ log2 𝑚 [bps],

where 𝑚 is the number of modulation states.

We know from previous chapters that an important feature of communication

channel is a limited bandwidth 𝐵. Relation between symbol rate and bandwidth
shows Nyquist rate:
𝑣𝑚 = 2 ⋅ 𝐵.
Ideally, the symbol rate should be twice the bandwidth. Substituting the formula
for the symbol rate, we get:
𝑣𝑝 = 2 ⋅ 𝐵 ⋅ log2 𝑚.

94
Introduction to digital slow-scan TV 9

Let’s look on the relationship between symbol rate and bit rate, because these two
term are often use interchangeably. E.g. packet radio on VHF has bit rate 1200 bps
and the used modulation is AFSK (Audio Frequency-Shift Keying). Frequencies
carrying information are two – 2200 Hz for mark (log. 1) and 1200 Hz for space (log.
0). We know 𝑣𝑝 = 1200 bps, 𝑚 = 2, so symbol rate is equal to bit rate:

𝑣𝑝 1 200
𝑣𝑚 = = Bd = 1 200 Bd.
log2 𝑚 log2 2

A packet radio is based on ITU-T V.23 specification for telephone modems, where
bandwidth is limited to about 4 kHz. Modern dial-up modems, but have a much
higher bit rates, up to 56 kbps and the bandwidth remains 4 kHz. How is that
possible?
It’s possible through the used advanced modulation, which has more modulation
states 𝑚 then two. For example, modems based on V.32 specification can use bit
rate up to 9,600 bps. There is used QAM (Quadrature Amplitude Modulation), which
in case of QAM-16 has 16 states per one modulation symbol. The symbol rate in
this case is:
𝑣𝑝 9 600
𝑣𝑚 = = Bd = 2 400 Bd.
log2 𝑚 log2 16

One could think that it’s possible to reach any speed because of improved modu-
lation and more states. Unfortunately not, because there are stark physical limits.
Maximal channel capacity 𝐶 (bit rate) in bps is given by Shannon’s law, which de-
pends on bandwidth 𝐵 (Hz) and channel parameters signal/noise ratio 𝑆/𝑁 (dB):

𝑆
𝐶 = 𝐵 ⋅ log2 (1 + ).
𝑁

As we can see the maximum bit rate speed is not affected by the used technology,
but the bandwidth 𝐵 and signal/noise ratio (SNR), which cannot be changed. SNR
is given in decibels (dB) and describes the ratio of a signal power to a noise power
of a processed bandwidth.

9.2 Error detection and correction

The error detection codes are used as a check of error-free transmission. The idea
is based on some extra data (redundancy) added to a message. The redundancy is
generated from some input data on a transmission side (FEC – Forward Error Cor-
rection) and on a reception side it is possible to check if the data was transferred

95
9 Introduction to digital slow-scan TV

without error. An used code may have also ability for an error correction, so data af-
fected during transmission can be repaired on reception side without retransmission.
There are several error detection codes, e.g. even parity described in chapter 3.6.2.
The codes have several parameters. First is the bit length of information 𝑘, which
we want to encode and the length of codeword 𝑛. The difference 𝑟 = 𝑛 − 𝑘 is the
length of redundancy data. Redundancy does not transfer any information, but it
is only used for error detection and possibly correction. The ratio of the number of
information symbols to the number of all symbols
𝑘 𝑛−𝑟
𝑅= =
𝑛 𝑛

expresses information ratio. In practice, we require that redundancy is minimized.

The ability of the code, how many errors should be detected or corrected is given
by Hamming distance. It is determined as the number of different symbols of two
different codewords. The most important is minimal Hamming distance 𝑑 of all
arbitrary codewords. E.g. Hamming distance of 0101000 and 0111001 is 𝑑 =
2. The errors during transmission cause replacement of one symbol to another
and Hamming distance indicates how many replacements may occur to change the
codeword to another valid codeword. Is is advantageous to have a Hamming distance
of codewords larger as possible. So if you want code that reveals just one error bit,
the minimum distance must be 𝑑 = 2. Block code with minimal distance 𝑑 detects
all 𝑡-multiple errors for 𝑡 < 𝑑. If there is too much errors that 𝑡 = 𝑑 or 𝑡 > 𝑑
there should be created a new valid word, so the error cannot be detected. The
code can correct errors for larger 𝑑, if for error word is found a valid codeword with
smaller Hamming distance between error and valid word. The block code corrects
all 𝑡-multiple error when
𝑑
𝑡< .
2

These findings can be demonstrated on a simple case of 2-bit code secured with even
parity. Two-bit code can have a total of 4 words of information, and a redundant
bit will be added, so that the number of log. ones in the codeword will be even.
The resulting code words have 3 bits and there are 23 = 8 different bit words
(code words are bold):
The minimum distance 𝑑 of our parity code is equal to 2, so the code is able to
detect just one error. When word 011 is sent and 010 is received we know that there
is an error. If there are two errors and 011 changes to 000, then there is a word that
belongs to a set of codewords and error isn’t detected.
In the following sections are described some commonly used error-detection and
correction codes.

96
Introduction to digital slow-scan TV 9

Information Parity Codeword

word
00 0 000
01 1 011
10 1 101
11 0 110
Table 9.1: none

000 001 010 011

100 101 110 111
Table 9.2: none

9.2.1 Cyclic redundancy check

The CRC is commonly used code. The systematic cyclic code, adds a fixed-length
check (checksum) value to message. The checksum is used for error detection.
CRC calculation is performed on block or section of data is stored in memory, the
𝑘-bit sequence is represented as a polynomial 𝐺(𝑥). This is polynomial is divided
by generating polynomial 𝑃 (𝑥) in arithmetic modulo 2. The result is polynomial
𝑄(𝑥) and the remainder after dividing 𝑅(𝑥). The remainder 𝑅(𝑥) is added to input
data and transmitted in message.
On the reception side the division with 𝑃 (𝑥) is computed again and new remainder
𝑅′(𝑥) is compared with transferred remainder 𝑅(𝑥). If both values are the same
transfer went without error, if not at least one bit was transferred incorrectly.

9.2.2 Hamming code

In the area of data communications (e.g. TV teletext) is sometimes used Hamming
code, which can detect up to two errors and in the case of a one error it is able to
determine at what point of codeword error occurred and it can fix received bits.
Basically, it uses for its purposes even parity. While the parity bits are in the final
codeword positioned at the serial number is equal to the square of 2 (1., 2., 4., 8.,…).
Under the control bit position is then selected certain sequence of information words,
which is used to determine the value of control bit.

97
9 Introduction to digital slow-scan TV

9.2.3 Reed-Solomon code

Hamming code works well in environments where errors occur randomly and their
incidence is low (e.g. in computer memory, which can detect a erroneous bit to
100 million). But even if that failure causes a greater number of adjacent bits
are corrupted (burst error), the Hamming code is useless. In the field of radio
transmission, where the signal is often affected by atmospheric disturbances, fade
outs and interference, then errors occur in clusters. This means that close to the
incorrect symbol are other symbols incorrect too. For burst error correction is
applied Reed-Solomon code (RS).
RS codes are the most widely used codes for detection and error correction. They
are characterized by having the largest possible minimum distance, and compared
to the previous code will not correct the individual bits, but all symbols. RS code
have found application in the number of areas – is used by NASA for space com-
munications, protects data on CD-ROMs and DVDs and is also used for terrestrial
transmission of HDTV or in the data modems for the cable television networks.
Like the CRC the RS code is systematic. For its generation are used the algebraic
calculations of Galois field.
Parameters of the RS(𝑛, 𝑘) are defined as follows:

⊳ 𝑠 – number of bits in one informational symbol,

⊳ 𝑘 – number of s-bit symbols in data block,
⊳ 𝑛 – number of bits in codeword.

RS(𝑛, 𝑘) is able to correct 𝑛−𝑘

2 errors in 𝑘 information symbols. Often used code
is RS(255, 223), it uses 223 8-bit symbols for creation of 255 symbols of codeword.
There is 32 symbols dedicated for error correction. RS(255, 223) is able to repair up
to 16 erroneous 8-bit symbols.

9.3 Data compression

The image with resolution 320×240 with a color depth of 16 million colors (2563 )
takes 230,400 Bytes (320×240×3) without compression. This file would be trans-
mitted fortyone minutes by RDFT with speed 92 Bytes per second! This time is
really scary in comparison with analog SSTV. It is really necessary to reduce the
file size and reduce the time required for transmission.
The data compression is widely used in such cases, where the data capacity of
communication channels or storage media and memory is limited.
The compression is the process where the physical data block size is reduces to
a certain level. Input data is compressed using the compression algorithm and

98
Introduction to digital slow-scan TV 9

then stored on media or transmitted via communication channel. The data are
decompressed in its original form, when a media is read or a signal received.
One of the important parameters of compression algorithms is lossy. While the
programs or text must by stored in perfect form, but in case of sound, images or
animations we can settle with the omission of certain details, then we’re talking
about lossy compression method.

9.3.1 Information entropy

When Clause E. Shannon was engaged in applied mathematic of communication
theory during 1940s, he started with definition of informational value of message
content. The message which is repeated often is less informative than the message,
which occurs sporadically. So, the often repeated message is more likely than the
unique. The probability in mathematic is expressed by real numbers in range from 0
(for a completely unlikely events) to 1 (for the phenomena that occur surely). Shan-
non defined the amount of information 𝐼(𝑥𝑖 ) for the message 𝑥𝑖 with the probability
of occurrence 𝑝(𝑥𝑖 ) as follows:
𝐼(𝑥𝑖 ) = − log2 𝑝(𝑥𝑖 ).

The graph of negative logarithm see on fig. 9.1 – if the message content is less likely
that its information value is higher.

I(xi )

I(xi ) = − log p(xi )

0 1 p(xi )
Figure 9.1: The relation between information con-
tent 𝐼(𝑥𝑖 ) and its probability 𝑝(𝑥𝑖 ).

Information entropy 𝐻 is defined as average rate of information value 𝐼(𝑥𝑖 ):

𝑁 𝑁
𝐻 = ∑ 𝑝(𝑥𝑖 )𝐼(𝑥𝑖 ) = − ∑ 𝑝(𝑥𝑖 ) log2 𝑝(𝑥𝑖 ) [bit]
𝑖=1 𝑖=1

99
9 Introduction to digital slow-scan TV

We show the entropy meaning in example. We need to transfer messages 𝑎1 , 𝑎2 , …𝑎8

and probability of their occurrence is same: 𝑝𝑖 = 1/8 = 0,125. The entropy of
source is
8
1 1
𝐻 = − ∑ 𝑝(𝑎𝑖 ) log2 𝑝(𝑥𝑖 ) = − (8 ⋅ log2 ) = 3 bits.
𝑖=1
8 8

The observed entropy determines how the message content can be encoded for data
transmission. The length of message in bits is greater then or equal to the entropy,
without loss of information. So the message can be encoded as word of 3bit length:
000, 001, 010,…Maximum entropy is reached when the probability of occurrence of
each message is the same.
But the messages have often different probabilities in many cases. In this example
we need to transfer messages 𝑎1 , 𝑎2 , …𝑎7 . Their probabilities are 𝑝(𝑎1 ) = 0.235,
𝑝(𝑎2 ) = 0.206, 𝑝(𝑎3 ) = 0.176, 𝑝(𝑎4 ) = 0.147, 𝑝(𝑎5 ) = 0.118, 𝑝(𝑎6 ) = 0.059,
𝑝(𝑎7 ) = 0.029, 𝑝(𝑎8 ) = 0.029. Entropy of source is
7
𝐻& = − ∑ 𝑝(𝑎𝑖 ) log2 𝑝(𝑎𝑖 ) =
𝑖=1
 & = −(0.235 ⋅ (−2.09) + 0.206 ⋅ (−2.28) + 0.176 ⋅ (−2.50) + 0.147 ⋅ (−2.76)+
&0.118 ⋅ (−3.08) + 0.059 ⋅ (−4.08) + 0.029 ⋅ (−5.08) + 0.029 ⋅ (−5.08)) bits
𝐻& ≈ −2.712 bits

We see, that the entropy of source is lower and because data bits are not divisible,
it is necessary to encode the message again to the words of length 3. But suspect
that such an encoding is no longer optimal. There is the idea to encode frequently
occurring words as the message of the shorter length. This idea was well-counseled
by David A. Huffman, the Shannon’s student.

9.3.2 Huffman coding

We can show an example of Huffman coding construction. The message we are going
to encode if following:
THE SHELLS SHE SELLS ARE SEASHELLS
This message contains 8 symbols (S, E, L, “ ”, H, A, T, R). The message can be
expressed with code words of 3bit length. Its whole length is 3 ⋅ 34 = 102 bits.
For Huffman coding we need to determine number of each symbol and their prob-
abilities.

100
Introduction to digital slow-scan TV 9

S E L H A T R
8× 7× 6× 5× 4× 2× 1× 1×
0.235 0.206 0.176 0.147 0.118 0.059 0.029 0.029
Table 9.3: none

There is used binary tree, it is a data structure often used in programming. The
symbols are sorted by their frequency and then each symbol represents a tree leaf,
and its weight is given by symbol occurrence. In first step join two leafs with the
lowest weight, in our case T and R and create a node. The node weight is sum of
weights T and R. In the next step join leafs or nodes with the lowest weight and
proceed as long as there is only one node (the root of binary tree).
Now, go from the root toward leafs by the edges and each edge label by 0 or
1, if the edge goes up or down (in tree terminology to left or right subtree). The
constructed tree with labeled edges see on fig 9.2. To find the code of each symbol
pass all ways from the root towards the leaves. The path going along the edges of
0, 0 ends in S, the path going along 1, 1, 1, 0 ends in A.

S 8 0
Symbol Code 1 15
E 7 0
S 00
E 10 L 6 0
11
34
1
L 010
0 1
5
19
011 H 4 0 1
H 110 A 2 0 8
1
A 1110 0 4
T 1 1
T 11110 1 2
R 1
R 11111
Figure 9.2: The results of Huffman encoding.

We see, that more frequent symbols with high probability of occurrence have shorter
code than sporadic symbols. Our message after encoding:

1111011001101001100110010000101
0011001101000110010000101111011
1110110100011110001100110010000

The message length was reduced from 102 to 93 bits. For decoding the binary tree
on 9.2 can be used again. We will start in the root and go along edges 1, 1, 1, 1,

101
9 Introduction to digital slow-scan TV

0 until we arrive to lead, here symbol T, then we return to the root and go along
1, 1, 0 and we arrive to leaf H. By this way we continue until the whole message
is decoded. Because Huffman coding is has unique prefixes for each code, and this
prefixes is not start of another codeword the decoding can not do mistake.
Other compression algorithms using dictionary methods. These methods are
based on fact that some words in the input file occur more frequently. Repeat-
ing words are stored in the dictionary. These words are replaced with their cor-
responding code words in output file. Among the representatives of this type of
compression belongs LZW (Lempel-Ziv-Welch) as used in the ZIP compression or
GIF or a variant of the TIFF formats.

9.3.3 Lossless data compression

Many applications needs for their requirements that data aren’t impaired if they
are compressed. E.g. for binary programs and data. Lossless compression has
its justification in the field of computer graphics and image storage too. Lossy
compression fits on “nature images” and photographs, but when it is used on a
computer-generated graphics such as diagrams and charts, the image distortion is
more noticeable on sharp edges and color gradients, even at low compress ratio (see
section 9.3.4.3.).
Many compression algorithms were developed for lossless compression. A sim-
ple algorithm is for example Run Length Encoding (RLE). This algorithm stores
repeated bytes as their value and number. E.g. AB AB AB CD EF EF EF EF EF is
stored as 03 AB 01 CD 05 EF, so instead of 9 bytes should be only 6 stored.
Other types of algorithms are based on statistical methods. Before or during
the compression process the algorithm determines the relative representations of
elements of the file, and those repeated frequently are expressed as a short code
word. Such algorithm is the Huffman coding described above. Also, Morse code
is one of those codes, frequently recurring characters such as E (.) A (.–), I (..)
have assigned shorter codes and the less frequent, such as H (....), J (.– – –), F (..–.)
longer codes.

9.3.3.1 Portable Network Graphics

The PNG is appropriate graphics format with lossless compression. PNG was cre-
ated to replace the outdated GIF format. PNG is not limited to a palette of 256
colors like GIF and allows to set a continuous level of transparency (alpha-channel)
compared to GIF, which has the option to choose only two levels (yes or no trans-
parency). If you want to save the lossless image just choose PNG.
The algorithm used in PNG is called deflate. This method is enhanced in some
ways, the image lines are firstly processed by filter, which tries to find a similar

102
Introduction to digital slow-scan TV 9

neighborhood for each pixel. After processing there is a large number of data with
zero value or a value close to zero (for same or similar values), so compression
algorithms finds in data areas with same value so it can shrinks the length of the
resulting file.

9.3.4 Lossy compression

The principle of lossy compression takes advantage of the processing equipment, in
the case of the human eye it is unable to process certain information, so it actually
would be an extra piece of information omitted.
A widely used method for lossy image compression format is JPEG (Joint Pho-
tographic Experts Group). The JPEG is the standard established by the ISO and
ITU, released in 1992 (later upgraded in 1997). A successor is upgraded format
JPEG2000. It was developed by JPEG committee since 1995, was released in De-
cember 2000 and further revised in 2003, but it is not so widespread as its predeces-
sor.

9.3.4.1 JPEG compression

JPEG usually does not use RGB color coding but use YCrCb, see chapter 3.5.2,
because th human eye perceives brightness and colors with different sensitivity. The
storage of YCrCb colors, mostly in the ration 4 : 2 : 0 reduces size of file, but itself is
not enough. The image is further transformed, see schema in fig. 9.3.
In first step the image is divided on square block of 8 × 8 pixels and these 64
points is transformed from spatial domain (𝑥, 𝑦) to frequency (𝑖, 𝑗) by discrete
cosine transform. Just for completeness, as follows:
7 7
1 (2x + 1)i𝜋 (2y + 1)j𝜋
DCT(i, j) = & C(i) ⋅ C(j) ∑ ∑ pixel(x, y) ⋅ cos [ ] ⋅ cos [ ],
4 x=0 y=0
16 16
⎧ √1 & for a = 0;
{
&where C(a) = ⎨ 2
{ 1 & in other cases.
⎩
The first position 𝑖 = 0, 𝑗 = 0 holds DC coefficient, the mean value of the waveform
for 64 values of block. The other positions contains AC coefficients and their value
is derived from deviations between each values and DC coefficient. Basically, the
DCT trying the block of 8×8 “to fit” a linear combination of shapes given byt the
previous formula.

103
9 Introduction to digital slow-scan TV

Original image
Macroblock
Dekompozice Y Y Cr

Y = 0,30R + 0,59G + 0,11B Y Cb

Cr = – 0,17R – 0,33G + 0,5B
Y
Cb = + 0,5R – 0,42G – 0,08B

DC AC01 AC07 every pixel – 128

705 109 –41 8 –1 7 14 1 107 103 98 84 45 8 4 1

–94 118 –23 1 3 9 6 9 AC37 109 93 114 97 90 66 11 –3

–70 52 –17 –11 10 1 –4 3 118 90 110 96 93 89 72 33

DCT
–12 16 24 –29 7 –16 –5 –10 106 97 99 100 96 97 92 79

–18 –11 13 –11 –1 4 –11 –7 98 108 106 97 93 106 94 94

3 –2 23 –12 2 6 –14 7 103 103 98 105 98 106 87 109
–5 –8 10 3 4 9 –1 –5 AC77 99 107 109 98 103 93 98 97
2 5 7 –3 –12 4 –4 8 84 80 89 103 100 105 92 81
AC70
Quantization table The coefficients of quantization
16 11 10 16 24 40 51 61
table, here for 50 % quality, they can
Quantization
12 12 14 19 26 58 60 55
be multiplied by constant α, which
14 13 16 24 40 57 69 56
gives a lossy ratio.
705 / 16 = 44, 14 17 22 29 51 87 80 62
109 / 11 = 10, … 18 22 37 56 68 109 103 77

24 35 55 64 81 104 113 92
Quality Q:
49 64 78 87 103 121 120 101 1 ≤ Q ≤ 50, α = 50 / Q,
44 10 –4 1 0 0 0 0 72 92 95 98 112 100 103 99 51 ≤ Q ≤ 100, α = 2 – Q / 50

–8 10 –2 0 0 0 0 0

–5 4 –1 0 0 0 0 0 Linearization 44, 10, –8, –5, 10, –4, 1, –2,

4, –1, –1, 1, –1, 0, 0, 0, 0, 0,
–1 1 1 –1 0 0 0 0 1, 0, 0, 0, 0, 0, –1, 0, 0, …
44 10 –4 1 0 0 0 0
–1 0 0 0 0 0 0 0 –8 10 –2 0 0 0 0 0

–5 4 –1 0 0 0 0 0
0 0 0 0 0 0 0 0 –1 1 1 –1 0 0 0 0 Run-length encoding
–1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
( zeros / categ. / value )
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Image block is linear combination of

patterns given by DCT.
44; (0/4/10), (0/4/–8), (0/3/–5),
(0/4/10), (0/3/–4), (0/1/1), (0/2/–2),
Huffman coding (0/3/4), (0/1/–1), (0/1/–1), (0/1/1),
(0/1/–1), (5/1/1), (5/1/–1), EOB
(only AC coefficients)
101111001011011110001010111100100011001010110010000000000100011110101111101001010
bit stream

Figure 9.3: The JPEG compression for one 8×8 block of brightness.

104
Introduction to digital slow-scan TV 9

Then follows a step that most affects the resulting image and a perception of the
lossy compression level. The quantization is carried out by individual members of
a predefined luminance quantization table (chrominance component has a different
predefined table). A member of the block at position 00 is divided by a member 00
of the quantization table and the position of the integer part of number is stored at
position 00, continues 01/01, 02/02,…up to each value is divided by its corresponding
coefficient. The result of this process is a square matrix, where most information is
stored in the upper left corner and around the lower right corner are just zeros.
This matrix is linearized into a sequence. Thanks to “zig-zag” reading the non-
zero values appear in front of the sequence and remain part is filled by unnecessary
zeroes.
Then the sequence is divided into categories, the first is DC coefficient and then
other values continue and for each is determined following values: (number of pre-
ceding zeros / category / intrinsic value). The redundant zeros are reduces by RLE
coding and from some place are presented only zeros. The all zeros are omitted and
replaced by EOB (end of block) mark. DC coefficient, brightness and chrominance
values have their codings.
For AC coefficients are zeroes labeled as category 0, for other integer values their
categories is given by bit length of value. For most common AC coefficients {−1, +1}
it is 1, these two values can be represented by value 0 or 1, for {−3, −2, +2, +3}
is length 2, and it is represented by {00, 01, 10, 11}, for {−7, …, −4, +4, …, +7} is
length 3, etc. The result code depends on number of preceding zeros and bit length,
so 0/1 (no zero/ length 1) has 00, 0/2 01, 0/3 100, 1/1 (1 zero/length 1) 1100, 5/1
1111010, etc. The results of Huffman coding for one block se in fig. 9.3.
We have an option to choose the image quality for JPEG files. For quality of 75 %
the distortion is not noticeable in most cases and compress ratio can be around 20 : 1
to 25 : 1. The results of different quality for image with 256 × 192 seen in fig. 9.4.
You can notice little distortion for quality about 50 %, mainly in areas with sharp
color gradients.
Lossy compression of JPEG is not suitable for all types of images. It is good
for “natural” images, but it is problematic for computer generated graphics like
schematic diagrams, 3D renders, etc., where there are many sharp color gradients.
The example of bad chose of compression see in fig. 9.6. The file size of both images
is almost same. While for lossless PNG we cannot see any distortion, in the right
image stored in JPEG format with compression set to the closest file size to PNG,
we see that a DCT transform cannot handle sharp edges and the bias around them
makes image heavily distorted.
There is also an option for storage data in progressive mode. In progressive mode,
in first step the DC coefficients of all image blocks are transferred, then first AC co-
efficients, second AC coef., etc. This allows a low detail preview after receiving only

105
9 Introduction to digital slow-scan TV

Quality 75% Quality 40%

15,970 B 9,564 B

Quality 10% Quality 3%

4,277 B 2,474 B
Figure 9.4: The file size as result
of JPEG compression loss degree.

Figure 9.5: The detail of image saved in 10% quality.

a portion of the data and during a reception more and more details are displayed.
The progressive mode is very useful for slow DSSTV transfer.

106
Introduction to digital slow-scan TV 9

PNG, 1,016 Bytes JPEG 1 % quality, 1,378 Bytes

Figure 9.6: The detail of image with unproper compression.

9.3.4.2 JPEG2000
When compared with original JPEG standard the new JPEG2000 has many im-
provements. There are used more sophisticate mathematical methods. The DCT
is not used, but discrete wavelet transformation (DWT). Wavelet transformation is
one of methods for frequency domain representation of signal and has some advan-
tages over DCT. A functions with defined wave shape are used instead of sinuses
and cosinuses.
Thanks to new transform method the compress ratio is better about 20 to 30 %.
The images with sharp edges and color gradients have lower distortion.
Users of new format appreciate the most a better compression ratio and higher
image quality when using the lossy compression. DCT in JPEG format requires the
division of the image into small 8 × 8 blocks, while JPEG2000 uses the whole image.
The RGB color coding is used. And in addition the user has the choice to mark
“area of interest”. These areas are part of the image, where is required to set lower

107
9 Introduction to digital slow-scan TV

or higher compression ratio. For use in DSSTV is advantageous fault tolerance of

the data stream. Only a small portion of the image displays poorly in the case of
faulty transmission, other sections carried well are not affected. For older JPEG,
the image part following the fault data of stream used to be completely discarded.

Ratio 10 : 1 Ratio 20 : 1
14,628 B 7,310 B

Ratio 29 : 1 Ratio 50 : 1
4,779 B 2,909 B
Figure 9.7: File size depends on
the compression ratio of JPEG 2000.

The new JPEG2000 has also progressive mode like an old JPEG. So the received
image can be viewed during reception. You can see phases of reception in fig. 9.8.

9.3.4.3 Lossy versus lossless image compression — conclusion

In JPEG section is described that lossy compression is not suitable for all types
of images. Charts, diagrams and other images featuring sharp color gradients get
significant loss, see fig. 9.6. Despite the significant quality loss the file size is not
considerable reduced. inTable[tab:comparison] contains a comparison of file size for
various formats. As the input file was used “smiling face” from fig. 9.6 stored at a
resolution of 256 × 192 in 16 colors.

108
Introduction to digital slow-scan TV 9

3% 10 % 26 %

42 % 54 % 67 %

80 % 92 % 100 %
Figure 9.8: Progressive display of JPEG2000 image depends on amount
of transfered data, original image has 400 × 298 resolution.

Even relatively dumb RLE algorithm for lossless compression, but maintaining a
100% quality beats JPEG. It is the user’s choice how to deal with the right choice of
format and select a suitable compromise between resolution, number of colors and
image quality.

109
9 Introduction to digital slow-scan TV

Format Quality File size

Windows Bitmap 100 % 24,654 B
JPEG 100 % 17,740 B
JPEG 75 % 7,300 B
JPEG 50 % 5,298 B
TIFF, PackBits compression 100 % 4,352 B
Windows Bitmap RLE 100 % 3,984 B
TIFF, komprese LZW 100 % 3,850 B
JPEG 25 % 3,766 B
GIF 100 % 1,569 B
JPEG 1% 1,378 B
Portable Network Graphics 100 % 1,111 B
Table 9.4: Comparison of file sizes for different graphic
formats.

110
 10
DSSTV transmission
systems
10.1 Redundant Data File Transfer
This communication mode uses phase shift keying (PSK) modulation. The simplest
PSK modulation changes subcarrier between two phase states (BPSK – biphase shift
keying) and these states corresponds to level of logical zero or one. This is used for
example for teletype mode PSK31.

Data 0 1 1 0 0 1 0

PSK

Figure 10.1: Two-phased digital PSK mod-

ulation.

RDFT uses similar principle but much more extended. The signal is composed from
eight subcarriers from 590 Hz to 2200 Hz with 230Hz steps. Each subcarrier uses
nine modulation states – eight informational states and one state with no change
meaning. The data from inner encoder are used for phase assignment.
The first step in modulation process is to take a cosine of modulation angle plus
1400Hz subcarrier angle. In next steps the energy around 1400 Hz is isolated and
translated onto right subcarrier. Then the subcarriers are compiled together and
resulting spectrum of signal is in fig. 10.3.
The signal contains two levels of error-coding. The outer coding scheme use RS
code (306, 𝑥), where 𝑥 is set by level of error control, see tab. 10.1. The symbol
numbers, produced by the outer encoder, are the input to the inner encoder. The
inner coding scheme uses RS((8, 4), where is 50% redundancy. All 8 symbols are
used for phase settings for each of 8 subcarriers, so the inner code-block is transferred
paralely. The decoder of inner code on reception side is able to correct whole block
if 6 of 8 symbols are transferred without error.

111
10 DSSTV transmission systems

+1
modulation –1
angle +2

 –2
+3
1400 subcarrier
Hz angle –3
+4

–4

low-pass
 filter


cos cos

1400 subcarrier
Hz frequency

590, 820, 1050,
sin 1280, 1510, 1740, sin
1970, 2200 Hz

low-pass
 filter 

Figure 10.2: Block diagram of RDFT modulator for one subcarrier.

Operator can choose one of four modes, in all cases the modulation speed is same
122,5 Bd, but the level of error control differs.
Table 10.1 are parameters of RDFT modes. You can choose lower error con-
trol level when band conditions are good or higher level in case of bad conditions
and big interference. The redundant data consume 70 % of all transferred data for
Wyman 14, so there is possible to apply an extensive error-correction. The Wyman
13 is recommended for long-distance contacts and Wyman 12 for intracontinental
QSOs.
Transferred data block consists of three parts:

⊳ The first is LEADER, it uses always same modulation scheme and error-coding.
It contains RDFT mode identification and it is used for detection of two para-
meter. The first parameter is a tuning deviation in Hz, because most SSB trans-
ceivers have smallest tunning step 10 Hz it is not possible to tune accurately.
The next parameter is clock rate difference, it is caused by inexact sample rates

112
 DSSTV transmission systems 10

0
−5
−10
−15
Amplitude [dB]

−20
−25
−30
−35
−40
−45
−50
250 590 820 1050 1280 1510 1740 1970 2200 2500 2750 3000
Frequency f [Hz]

Figure 10.3: RDFT frequency spectrum

RDFT Subcarrier RS code Redun- Maximum Band- Bit rate Symbol

mode number dancy errors width rate
Wyman 11 8 (306, 274) 10 % 16 1840 Hz 866 bps 122,5 Bd
Wyman 12 8 (306, 242) 20 % 32 1840 Hz 765 bps 122,5 Bd
Wyman 13 8 (306, 178) 40 % 64 1840 Hz 563 bps 122,5 Bd
Wyman 14 8 (306, 92) 70 % 107 1840 Hz 291 bps 122,5 Bd

Table 10.1: RDFT communication modes.

of sound cards and there is also small difference on receiver and transmitter side.
Both these parameters are dynamically found during transmission and they are
used in demodulation process.
⊳ The next part is CODEBLOCK, it is sequence of data frames of transferred file
with redundancy symbols for error correction.
⊳ The transfer ends with TRAILER, it contains mode identification like the first
part.

The average bit rate of transfer is about 736 bps (92 bytes per second).
The input of demodulator are samples of RDFT signal, the output are some phase
states of each subcarrier. The block circuit of demodulator is in fig. 10.2. The delay
block provides a delay of one symbol period. Subtracting the angle values separated
in time by one symbol period is the “differential” portion of this “differential phase
demodulator”. The average block averages 24 adjacent differences and divides by
the unit phase step, to produce the final demodulator output. This averaging helps
reduce the intersymbol interference produced by the low pass filters.

113
 10 DSSTV transmission systems

formater tuning
binary
and outer receiver parameter demodulator
data file
encoder extractor

RS(306, x)

inner inner outer binary

encoder & transmitter decoder decoder data file
modulator
RS(8,4)
Figure 10.4: RDFT communication channel.

low-pass
 filter

cos

subcarrier
frequency arctan

sin

low-pass
 filter


delay average
–

Figure 10.5: The RDFT demodulator block diagram.

10.1.1 RDFT operations

The audio level of sound card is the most important setting on TX side. It must be
inside linear range, it is usually the middle half of its dynamic range. On RX side
should be sound card level adjusted too. The input signal should not be overdriven,
it causes unwanted nonlinearities.
The stations should be very precisely tuned to each other for making successful
contact. The station before the data transfer should send tunning signal. This

114
DSSTV transmission systems 10

signal consist of two tones 1180 and 1520 Hz. These frequencies are labeled on
spectroscope, so you can fine tune and align the frequency peeks with the marks.

Tunning signal Spectroscope with

and station id. data transmission.
Figure 10.6: The tuning
spectrospoce in DIGTRX.

Following operations must be done for transmission and reception of images:

1. Original data file is processes, in case of images the resolution and compression
level are set. Then is generated WAV audio file, which contains a audio signal
for radio transmission. The time spent to signal generation is derived from input
file size and computer configuration. It can took form a second or two (2GHz
and faster CPUs) up to several minutes on slower systems (400 MHz).
2. The audio file is played and transmitted. The reception station records sound
and store it on a hard drive.
3. The software process recorded WAV file and reconstructs original file. This
step is also computationally very demanding and the time needed for decoding
depends not only on the volume of data and processor speed, but also on how
much is necessary to use the error-correction algorithm. This step may take
several minutes on a slow computer, on a 2GHz machine it takes 15 seconds.

Barry Sanderson, KB4VAK, developed programs for RDFT encoding. These pro-
gram are command line driven and it is available as open source under GNU GPL
license. So programmers can implement it to several computer platforms. So thanks
to open source idea, there is few programs where is the RDFT mode available.

115
10 DSSTV transmission systems

RDFT presentation in Dayton conference: http://replay.web.archive.org

/20080528090630/http://www.svs.net/wyman/examples/hdsstv/

10.2 HamDRM system

Communication system HamDRM is derived from open standard Digital Radio Mon-
diale [12], which was created for digitalization of radio broadcast on medium-wave
and short-wave bands. Normal DRM use bandwidth 4.5 kHz to 20 kHz for sound
quality similar to FM broadcast on VHF. The hamradio version HamDRM was cre-
ated by Francesco Lanza, HB9TLK. It is modified for usage in SSB channel with
2.5 kHz bandwidth. HamDRM can be used for image and data file transfer and also
for voice communication, so it should be competitor for analog SSB in future.
The used modulation is COFDM (Coded Orthogonal Frequency Division Multi-
plex), which has maximal utilization of communication channel. The Reed-Solomon
code is used for error correction.
−10

−20

−30
Amplitude [dB]

−40

−50

−60

−70

−80
0 500 1000 1500 2000 2500 3000
Frequency f [Hz]

Figure 10.7: The frequency spectrum of HamDRM system.

The OFDM signal consists of a huge number of subcarriers in baseband. There are
from 29 to 57 subcarriers in case of HamDRM. An each subcarrier is modulated
independently with quadrature amplitude modulation (QAM) and together with
error-correction code creates COFDM. This modulation is well resistant to phase
distortion, attenuation, selective fading and pulse interference. The used modulation
techniques are described later in section 10.2.2.

116
DSSTV transmission systems 10

Figure 10.8: The tuning spectroscope

in DIGTRX. Note three frequency peeks,
that is used as guidelines for proper tun-
ing.

Unlike RDFT, which needs 3 partial operations – coding, recording of broadcast

and decoding, the HamDRM doesn’t transmit data in whole block, but the file is
divided into separate segments, so the image can be decoded and displayed during
transmission.
HamDRM can be used in three basic modes. The Mode A allows the fastest
transmission, but does not protect against the negative effects caused by selective
fading. The Mode B is slower than the first mode, but is resistant against the
negative impacts and it’s much more robust. The last mode is Mode E, which is
designed for communication through a channel with large delay and Doppler effect.
The QAM modulation is used with 4, 16 or 64 states. Modulation QAM-64 is the
fastest but it needs a very good level of signal-noise ratio, at least 18 dB. Modulation
QAM-4 is slower, but is more resistant to interference and requires a lower signal-
noise ratio, about 5–6 dB. Minimum SNR for QAM-16 should be about 8–10 dB.
The selection of modulation depends on an user and an actual conditions prevailing
on the band. Other HamDRM features that can be set by user are following:

⊳ Interleaving is used for change of symbol sequence, it is a way to arrange data in a

non-contiguous way to increase performance. The long interleave has 2 seconds,
it supports better error-correction but causes longer delay during decoding. The
short interleave take 400 ms.
⊳ Bandwidth can be changed to 2.3 kHz or 2.5 kHz. A narrower SSB filter can be
used for lower bandwidth, but transfer speed is little lower.
⊳ Amount of instances is value that gives number of file repetitions during trans-
mission. If there is more than one instance then all segments will be repeated

117
10 DSSTV transmission systems

and the error parts can be corrected automatically on reception side during sec-
ond or third instance. The number of instances makes transmission time longer.
⊳ Leadin is broadcast at beginning of transmission. This initialization is used to
receiver synchronization, extra time allows better synchronization and automatic
set up of reception settings.

Details of the mode and its parameters, along with the call sign is broadcast
throughout the transmission with QAM-4 modulation, so it is possible to tune to
signal during transmission, but the complete data will be received if at least one
complete instance of the transferred file is received.

Mode Bandwidth Number of Level of Transmission speed [bps]

subcarriers MSC FEC QAM-4 QAM-16 QAM-64
A 2,3 kHz 53 normal 1480 2710 4170
low 1900 3470 5340
2,5 kHz 57 normal 1760 3220 4960
low 2260 4130 6360
B 2,3 kHz 45 normal 1070 1950 3000
low 1370 2500 3850
2,5 kHz 51 normal 1270 2320 3570
low 1630 2970 4580
E 2,3 kHz 29 normal 690 1270 1950
low 890 1620 2500
2,5 kHz 31 normal 820 1510 2320
low 1060 1930 2970
Table 10.2: The parameters of HamDRM
modes and their transmission speed.

Parameter selection of HamDRM modes affects the transmission performance, and

hence the transmission speed, which depends on the settings, see table 10.2. The
two corresponding stations should not communicate with each other in the same
mode. E.g. station X has a considerable local interference, so station Y sends in a
more resistant mode, but Y hasn’t this problem, so X can easily transmit in a faster
but less resistant mode.
If the transfer of some segments fails completely, it is not lost at all, because your
QSO partner can send bad segment report (BSR) and you can resend bad segments
again. It’s important to send BSR in same mode. The repeated segments can be
received third station too and if have not all segments received it can complete

118
DSSTV transmission systems 10

whole data. When band conditions are really bad and part of resend fails again, it
is possible to generate the new BSR, so the amount of transfered data will be lower
in next resent.
The DRM transfer consist of three channels – MSC, SDC and FAC. Each channel
is dedicated for transmission of certain data or service information and also for each
is used different coding and modulation scheme.

MSC – Main Service Channel contains data for all services of DRM mul-
tiplex. The multiplex can contain one to four services and each can
transfer data or service information.

FAC – Fast Access Channel, is support channel. It uses QAM-4 and

broadcast callsign, DRM mode identification (band spectrum occu-
pancy, interleaving, mode of MSC and SDC modulation,…).
FAC channel with service information transfer packet with 40 bit
size:

⊳ 2 bit FRAME-ID, identifies a frame in a superframe, value 0, 1,

2
⊳ 1 bit Spectrum Occupancy (2.3 / 2.5 kHz)
⊳ 1 bit Interleaver Depth (400 ms / 2 s)
⊳ 1 bit MSC Mode (QAM-16 / QAM-64)
⊳ 1 bit Protection Level (amount of FEC used)
⊳ 1 bit Audio/Data
⊳ if audio is used, then follows:
⊳ 2 bit, audio codec: LPC, unused, SPEEX;
⊳ 1 bit, text flag;
⊳ if data is used, then follows:
⊳ 2 bity, Packet ID;
⊳ 1 bit, extended MSC mode (QAM-4);
⊳ 21 bits, Label, consisting of 3 × 7 bit ASCII characters (9 char-
acters in superframe)
⊳ 1 bit, dummy
⊳ 8 bitů, CRC, used polynomial 𝐺(𝑥) = 𝑥8 + 𝑥4 + 𝑥3 + 𝑥2 + 1.

SDC – Service Description Channel contains information of MSC decod-

ing scheme and broadcast service attributes during multiplexing.

10.2.1 Comparison of HamDRM and RDFT

There are several software products for RDFT and HamDRM, but preference of
users inclines to HamDRM. Main reason for HamDRM popularity over RDFT are:

119
10 DSSTV transmission systems

1. subcarrier f 51. subcarrier

signal bandwidth

*....T0f.TT.0T..TT0.X..f0.TT.T0fTT..0T.TT.0T....*..
..*....f0.X...0.....0.Xf..0....f0.X...0.....0.....*
....0..f..0.X...0.....0fX...0..f..0.X...0.....0....
*.....0f....0.X...0....f0.X...0f....0.X...0.....*..

OFDM symbol
..*.X..f0.....0.X...0..f..0.X..f0.....0.X...0.....*
....0.Xf..0.....0.X...0f....0.Xf..0.....0.....0....
400 ms frame

t *.....0fX...0.....0.X..f0.....0fX...0.....0.....*..
..*....f0.X...0.....0.Xf..0....f0.X...0.....0.....*
....0..f..0.X...0.....0fX...0..f..0.X...0.....0....
*.....0f....0.X...0....f0.X...0f....0.X...0.....*..
..*.X..f0.....0.X...0..f..0.X..f0.....0.X...0.....*
....0.Xf..0.....0.X...0f....0.Xf..0.....0.....0....
*.....0fX...0.....0.X..f0.....0fX...0.....0.....*..
..*....f0.X...0.....0.Xf..0....f0.X...0.....0.....*
....0..f..0.X...0.....0fX...0..f..0.X...0.....0....

freq. pilots for manual tunning

Figure 10.9: Example of frame for Mode B, spectrum occupancy 2.5 kHz. Leg-
end: : – DC-carrier, . – MCS cells, X – FAC cells, T – time pilots, f – frequency
pilots, 0 – scattered pilots, * – boosted scattered pilots.

⊳ it is possible to decode and display image during transmission;

⊳ transfer speed is better up to 3×;
⊳ HamDRM continuously broadcasts station identification, so receiving operator
can start reception and direct yagi;
⊳ thanks to several instances, it isn’t necessary to record the transmission from
beginning to end;
⊳ when reception failed, only bad segments can be repeated, not whole transmis-
sion;
⊳ main disadvantage of HamDRM is, that the powerful PC configuration and OS
better then Windows 2000 is a must.

10.2.2 Quadrature amplitude modulation — QAM

Quadrature amplitude modulation (QAM) uses amplitude and phase modulation
together. HamDRM for each subcarrier (OFDM cells) can use several modulation
schemes, which differ in number of modulation states – QAM-4, QAM-16 and QAM-
64. √
The number
√ of modulation states QAM-𝑚 is divided into 𝑚 states for phase
keying and 𝑚 amplitude levels. Thanks to multistate modulation it is not required

120
 DSSTV transmission systems 10

so huge bandwidth, on the other hand, a growing number of states of modulation

makes the signal less resistant to interference.

pulse I
low-pass I cos(2πft)
amplitude
filter 
modulation

cos

mapping carrier QAM
data circuit frequency
–
signal
sin

pulse Q
low-pass
amplitude
filter
 Q sin(2πft)
modulation

Figure 10.10: The QAM modulator.

An modulation state is created from combination of amplitude and phase, which can
define a bit word of length 𝑙. For QAM-4 is the word length 𝑙 = log2 𝑚 = log2 4 = 2,
for QAM-16 is 𝑙 = 4 and for QAM-64 it is 6. The modulation changes between these
states:
√ √
𝐴𝑘 = 2𝑘 − 1 − 𝑚 pro 𝑘 = 1, 2, … 𝑚.
E.g. for QAM-16 levels are −3, −1, 1, 3.
The signal, which can be presented like
𝑆𝑘 (𝑡) = 𝐴𝑘 cos(2𝜋𝑓𝑡 + 𝜙𝑘 )
has 16 combinations of amplitudes 𝐴𝑘 and phases 𝜙𝑘 .
The block diagram of QAM modulator see in fig. 10.10. Now, we describe how
QAM-16 modulates data sequence 𝑁 = {0, 13, 5, 2, 10, 7, 6, 5, 1, 15}. The
result is fig. 10.12. The information words with 4bit length are divided on two
parts in mapping circuit and first 2bit combination is coded in pulse amplitude
modulation (PAM) into one of four levels. The way how to code input bit quaternion
{𝑖0 , 𝑖1 , 𝑞0 , 𝑞1 } is defined by constellation diagram, fig. 10.11. E.g. for input 0
it is 𝑖0 𝑖1 = 00, 𝑞0 𝑞1 = 00 and this corresponds to 𝐼 = 3, 𝑄 = 3, the next value
13, in binary 1101 corresponds 𝑖0 𝑖1 = 11 output 𝐼 = −3 and for 𝑞0 𝑞1 = 01 output
𝑄 = −1, etc.
The results of PAM are pulses with given amplitudes and they are filtered with
low-pass filter for the bandwidth reduction and for in phase path 𝐼 and similarly
for quadrature path 𝑄. The 𝐼 and 𝑄 are input signals for modulators with carrier

121
10 DSSTV transmission systems

frequency 𝑓. This way there is a phase of 90∘ between them. Output signal is made
by joining of both paths together:
𝑆𝑘 (𝑡) = 𝐼𝑘 cos(2𝜋𝑓𝑡) − 𝑄𝑘 sin(2𝜋𝑓𝑡).

q0q1
1100 0100 1000 0000
3 00

1110 0110 1010 0010

1 10
I

–3 –1 1 3
–1 01
1101 0101 1001 0001

–3 11
1111 0111 1011 0011

i0 1 0 1 0
i1 1 1 0 0

Figure 10.11: The constellation diagram for QAM-16 with bit

order {𝑖0 , 𝑖1 , 𝑞0 , 𝑞1 } used in DRM.

10.2.3 Orthogonal frequency-division multiplexing — OFDM

OFDM is a representative of the modulation scheme with multiple carriers MCM
(Multicarrier Modulation). Thanks to its properties the OFDM found application in
many modern technologies, i.e. ADSL, WiFi (IEEE 802.11a/g) networks, WiMAX
and standards for digital broadcast and terrestrial digital television DVB-T, etc.
OFDM has very good spectral performance and it is resistant to pulse interfer-
ence, because transfered information is dispersed in wide frequency spectrum, so
interference disturb only few nearby symbols. It’s also resistant to inter-symbol in-
terference, fade outs caused by multipath spreading and has low sensitivity to errors
in time synchronization.

122
DSSTV transmission systems 10

16
14
Input data N

12
10
8
6
4
2
0

3
Signal I

1
0
−1
−3

3
Signal Q

1
0
−1
−3

4
Signal QAM

2
0
−2
−4

Time t
Figure 10.12: The example of QAM-16
modulation for input data sequence.

The OFDM generates a huge number of subcarrier waves and in case of HamDRM
there are for best performance only 57 subcarriers. Many other applications like
digital video broadcast or wideband data communication uses hundreds or thousands
of subcarriers! These subcarriers have very small distances, even those, that the
overlap the range of others. An example of OFDM spectrum is in fig. 10.13, as
spectrum of each subcarrier is considered the spectrum of rectangular signal, which
is expressed by sin(𝑥)/𝑥 function.
The subcarriers has exact distances, so maximal level of spectrum of each subcar-
rier is null in maximal levels of other subcarriers, so they are mutually orthogonal.

10.2.3.1 OFDM transfer

The modulator block diagram is in fig. 10.14. Input data stream comes to serial-
parallel converter and it is cyclically distributed to a larger number of parallel com-

123
10 DSSTV transmission systems

Amplitude A

Frequency f
Figure 10.13: The frequency
spectrum of orthogonal carrier waves.

ponents. The parallel component transmitted simultaneously creates a complete

OFDM symbol. Components are also modulated to the orthogonal system of 𝑁
subcarriers, the frequencies are distributed to ensure their orthogonality. Subcarrier
waves in our case use modulation QAM-4, QAM-16 or QAM-64, but for some other
applications there are used multiphase BPSK or QPSK.
A signal processor provides modulation of huge number of subcarriers, in our case
it is software, which implements algorithms for inverse discrete Fourier transform
(DFT−1 ). Because DFT algorithm has big computing complexity there is used its
faster variant FFT (Fast Fourier Transform). The inverse FFT (FFT−1 ) transform
input data from frequency domain to time domain. The process on a receiver side
vice-versa use direct FFT to obtain individual subcarriers.
Two data stream are outputs of FFT−1 , which are converted with digital/analog
converters on two analog signals. Then these signals are modulated to main carrier
and there is a phase of 90∘ between them. The Re signal presents amplitude com-
ponent and Im signal phase component. Both joined together creates transmitted
OFDM signal.
Everything on reception side goes in opposite way. The received signal is amplified
and converted to lower frequency. Then signals Re and Im go through low-pass
filters to analog/digital converters and data from them is processed by DSP with
direct FFT and divided into individual subcarriers. The output data are compiled
in parallel-serial converter.

124
DSSTV transmission systems 10

constellation
mapping
D/A Re
convertor

serial
parallel
convertor cos
Inverse
Discrete carrier
Fourier  OFDM
data frequency
Transformation signal

FFT–1 sin

Im
D/A
convertor


Figure 10.14: The OFDM modulator use fast Fourier transform (FFT−1 ) for
making a huge number of modulated subcarriers.

10.3 DSSTV software selection

There is several programs available supporting HamDRM and RDFT.
Software RDFT HamDRM Web page
DIGTRX 3.11 ⋆ ⋆ http://www.qslnet.de/member/py4zbz
/hdsstv/HamDRM.htm
DigiACE V1.9 ⋆ http://homepage.ntlworld.com/mhemmerson/
DigPAL ⋆ http://www.home.bellsouth.net/p/PWP-hampal
EasyPAL ⋆ http://vk4aes.com/
HamPAL ⋆ http://www.home.bellsouth.net/p/PWP-hampal
RDFT ⋆ http://www.svs.net/wyman/examples/hdsstv/
RXAMADRM (Linux) ⋆ http://pa0mbo.nl/ties/public_html
/hamradio/rxamadrm/index.html
SSTV-PAL Multimode ⋆ http://f6baz.free.fr/FTP/SSTVPalPlus/
WinDRM ⋆ http://n1su.com/windrm/

10.4 Making QSO

Digital SSTV is not spread too far. There are few station found sporadically on the
14MHz band. But there is also working party of German stations on the 3.7MHz
band around the frequency 3.733 kHz almost daily in the evening. Stations use
only HamDRM system. Listening to their signals is a good opportunity to try
DSSTV reception and get some practice with it, also try to make contact. After

125
10 DSSTV transmission systems

that, you already know how there is used special modulation schematic and error-
correcting coding it is important to see if it at all works and how. Will be there
SSTV digitalization boom?
Some opponents of digital video broadcast claim, that in conditions where we can
receive noisy, but still usable analog TV signal, the digital TV cannot be received at
all. And same argument can say opponents of digital-SSTV. When there are good
conditions, it is possible only to tune on channel, images are received automatically
and operator should not do anything. When interference gets stronger and signal
weaker there can help more data instances or bad segment report and additional
repetition of bad segments. But when we only guess HamDRM signal drowned in
high noisy level the reception is impossible.
The DSSTV traffic can be found on band near the centre of SSTV activity. Also
hamspirit rules should be observed and we should be considerate to another traffic
on the band. Sometimes it takes a little tact to explain to uninformed station, that
the strange rattling sound is the digital signal from your QSO partner.
A CQ call can be done by sending picture on free frequency. HamDRM during
transmission broadcast station id, so is you don’t receive complete data you can see
what station is transmitting. After the end of transmission you can call the station
by voice.
For reception confirmation or short message transfer there is used waterfall mes-
sages – messages displayed in tuning indicator. Principle of these messages is de-
scribed in next section 10.5 and example of some message see in fig. 10.15.

Figure 10.15: The confirmation of successful re-

ception displayed in tuning indicator

The reports are same as for the phone operation in the RS (readability and strength)
code. The V (View) value representing image quality of digital transmission is losing
its importance. Readability is measured on a scale of 1 to 5, so level 5 stands for
for a perfect error-free transfer, level 4 is 4 still acceptable, with occasional failure

126
DSSTV transmission systems 10

segments and potentially it’s needed to increase the number of instances. Report
the worst level 1 if can not receive any digital data.
Contrary to popular SSTV operations when stations restrict only to the exchange
of images, the phone mode is much more used in case of DSSTV.
The choice of images is not limited to the usual 320 × 240 resolution, but there
can be used any resolution. The limiting factor is only time of transmission, e.g. in
the DIGTRX software the broadcast time is already known, so you can play around
with the compression level, resolution, or number of colors and achieve a reasonable
compromise.
Also, the transmitted data file format can be any. Listen on the band and you
will make sure that JPEG2000 is often used, but also animated GIFs, or text files
with ASCII art.

10.5 Waterfall images

For digital SSTV and RDFT or HamDRM system is used tuning indicator, which
displays spectrum of SSB channel. The image showed by indicator is created using
discrete Fourier transformation. The indicator displays new samples on top and the
old samples disappear at the bottom and the whole spectrogram is moving down so
the indicator was nicknamed waterfall.
In the fig. 10.6 and fig. 10.8 you can see station and software identification
and also messages about reception confirmation, request for repeat or more complex
pictures also.
The principle of “waterfall images” is based on Fourier transformation and the
fact, that the signal can be compiled from a huge number of harmonic waves. If the
proper harmonic are compiled, so the resulting carrier wave has frequency spectrum,
that will look like desired image.
The utility PicFall.exe can be used for generating sound file from picture. You
can find it on website of DIGTRX author. The input file is a bitmap in BMP format
and output is WAV audio file.
Generate waterfall image by using PicFall.exe:
http://www.qsl.net/py4zbz/tutsstv14.htm

127
 10 DSSTV transmission systems

Spectrum Carrier

f
0 500 1000 1500 2000 2500 3000

Figure 10.16: The principle of waterfall image display.

128
 11
Facsimile — Radiofax

Facsimile (from latin facere – make and simile – similar) is one of the oldest com-
munication modes and it is used for an image transmission. The facsimile is mainly
used by professional services for wireless distribution of meteorological maps and
informations, hence the name Weather Facsimile (WEFAX) follows from it. The
radiofax can be used by radio amateurs too.

11.1 The history of image transmission

Already in 1843 a Scottish clockmaker Alexander Bain suggested that some images
can be broadcast via electric lines, when it is electrically scanned by rows and point
by point. That’s the basic idea of image transmission.
Bain’s example was soon followed by other entrepreneurs. In 1847 an Englishman
Frederick Collier Bakewell reeled an image in the transmitter and sheet of paper to
cylinder in the receiver, which was turned by the clock machine. The picture were
printed with fat on a tinfoil sheet.

Figure 11.1: Bakewell’s

image telegraph.

129
11 Facsimile — Radiofax

A honor for the first fax service, however belongs to Giovanni Caselli, Italian catholic
priest and physicist. In 1856 he built device named panthèlègraph. With pantèlègraph
could be sent images or texts.
Caselli received enthusiastic support from the French emperor Napoleon III. The
emperor personally visited his workshop in 1860 and enabled him an access to teleg-
raphy lines. The first commercial fax service started in 1865 in Paris and it was
connecting some major cities of France. Indeed the transmission was very slow and
the fees were high so there was only few clients. The service wasn’t profitable and
had to be stopped.
In 1901 a German scientist Dr. Arthur Korn invented the principle of the pho-
toelectric reading and began to transfer some positive photographic slides on a
transparent substrate. These slides were illuminated point by point and row by
row and light passed through transparent slide influence the selenium cell. The cell
changes its resistance depending on the light intensity and transform image pixel
shade to electric current. The receiver contained “light relay”, a device with early
gas-discharge lamp. The intensity of light exposed present point on photographic
paper and it varied according to current flowed form receiver.
Dr. Korn designed the first phototelegraph device in 1902 and already in March
1904 he managed to reproduce photo transfered form Munich to Nuremberg. The
transfer of postcard size photo took 24 minutes. In 1907 major cities Berlin, Munich,
Paris and London were linked and his devices were bought by newspaper agencies
and the first phototelegraphic service was founded.
The transfer was simplified by using of an electric photocell. Thanks to the
photocell the transfer speed increased and the preparation of transparent slides
wasn’t necessary. The photocell is so sensitive that it is influenced by reflected
light and some photos could be scanned directly. The photocell was used for the
first time by American captain Richard H. Ranger for test transmission between
Cleveland and New York. In November 30th, 1924 was wirelessly transfered photo
of the British royal couple from London to New York. The first phototelegraphic
service between America and Europe was established in May 1st, 1926.

11.2 The fax mode

The modern facsimile (fax) is used for transmitting images in the high resolution
(usually 1810 dots per line) with image size up to several thousand lines. An rela-
tively long time of transmission is used due to small bandwidth. In dependency on
image size and transmission speed it takes from 3 to 20 minutes.

130
Facsimile — Radiofax 11

11.2.1 Image transmission

A typical mechanical transmitter consist of cylinder rotated by crystal controlled
synchronous motor. Broadcast material is attached to the cylinder, which rotates
in a constant speed. A small ray of light is focused on the broadcast medium
(map, text, photos, etc.). The light reflected from the medium is processed by a
photoelectric sensor. The sensor bears the light source, photocell and moves along
rolled in a constant speed. The sensor moves from one end to the other and captures
the image line by line. Voltage difference from the photo sensor are amplified and
it is used to modulate the signal carrier.

direction of
rotation

aperture
photo cell
exciter
lamp

Figure 11.2: The principle of an

electromechanical image capture.

The fax transmission on shortwaves has only few similarities with the fax machines
you know from home and offices. The fax transmission is based on CCITT recom-
mendations Facsimile CCITT Group 1 (T.2) of 1968, the short wave recommenda-
tion is described chapter 11.6. The CCITT recommends the frequency 1500 Hz for
white and 2300 Hz for black. In North America it is 1500 Hz for white and 2300 Hz
or 2400 Hz for black. The transmission speed is 180 lpm. The fax machines of this
type could be adjusted for amateur operation. The later recommendations T.3, T.4
or T.30 can not be used on HF and there are used in telephone lines.
The frequency modulation F3C is used for shortwave transmission (𝐹 – frequency
modulation, 3 – single channel containing analog information, 𝐶 – facsimile). The

131
11 Facsimile — Radiofax

transmitter modulates the frequency of carrier in the range ±400 Hz on shortwaves

and ±150 Hz on long waves. This range is called signal deviation.
The fax signal can be created by direct modulation of broadcast frequency or by
frequency modulation of subcarrier 1900 Hz. Then the transmitter changes fre-
quency between black and white colour. Black color corresponds to 1500 Hz and
white 2300 Hz.
The amplitude modulation (AM) used for image transmission on VHF and mi-
crowave meteo satellite downlinks. There is used positive AM and level of modu-
lation determines the brightness. For black colour it is 4 % level and for white it
is 90 % to 100 %. The negative modulation inverts levels, the minimal value is for
white and the maximal value for black.
Most commercial stations use APT – Automatic Picture Transmission for a fully
automated reception without requiring the presence of an receiver operator. The
image transmission begins with start tone, when transmitter modulates the carrier
with some frequency, mostly 300 Hz (changing the maximal levels of modulation
300× in a second). This signal is recognized be the receiving device and it switches
from stand-by mode to working mode and waits for phasing signal.
Starting APT signal

Phasing signal
Image edge

Figure 11.3: The start of facsimile transmission.

The phasing signal is used for synchronization and it is broadcast few second before
an image. Normally consist of rows of 95 % black and 5 % white. It generats a
vertical white line, which identifies the edge of the transmitted image.
After the end of image transmission another APT signal is sent, it has modulation
450 Hz and switches the receiving device back to stand-by mode.

11.2.2 The reception

Shortwave facsimile reception can be realized by using upper single sideband (USB)
receiver. For this reason, you’ll have to tune on the frequency, which si about
1900 Hz below the station frequency. So if the station uses frequency 3855.0 kHz,
you must tune in USB on 3853.1 kHz (i.e. 1900 Hz below), see 11.4. All facsimile
reception software is equipped by spectroscope, the same as for SSTV, so precise
tuning should not be a problem. You can control tuning by the fact that the largest
portion in the fax image is mostly white.

132
Facsimile — Radiofax 11

Frequency of USB carrier,

displayed by receiver Frequency of transmitter,
3853.1 kHz i.e. 3855 kHz (DDH3)

−400 Hz +400 Hz
deviation deviation

1500 Hz
1900 Hz
2300 Hz

black white
Upper Side Band (USB) channel width

Figure 11.4: Relation between

reception and transmission frequencies.

There are still used analog WEFAX receivers with electromechanical printers, but
in our case we can use only PC with sound card and proper software.

Figure 11.5: Modern receiver Sony CRF-V21 for WEFAX

and RTTY reception equipped with printer.

The most important parameters of the transmission are the speed and index of
cooperation – IOC. The IOC relates with a horizontal scan rate and can be converted
to number of pixels by simple formula:

133
11 Facsimile — Radiofax

line = 𝜋 × IOC pixels

The most frequently used IOC is 576 (1810 pixels), then 288 (900 pixels).

Figure 11.6: Typical synoptic map transmitted on HF by facsimile,

here from station DDH3, line speed 120 lpm and IOC 576.

The speed of transmission is given by rotation of cylinder (round per minute, rpm)
and it is equivalent to number of lines per minute, lpm. Professional stations use
most often 120 lpm, in eastern European countries and in post-soviet states it is
90 lpm, news agencies use 60 lpm and meteo satellites 240 lpm.
The most common fax images (synoptic maps) are transmitted only in black and
white, but some images like retransmission of weather satellite images are in gray
scale.

11.2.2.1 Facsimile transmission modes

11.3 Professional stations

There are dozens of stations operating on high frequency bands. It depends on
your location, some of them are well-catchable all day, others only when conditions
improve. Broadcast images are in most cases a variety of meteorological maps,
synoptic charts, graphs of pressure and altitude, direction of wind, weather forecasts,

134
Facsimile — Radiofax 11

Name IOC lpm APT start d APT stop d Note

Wefax 288 288 120 / 90 / 60 675 Hz 3s 450 Hz 3s
Wefax 576 576 120 / 90 / 60 300 Hz 3s 450 Hz 3s
Ham Color 204 360 200 Hz 5s 450 Hz 5 s color
Ham 288b 240 240 / 120 675 Hz 5s 450 Hz 5s
Color 240 288 240 200 Hz 3s 450 Hz 5 s color
FAX 480 204 480 500 Hz 3s 450 Hz 3s
Photopress 352 60 ? ? 450 Hz ?
𝑑 – minimal duration of APT signal
Table 11.1: Facsimile transmission modes.

cyclone or typhoon warnings, retransmissions of satellite imagery and broadcast of

news agencies.
Even today when Internet is almost everywhere the fax broadcast has still its
foundation. The main customers are naval ships, military, remote airports and
islands, where the shortwave transmission is only way how to get actual information.
They are often very important, because station distributes also weather warning of
upcoming storm and hurricanes.
Each station has given its daily broadcast schedule, for example see 11.7. You
can find here what images will be transmitted in a time of day.
For first experiments with facsimile reception are suitable strong local stations.
Here in Europe it is the German station DDHx, which is active on frequencies
3855.0, 7880,0 a 13 882,5 kHz. As already announced, station always receive in USB
mode and tune it 1900 Hz below. Therefore DDH3 tune on 3853.1 kHz. Its speed is
120 lpm and IOC 576.
Another station is an Englich GYA broadcasting from Northwood (120/576) on
frequencies 2618.5, 4610.0, 8040.0, 11085.5 kHz (active are at least two channels
simultaneously).
From long distance stations cab be received almost daily Tokyo station JMH4.
Of the three transmitters is the best JMH4 on 13597.0 kHz with 5 kW output power.
Besides the usual synoptic chart there are available satellite imagery too.
Next Tokyo station is JJC, it is Kyodo News Agency. Transmission speed is
60 lpm and IOC 576, sometimes 120 lpm when weather charts are posted. The
station broadcast simultaneously on several frequencies: 12745.5, 16971.0, 17069.6,
22542.0 kHz, by listening find the active frequency. Interestingly, the owner of JJC
station was in 1997 thinking about closure, if they find some other way for news

135
11 Facsimile — Radiofax

Figure 11.7: The station schedule of GYA.

Figure 11.8: Test chart of DDH3

transmitted daily in 11:10 UTC.

distribution. It didn’t stop yet and Kyodo News Agency still distributes newspapers
via facsimile.

136
Facsimile — Radiofax 11

Figure 11.9: Test chart from

JMH4 available daily at 13:00 UTC.

Figure 11.10: Typical JJC transmission.

For those who deal with DX radio reception on the HF While listening to the
interesting DXů will be rewarded by received image. Complete list of stations sorted
by country or frequency see chapter 12.
The detail list of frequencies and station schedule can be find in publication Ma-
rine Worldwide Radiofacsimile Broadcast Schedules, which is published by the Na-
tional Oceanic and Atmospheric Administration’s (NOAA) and it’s freely available
to download.
Radiofacsimile Worldwide Marine Broadcast Schedules:
http://www.nws.noaa.gov/om/marine/rfax.pdf
Another excellent source of information for those who are interested in receiving of
meteorological data is the website of the World Meteorological Organizations W.M.O.
Besides general information about meteorology there are lists of frequencies and

137
11 Facsimile — Radiofax

schedules, not only for the fax, but also for other professional stations which use for
distribution radio teletype (RTTY), NAVTEX and other types of digital modes.
World Meteorological Organization: http://www.wmo.ch/

11.4 Satellite imagery retransmission

Following list is compiled from It’s the list of stations retransmitting meteo-satellites
imagery. All listed stations transmit in 120 lpm and IOC 576.
UTC Station Ident. Frequencies (kHz) Note
00:34 Hawai, USA KVM70 9982.5; 11090; 16135 East pacific GOES IR
00:48 Hawai, USA KVM70 9982.5; 11090; 16135 SW pacific GOES IR
01:01 Halifax, Canada CFH 122.5; 4271; 6496.4; 10536; 13510 IR
01:10 Tokyo, Japan JMHx 3622.5; 7795; 13988.5 MTSAT
01:20 Taipei, China BMF 4616; 8140; 13900; 18560 GMS
01:43 California, USA NMC 8682; 12786; 17151.2; 22527 NE GOES IR
01:54 California, USA NMC 8682; 12786; 17151.2; 22527 Pacific GOES IR
02:00 Lousiana, USA NMG 4317.9; 8503.9; 12789.9; 17146.4 Tropical GOES IR
03:51 Massachusetts, USA NMF 4235; 6340.5; 9110; 12750
05:06 Alaska, USA NOJ 2054; 4298; 8459; 12412.5 GOES IR
06:35 Hawai, USA KVM70 9982.5; 11090; 16135 East pacific GOES IR
06:49 Hawai, USA KVM70 9982.5; 11090; 16135 SW pacific GOES IR
07:10 Tokyo, Japan JMHx 3622.5; 7795; 13988.5 MTSAT
07:20 Taipei, China BMF 4616; 8140; 13900; 18560 GMS
07:37 California, USA NMC 8682; 12786; 17151.2; 22527 Tropical GOES IR
08:00 Lousiana, USA NMG 4317.9; 8503.9; 12789.9; 17146.4 Tropical GOES IR
09:06 Hawai, USA KVM70 9982.5; 11090; 16135 Pacific GOES IR
09:08 California, USA NMC 8682; 12786; 17151.2; 22527 Pacific GOES IR
09:43 Hawai, USA KVM70 9982.5; 11090; 16135 Tropical GOES IR
09:51 Massachusetts, USA NMF 4235; 6340.5; 9110; 12750
10:22 Halifax, Canada CFH 122.5; 4271; 6496.4; 10536; 13510 IR
11:17 Alaska, USA NOJ 2054; 4298; 8459; 12412.5 GOES IR
11:30 Playa Ancha, Chile CBV 4228; 8677; 17146.4
12:32 Hawai, USA KVM70 9982.5; 11090; 16135 East pacific GOES IR
12:48 Hawai, USA KVM70 9982.5; 11090; 16135 SW pacific GOES IR

138
Facsimile — Radiofax 11

UTC Station Ident. Frequencies (kHz) Note

13:10 Tokyo, Japan JMHx 3622.5; 7795; 13988.5 MTSAT
13:20 Taipei, China BMF 4616; 8140; 13900; 18560 GMS
14:00 Lousiana, USA NMG 4317.9; 8503.9; 12789.9; 17146.4 Tropical GOES IR
14:03 California, USA NMC 8682; 12786; 17151.2; 22527 NE GOES IR
14:14 California, USA NMC 8682; 12786; 17151.2; 22527 Pacific GOES IR
15:03 Massachusetts, USA NMF 4235; 6340.5; 9110; 12750
16:45 Playa Ancha, Chile CBV 4228; 8677; 17146.4
17:06 Alaska, USA NOJ 2054; 4298; 8459; 12412.5 GOES IR
18:35 Hawai, USA KVM70 9982.5; 11090; 16135 East pacific GOES IR
18:49 Hawai, USA KVM70 9982.5; 11090; 16135 SW pacific GOES IR
19:02 California, USA NMC 8682; 12786; 17151.2; 22527 Tropical GOES IR
19:10 Tokyo, Japan JMHx 3622.5; 7795; 13988.5 MTSAT
19:20 Taipei, China BMF 4616; 8140; 13900; 18560 GMS
19:30 Playa Ancha, Chile CBV 4228; 8677; 17146.4
20:00 Lousiana, USA NMG 4317.9; 8503.9; 12789.9; 17146.4 Tropical GOES IR
21:06 Hawai, USA KVM70 9982.5; 11090; 16135 Pacific GOES IR
21:13 California, USA NMC 8682; 12786; 17151.2; 22527 Pacific GOES IR
21:43 Hawai, USA KVM70 9982.5; 11090; 16135 Tropical GOES IR
21:51 Massachusetts, USA NMF 4235; 6340.5; 9110; 12750
23:17 Alaska, USA NOJ 2054; 4298; 8459; 12412.5 GOES IR
23:25 Playa Ancha, Chile CBV 4228; 8677; 17146.4

It depends on your own position if you have interest in reception of images from
these stations. For long distance stations should not be reception conditions so good
every day. Their images for me in Europe is not interesting for weather forecast, but
there are images of hurricanes, typhoons and other unusual weather phenomena.

11.4.1 Meteorologic satellites

If you are interested in reception of satellite imagery, there is few basic informations.
For amateur reception it is possible to use NOAA satellites on low Earth orbit.
These satellites transmit in 137MHz band in WEFAX mode with amplitude modu-
lation, so an sound card can be used as signal decoder with some dedicated software
like JVComm32 or WXtoIMG. For the best reception should be used a receiver with
30 kHz intermediate frequency (IF) width, which unfortunately common receivers
and TRXes don’t support. The narrower IF causes image distortion and receiver for

139
11 Facsimile — Radiofax

Figure 11.11: Restransmission of satellite image from Boston NMF station re-
ceived on 12.750 kHz, hurricane Katrina is devastating Mexican gulf

wideband FM (about 200 kHz) is not very suitable due to more noise that affecting
signal. Also it is necessary to use an antenna with right-handed circular polariza-
tion as turnstile antenna (crossed dipoles) or QFH (Quadrifillar Helix Antenna).
Satellites NOAA, MetOp and Fengyun are carrying high resolution scanners and
digital transmitters – HRPT (High Resolution Image Transmission). They broadcast
on 1.6GHz band, but used system is digital and for data reception must be used
band converter, special modem and main difficulty rests in need of antenna aiming.
A computer controlled rotator is needed for aiming of azimuth and elevation.
Next option is a reception of geostationary satellites of Meteosat and GOES fam-
ily. The analog WEFAX broadcast in 1.6GHz band was discontinued. For analog
WEFAX reception could be used NOAA receiver with band converter and dish or
yagi antenna.
The follower of the analog broadcasst is new system MSG (Meteosat Second
Genereation), it is fully operational since 2004. Digital data LRIT (Low Rate
Information Transmission) and HRIT (High Rate Information Transmission) are
broadcast via television transponder EuroBird 9 (Ku band 10.7 – 12.75 GHz) on
11,976.82 MHz (EUMETCAST). An extension PCI card for DVB-S reception is used
for data reception, e.g. SkyStar 2 card for PCI or external version for USB. The

140
Facsimile — Radiofax 11

data are decoded with tq-TELLICAST software. The disadvantage is that almost
all data is distributed encrypted, so it is necessary to register at Eumetsat provider
and buy the hardware decoding key. The price for hobby purposes is €100 (software
is for €60 and key is for €40). The are also higher demands on PC configuration:
2GHz CPU, 1 GB RAM and 36 GB hard disk.
The radiometer of Meteosat 8 and 9 provides images in 11 spectral channels in
3 km/pixel resolution and in HRV (High Resolution Visible channel it is 1 km/pixel,
although in regard to slant projection of the Earth’s surface the resolution for Europe
and Globe edges is lower. Image data (High Rate SEVIRI) have a standard size
3712×3712 pixels and for HRV it is 5568×11136 pixels. Data from satellite are send
first to the primary station in German Darmstadt and then they are processed and
they are distributed via EuroBird 9.
The Meteosat 9 provides image of Earth globe every 15 minutes, the Meteosat 8
sends data every 5 minutes (Rapid Scanning Service), but only European part of
globe is sent. In addition to these data via EUMETCAST are broadcast further
meteorological products, such as NOAA and MetOp HRPT imagery and processed
data from other satellite sensors (infrared, microwaves,…).

11.4.2 Essential Services

The unencrypted data from Meteosat 9 and Meteosat 7 are transmitted every six
hours, also GOES and MTSAT-1R is available every 3 hours.

Satellite Interval Transmittion times (UTC)

Meteosat 9 HRIT/LRIT 6 hours 05:45, 11:45, 17:45, 23:45
Meteosat 7 6 hours 00:00, 06:00, 12:00, 18:00
GOES 9, 10, 12; MTSAT-1R 3 hours 00:00, 03:00, 06:00, 09:00, 12:00, 15:00, 18:00, 21:00

Table 11.2: Essential services data.

11.5 Hamradio facsimile operations

The facsimile operation did not spread as SSTV in the past years. The reason for
this is using of complex mechanical scanners and recorders and also a relatively long
time of image transfer. A little development came with the computer software and
cheap interfaces, but it is very rare to heard amateur fax on bands.
The use of facsimile fits for very high resolution images, which is better than any
SSTV mode. The number of lines for image is not given, but aspect ratio 4:3 should
be used. Hamradio operators use IOC 576 or 288 and speed 120 lpm or 240 lpm, the
usage of other modes depends on agreement of both parties. A report is given in
common RST code (Readibility, Strenght, Tone).

141
11 Facsimile — Radiofax

Figure 11.12: Hamradio facsimile from ON7BW received on 14MHz band, speed
240 lpm, IOC 288.

An opportunity to receive rare facsimile amateur stations is 3rd weekend in Au-

gust, when The International HF – FAX – Contest by DARC is active. Those who
are interested in this kind of communication mode may also try to obtain a diploma
awarded by the DARC for two-way contacts.

11.5.1 EU – FAX – Diplom

The diplom is award for two-way facsimile contacts with European countries. There
are three degrees for QSO with 10 prefixes in 5 countries, 20 prefixes in 10 countries
and 40 prefixes in 20 countries. European countries are given by WAE list. Valid
QSOs are from 1/1/1980 and QSL cards must have note 2-WAY FAX. A confirmed
list of QSL cards and €5 send to: DARC FAX Manager, Werner Ludwig DF5BX,
Post Box 1270, D-49110 Georgmarienshutte, Germany.

11.5.2 The International HF – FAX – Contest by DARC

The contest is organized by the Deutscher Amateur Radio Club. Ongoing 3rd week-
end of the August, starts at 8:00 UTC on Saturday and ends on Sunday at 20:00

142
Facsimile — Radiofax 11

UTC. It progress on all HF bands except the WARC. Assessed by two classes – lis-
teners and one operator. All QSOs must be done in facsimile mode and image calls
should include CQ FAX TEST. Report the RST and the number of connections from
001. Any QSO is a point valued, multipliers are countries WAE/DXCC and dis-
tricts W, VE, and JA. QSO with a same station are valid on more bands. The log
should be sent within 2 weeks after the contest to: Werner Ludwig DF5BX, Post
Box 1270, D-49110 Georgsmarienshuette, Germany, email: df5bx@darc.de.

11.6 International facsimile standard recommedation

1. Drum speed:
60, 90, 120, 240 revolutions per minute, if speeds greater then 120 rpm are
used, they should be multiples of 60 rpm.
2. Diameter of drum:
152 mm, in the case of flat-bed scanners this will be length of the scanning
line (including the dead sector).
3. Index of co-operation (IOC):
⊳ 576 for minimum black or white picture elements of 0.4 mm and
⊳ 288 for minimum picture elements of 0.7 mm.
4. Length of drum:
the length of the drum should be at least 550mm
5. Spanning density
Scanning density = IOC / diameter of drum
It is approximately: 4 lines per mm for index 576, 2 lines per mm for index
288.
6. Direction of scanning:
at the transmitter, the plane (developed in the case of drum transmitter) of
the message area is scanned along lines running from left to right commencing
in the left hand corner at the bottom and this is equivalent of scanning over a
left hand helix.
7. Dead sector:
4.5 % ± 0.5 % of the length of the scanning line. The signal transmitted during
the passage of the dead sector should correspond to white but it is permitted
that a black pulse be transmitted within and not exceeding one half length of
the dead sector.
8. Selection of index of co-operation:
a five second transmission of alternating black and white signal at 300 Hz for
index 576, 675 Hz for index 288. The envelopes of the signals transmitted will
be roughly rectangular.
9. Synchronization:

143
11 Facsimile — Radiofax

the scanning speed should be maintained within 5 part in 106 of the normal
value.
10. Starting recorders:
recorders should be designed to start upon receipt of either the index selection
signal or the phasing signal and no special signal for starting will be transmitted.
11. Phasing:
a 30 second transmission of alternating white and black signal at the following
frequencies:
⊳ 1.0 Hz for speed of 60 rpm,
⊳ 1.5 Hz for speed of 90 rpm,
⊳ 2.0 Hz for speed of 120 rpm.
12. Adjustment of recording levels:
adjustment of recording level when used should be effected by reference to
phasing signal.
13. Stopping recorders:
a 5 second transmission of alternating black and white signals at 450 Hz fol-
lowed by 10 seconds of signals corresponding to continuous black.
14. Modulation characteristics
⊳ Amplitude Modulation
The maximum amplitude of the carrying frequency should correspond to
the transmission of signal black. Value of the carrying frequency: 1800 Hz.
⊳ Modulation by frequency deviation
⊳ Value of central frequency: 1900 Hz.
⊳ Value of frequency for black: 1500 Hz.
⊳ Value of frequency for white: 2300 Hz.
The frequency for black and white should not vary by over 8 Hz over a
period of 30 s and by more than 16 Hz over a period of 15 minutes.
15. Levels of signals in case of AM:
receiving equipment should accept any level between +5 dB and −20 db, zero
reference level corresponding to a power of one milliwatt dissipated in a resistence
of 600 ohms.
16. Contrast ratio
contrast ratio for picture signals and control signals will be the same for any
transmission and will be between 12 and 25 dB.
17. Facsimile transmission op meteorological charts over radio circuits:
⊳ When frequency modulation of the sub-carrier is employed for the facsimile
transmission of meteorological charts over radio circuits, the following char-
acteristics should be used:
⊳ Centre frequency: 1900 Hz.
⊳ Frequency corresponding to black: 1500 Hz.
⊳ Frequency corresponding to white: 2300 Hz.

144
Facsimile — Radiofax 11

⊳ When direct frequency modulation (FSK) is employed for the facsimile trans-
mission of meteorological ctorts over radio circuits, the following character-
istics apply:
⊳ Decametric waves (3 MHz – 30 MHz)
⊳ Centre frequency: 𝑓0 .
⊳ Frequency corresponding to black: 𝑓0 − 400 Hz.
⊳ Frequency corresponding to white: 𝑓0 + 400 Hz.
⊳ Kilometric waves (30 kHz – 300 kHz)
⊳ Centre frequency: 𝑓0 .
⊳ Frequency corresponding to black: 𝑓0 − 150 Hz.
⊳ Frequency corresponding to white: 𝑓0 + 150 Hz.

145
 12
List of professional
stations
There is list of HF professional stations. Be sure to tune in USB mode for 1.9 kHz
lower! Time values are given in UTC Coordinated Universal Time.

12.1 Europe

12.1.1 Athens, Greece

Ident. Frequency lpm IOC Power Note

SVJ4 4,481.0 kHz 120 576 8 kW 08:45–10:44
SVJ4 8,105.0 kHz 120 576 8 kW 08:45–10:44
Broadcast time: 08:45–10:44.

12.1.2 Hamburg/Pinnenberg, Germany

Ident. Frequency lpm IOC Power Note

DDH3 3,855.0 kHz 120 576 10 kW
DDK3 7,880.0 kHz 120 576 20 kW
DDK6 13,882.5 kHz 120 576 20 kW
Broadcast time: 04:30–11:45, 15:20-22:00.

12.1.3 Roma, Italy

Ident. Frequency lpm IOC Power Note

IMB51 4,777.5 kHz 120 576 5 kW
IMB55 8,146.6 kHz 120 576 5 kW
IMB56 13,597.4 kHz 120 576 5 kW

146
List of professional stations 12

Broadcast time: continuous.

12.1.4 Moscow, Russia

Ident. Frequency lpm IOC Power Note

3,830.0 kHz 90, 120 576, 288
5,008.0 kHz 90, 120 576, 288
6,987.0 kHz 90, 120 576, 288
7,695.0 kHz 90, 120 576, 288
RCC76 10,980.0 kHz 90, 120 576, 288
RDD78 11,617.0 kHz 90, 120 576, 288
RCC76 10,980.0 kHz 90, 120 576, 288
Broadcast time: continuous.

12.1.5 Murmansk, Russia

Ident. Frequency lpm IOC Power Note

RBW41 5,336.0 kHz 90, 120 576
RBW41 6,445.5 kHz 90, 120 576 main frequency
RBW41 7,908.8 kHz 90, 120 576 19:00–06:00
RBW48 10,130.0 kHz 90, 120 576 06:00–19:00
Broadcast time: 07:00–08:00, 14:00–14:30, 18:50, 20:00.

12.1.6 Northwood, The United Kingdom

Ident. Frequency lpm IOC Power Note

GYA 2,618.5 kHz 120 576 10 kW 20:00–06:00
GYA 4,610.0 kHz 120 576 10 kW
GYA 8,040.0 kHz 120 576 10 kW
GYA 11,086.5 kHz 120 576 10 kW 06:00–20:00
Broadcast time: continuous, at least on two frequencies.

147
12 List of professional stations

12.2 Africa

12.2.1 Cape Naval, South Africa

Ident. Frequency lpm IOC Power Note

ZSJ 4,014.0 kHz 120 576 10 kW pouze od 16 h do 06 h
ZSJ 7,508.0 kHz 120 576 10 kW
ZSJ 13,538.0 kHz 120 576 10 kW
ZSJ 18,238.0 kHz 120 576 10 kW pouze od 06 h do 16 h
Broadcast time: 04:30–11:00, 15:30, 17:00, 22:30.

12.3 Asia

12.3.1 Beijing, China

Ident. Frequency lpm IOC Power Note

BAF6 5,526.9 kHz 120 576 6-8 kW
BAF36 8,121.9 kHz 120 576 6-8 kW
BAF4 10,116.9 kHz 120 576 10 kW
BAF8 14,366.9 kHz 120 576 15 kW
BAF9 16,025.9 kHz 120 576 ?? kW
BAF33 18,236.9 kHz 120 576 6-8 kW
Broadcast time: 00:08–11:58, 13:40, 19:04–22:40.

12.3.2 Beijing, China

Ident. Frequency lpm IOC Power Note

3SD 8,461.9 kHz 120 576 10 kW
3SD 12,831.9 kHz 120 576 10 kW
3SD 16,903.9 kHz 120 576 30 kW
Broadcast time: 07:55, 11:30.

148
List of professional stations 12

12.3.3 Shanghai, China

Ident. Frequency lpm IOC Power Note

BDF 3,241.0 kHz 120 576
BDF 5,100.0 kHz 120 576
BDF 7,420.0 kHz 120 576
BDF 11,420.0 kHz 120 576
BDG 18,940.0 kHz 120 576
Broadcast time: 00:10, 01:30, 18:10, 20:30.

12.3.4 New Delhi, India

Ident. Frequency lpm IOC Power Note

ATP57 7,404.9 kHz 120 576 10 kW 14:30–02:30
ATP65 14,842.0 kHz 120 576 10 kW 02:30–14:30
Broadcast time: continuous.

12.3.5 Tokyo, Japan

Ident. Frequency lpm IOC Power Note

JMH 3,622.5 kHz 120 576 5 kW
JMH2 7,305.0 kHz 120 576 5 kW
JMH4 13,597.0 kHz 120 576 5 kW
Broadcast time: continuous.

12.3.6 Pevek, Chukotka peninsula

Ident. Frequency lpm IOC Power Note

148.0 kHz 90 576
Broadcast time: 05:30–07:30, 11:30–13:30, 14:30–16:30.

149
12 List of professional stations

12.3.7 Taipei, China

Ident. Frequency lpm IOC Power Note

BMF 4,616.0 kHz 120 576 10 kW
BMF 5,250.0 kHz 120 576 10 kW
BMF 8,140.0 kHz 120 576 10 kW
BMF 13,900.0 kHz 120 576 10 kW
BMF 18,560.0 kHz 120 576 10 kW
Broadcast time: 00:40–10:10, 13:10–22:10.

12.3.8 Seoul, Republic of Korea

Ident. Frequency lpm IOC Power Note

HLL1 3,585.0 kHz 120 576 3 kW
HLL2 5,857.5 kHz 120 576 3 kW
HLL3 7,433.5 kHz 120 576 3 kW
HLL4 9,165.0 kHz 120 576 3 kW
HLL5 13,570.0 kHz 120 576 3 kW
Broadcast time: 00:00–10:40, 12:00–22:40.

12.3.9 Bangkok, Thailand

Ident. Frequency lpm IOC Power Note

HSW64 7,396.8 kHz 120 576 3 kW
Broadcast time: 00:50–06:00, 07:20–10:20, 13:00, 17:00, 23:00.

12.3.10 Kyodo News Agency, Japan

Ident. Frequency lpm IOC Power Note

JJC 4,316.0 kHz 60 576 5 kW meteo. maps 120 lpm
JJC 8,467.5 kHz 60 576 10 kW meteo. maps 120 lpm
JJC 12,745.5 kHz 60 576 15 kW meteo. maps 120 lpm
JJC 16,971.0 kHz 60 576 15 kW meteo. maps 120 lpm
JJC 17,069.6 kHz 60 576 15 kW meteo. maps 120 lpm
JJC 22,542.0 kHz 60 576 15 kW meteo. maps 120 lpm

150
List of professional stations 12

Broadcast time: 01:45–07:45, 11:00–11:30, 13:35–22:15.

12.3.11 Kyodo News Agency, Singapore

Ident. Frequency lpm IOC Power Note

9VF/252 16,035.0 kHz 60 576 10 kW 07:40–10:10, 14:15–18:15
9VF/252 17,430.0 kHz 60 576 10 kW 07:40–10:10, 14:15–18:15
Broadcast time: 07:40–10:10, 14:15–18:15.

12.3.12 Northwood, Persian Gulf Base

Ident. Frequency lpm IOC Power Note

GYA 6,834.0 kHz 120 576 10 kW 18:00–08:00
GYA 14,436.0 kHz 120 576 10 kW nepřetržitý provoz
GYA 18,261.0 kHz 120 576 10 kW 08:00–18:00

12.4 South America

12.4.1 Rio de Janeiro, Brazil

Ident. Frequency lpm IOC Power Note

PWZ-33 12,665.0 kHz 120 576 1 kW
PWZ-33 16,978.0 kHz 120 576 1 kW
Broadcast time: 07:45–08:50.16:30–17:35.

12.4.2 Valparaiso Playa Ancha, Chile

Ident. Frequency lpm IOC Power Note

CBV 4,228.0 kHz 120 576 1 kW
CBV 8,677.0 kHz 120 576 1 kW
CBV 17,146.4 kHz 120 576 1 kW
Broadcast time: 11:15–23:25.

151
12 List of professional stations

12.5 North America

12.5.1 Halifax, Nova Scotia Canada

Ident. Frequency lpm IOC Power Note

CFH 122.5 kHz 120 576 10 kW
CFH 4,271.0 kHz 120 576 6 kW
CFH 6,496.4 kHz 120 576 6 kW
CFH 10,536.0 kHz 120 576 6 kW
CFH 13,510.0 kHz 120 576 6 kW
Broadcast time: continuous.

12.5.2 Iqaluit, NWT Canada

Ident. Frequency lpm IOC Power Note

VFF 3,253.0 kHz 120 576 5 kW 21:00–23:30
VFF 7,710.0 kHz 120 576 5 kW 00:10–09:00
Broadcast time: since middle of June till the end of November.

12.5.3 Resolute, NWT Canada

Ident. Frequency lpm IOC Power Note

VFR 3,253.0 kHz 120 576 5 kW 00:10–09:00
VFR 7,710.0 kHz 120 576 5 kW 21:00–23:30
Broadcast time: since middle of June till the end of November.

12.5.4 Sydney, Nova Scotia Kanada

Ident. Frequency lpm IOC Power Note

VCO 4,416.0 kHz 120 576 11:21, 11:42, 17:41
VCO 6,915.0 kHz 120 576 22:00, 23:31
Broadcast time: by frequency.

152
List of professional stations 12

12.5.5 Kodiak, Alaska USA

Ident. Frequency lpm IOC Power Note

NOJ 2,054.0 kHz 120 576 7.5 kW
NOJ 4,298.0 kHz 120 576 7.5 kW
NOJ 8,459.0 kHz 120 576 7.5 kW
NOJ 12,412.5 kHz 120 576 7.5 kW
Broadcast time: 04:00–11:59, 16:00-00:18.

12.5.6 Pt. Reyes, California USA

Ident. Frequency lpm IOC Power Note

NMC 4,346.0 kHz 120 576 4 kW 01:40–16:08
NMC 8,682.0 kHz 120 576 4 kW
NMC 12,786.0 kHz 120 576 4 kW
NMC 17,151.2 kHz 120 576 4 kW
NMC 22,527.0 kHz 120 576 4 kW 18:40–23:56
Broadcast time: continuous.

12.5.7 New Orleans, Louisiana USA

Ident. Frequency lpm IOC Power Note

NMG 4,317.9 kHz 120 576 4 kW
NMG 8,503.9 kHz 120 576 4 kW
NMG 12,789.9 kHz 120 576 4 kW
NMG 17,146.4 kHz 120 576 4 kW 12:00–20:45
Broadcast time: 00:00–08:45, 12:00–20:45.

12.5.8 Boston, Massachusetts USA

Ident. Frequency lpm IOC Power Note

NMF 4,235.0 kHz 120 576 4 kW 02:30–10:28
NMF 6,340.5 kHz 120 576 4 kW
NMF 9,110.0 kHz 120 576 4 kW
NMF 12,750.0 kHz 120 576 4 kW 14:00–22:28

153
12 List of professional stations

Broadcast time: 02:30–10:28, 14:00–22:28.

12.5.9 Inuvik, Canada

Ident. Frequency lpm IOC Power Note

VFA 8,457.8 kHz 120 576 1 kW 02:00, 16:30
Broadcast time: 02:00, 16:30.

12.6 Australia and Oceania

12.6.1 Charleville, Australia

Ident. Frequency lpm IOC Power Note

VMC 2,628.0 kHz 120 576 1 kW 09:00–19:00
VMC 5,100.0 kHz 120 576 1 kW
VMC 11,030.0 kHz 120 576 1 kW
VMC 13,920.0 kHz 120 576 1 kW
VMC 20,469.0 kHz 120 576 1 kW 19:00–09:00
Broadcast time: continuous.

12.6.2 Wiluna, Australia

Ident. Frequency lpm IOC Power Note

VMW 5,755.0 kHz 120 576 1 kW 11:00–21:00
VMW 7,535.0 kHz 120 576 1 kW
VMW 10,555.0 kHz 120 576 1 kW
VMW 15,615.0 kHz 120 576 1 kW
VMW 18,060.0 kHz 120 576 1 kW 21:00–11:00
Broadcast time: continuous.

154
List of professional stations 12

12.6.3 Wellington, New Zealand

Ident. Frequency lpm IOC Power Note

ZKLF 3,247.4 kHz 120 576 5 kW 09:45–17:00
ZKLF 5,807.0 kHz 120 576 5 kW
ZKLF 9,459.0 kHz 120 576 5 kW
ZKLF 13,550.5 kHz 120 576 5 kW
ZKLF 16,340.1 kHz 120 576 5 kW 21:45–05:00
Broadcast time: 00:00–04:00, 09:00–16:00, 21:00–23:00 (beginning every whole hour).

12.6.4 Honolulu, Hawaii USA

Ident. Frequency lpm IOC Power Note

KVM70 9,982.5 kHz 120 576 5 kW 05:19–15:56
KVM70 11,090.0 kHz 120 576 5 kW
KVM70 16,135.0 kHz 120 576 5 kW 17:19–03:56
Broadcast time: continuous, by frequency.

155
12 List of professional stations

12.7 List by frequency

Freq. Ident. Station lpm IOC Note
kHz Call sign
122.5 CFH Halifax, Nova Scotia Canada 120 576
148.0 Pevek, Chukotka peninsula 90 576
2,054.0 NOJ Kodiak, Alaska USA 120 576
2,618.5 GYA Northwood, The United Kingdom 120 576 20:00–06:00
2,628.0 VMC Charleville, Australia 120 576 09:00–19:00
3,241.0 BDF Shanghai, China 120 576
3,247.4 ZKLF Wellington, New Zealand 120 576 09:45–17:00
3,253.0 VFF Iqaluit, NWT Canada 120 576 21:00–23:30
3,253.0 VFR Resolute, NWT Canada 120 576 00:10–09:00
3,585.0 HLL1 Seoul, Republic of Korea 120 576
3,622.5 JMH Tokyo, Japan 120 576
3,830.0 Moscow, Russia 90, 120 576 IOC 288
3,855.0 DDH3 Hamburg/Pinnenberg, Germany 120 576
4,228.0 CBV Valparaiso Playa Ancha, Chile 120 576
4,235.0 NMF Boston, Massachusetts USA 120 576 02:30–10:28
4,271.0 CFH Halifax, Nova Scotia Canada 120 576
4,298.0 NOJ Kodiak, Alaska USA 120 576
4,316.0 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm
4,317.9 NMG New Orleans, Louisiana USA 120 576
4,346.0 NMC Pt. Reyes, California USA 120 576 01:40–16:08
4,416.0 VCO Sydney, Nova Scotia Kanada 120 576 11:21, 11:42, 17:41
4,481.0 SVJ4 Athens, Greece 120 576 08:45–10:44
4,610.0 GYA Northwood, The United Kingdom 120 576
4,616.0 BMF Taipei, China 120 576
4,777.5 IMB51 Roma, Italy 120 576
5,008.0 Moscow, Russia 90, 120 576 IOC 288
5,100.0 BDF Shanghai, China 120 576
5,100.0 VMC Charleville, Australia 120 576
5,250.0 BMF Taipei, China 120 576
5,526.9 BAF6 Beijing, China 120 576

156
List of professional stations 12

Freq. Ident. Station lpm IOC Note

kHz Call sign
5,755.0 VMW Wiluna, Australia 120 576 11:00–21:00
5,807.0 ZKLF Wellington, New Zealand 120 576
5,857.5 HLL2 Seoul, Republic of Korea 120 576
6,340.5 NMF Boston, Massachusetts USA 120 576
6,496.4 CFH Halifax, Nova Scotia Canada 120 576
6,834.0 GYA Northwood, Persian Gulf Base 120 576 18:00–08:00
6,915.0 VCO Sydney, Nova Scotia Kanada 120 576 22:00, 23:31
6,987.0 Moscow, Russia 90, 120 576 IOC 288
7,305.0 JMH2 Tokyo, Japan 120 576
7,396.8 HSW64 Bangkok, Thailand 120 576
7,420.0 BDF Shanghai, China 120 576
7,433.5 HLL3 Seoul, Republic of Korea 120 576
7,535.0 VMW Wiluna, Australia 120 576
7,695.0 Moscow, Russia 90, 120 576 IOC 288
7,710.0 VFF Iqaluit, NWT Canada 120 576 00:10–09:00
7,710.0 VFR Resolute, NWT Canada 120 576 21:00–23:30
7,880.0 DDK3 Hamburg/Pinnenberg, Germany 120 576
8,040.0 GYA Northwood, The United Kingdom 120 576
8,105.0 SVJ4 Athens, Greece 120 576 08:45–10:44
8,121.9 BAF36 Beijing, China 120 576
8,140.0 BMF Taipei, China 120 576
8,146.6 IMB55 Roma, Italy 120 576
8,457.8 VFA Inuvik, Canada 120 576 02:00, 16:30
8,459.0 NOJ Kodiak, Alaska USA 120 576
8,461.9 3SD Beijing, China 120 576
8,467.5 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm
8,503.9 NMG New Orleans, Louisiana USA 120 576
8,677.0 CBV Valparaiso Playa Ancha, Chile 120 576
8,682.0 NMC Pt. Reyes, California USA 120 576
9,110.0 NMF Boston, Massachusetts USA 120 576
9,165.0 HLL4 Seoul, Republic of Korea 120 576

157
12 List of professional stations

Freq. Ident. Station lpm IOC Note

kHz Call sign
9,459.0 ZKLF Wellington, New Zealand 120 576
9,982.5 KVM70 Honolulu, Hawaii USA 120 576 05:19–15:56
10,116.9 BAF4 Beijing, China 120 576
10,536.0 CFH Halifax, Nova Scotia Canada 120 576
10,555.0 VMW Wiluna, Australia 120 576
10,980.0 RCC76 Moscow, Russia 90, 120 576 IOC 288
10,980.0 RCC76 Moscow, Russia 90, 120 576 IOC 288
11,030.0 VMC Charleville, Australia 120 576
11,086.5 GYA Northwood, The United Kingdom 120 576 06:00–20:00
11,090.0 KVM70 Honolulu, Hawaii USA 120 576
11,420.0 BDF Shanghai, China 120 576
12,412.5 NOJ Kodiak, Alaska USA 120 576
12,745.5 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm
12,750.0 NMF Boston, Massachusetts USA 120 576 14:00–22:28
12,786.0 NMC Pt. Reyes, California USA 120 576
12,789.9 NMG New Orleans, Louisiana USA 120 576
12,831.9 3SD Beijing, China 120 576
13,510.0 CFH Halifax, Nova Scotia Canada 120 576
13,538.0 ZSJ Cape Naval, South Africa 120 576
13,550.5 ZKLF Wellington, New Zealand 120 576
13,570.0 HLL5 Seoul, Republic of Korea 120 576
13,597.0 JMH4 Tokyo, Japan 120 576
13,597.4 IMB56 Roma, Italy 120 576
13,882.5 DDK6 Hamburg/Pinnenberg, Germany 120 576
13,900.0 BMF Taipei, China 120 576
13,920.0 VMC Charleville, Australia 120 576
14,366.9 BAF8 Beijing, China 120 576
14,436.0 GYA Northwood, Persian Gulf Base 120 576 nepřetržitý provoz
14,842.0 ATP65 New Delhi, India 120 576 02:30–14:30
15,615.0 VMW Wiluna, Australia 120 576
16,025.9 BAF9 Beijing, China 120 576

158
List of professional stations 12

Freq. Ident. Station lpm IOC Note

kHz Call sign
16,035.0 9VF/252 Kyodo News Agency, Singapore 60 576 07:40–10:10, 14:15–18:15
16,135.0 KVM70 Honolulu, Hawaii USA 120 576 17:19–03:56
16,340.1 ZKLF Wellington, New Zealand 120 576 21:45–05:00
16,903.9 3SD Beijing, China 120 576
16,971.0 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm
17,069.6 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm
17,146.4 CBV Valparaiso Playa Ancha, Chile 120 576
17,146.4 NMG New Orleans, Louisiana USA 120 576 12:00–20:45
17,151.2 NMC Pt. Reyes, California USA 120 576
17,430.0 9VF/252 Kyodo News Agency, Singapore 60 576 07:40–10:10, 14:15–18:15
18,060.0 VMW Wiluna, Australia 120 576 21:00–11:00
18,236.9 BAF33 Beijing, China 120 576
18,261.0 GYA Northwood, Persian Gulf Base 120 576 08:00–18:00
18,560.0 BMF Taipei, China 120 576
18,940.0 BDG Shanghai, China 120 576
20,469.0 VMC Charleville, Australia 120 576 19:00–09:00
22,527.0 NMC Pt. Reyes, California USA 120 576 18:40–23:56
22,542.0 JJC Kyodo News Agency, Japan 60 576 meteo. maps 120 lpm

159
 13
Computer image
processing
This chapter focuses on a preparation of our broadcast contents – image and photo
editing before the transmission.
There are available many programs for image editing, from complex editors for
bitmap images to simple viewers with few editing functions. Some SSTV programs
have also image editing functions.
The selection of suitable program depends on a user and his needs and although
a control of programs can be slightly different the procedures described later are
so general that it would not be a problem to achieve results using your favorite
editor. The appropriate editor is generally any program for editing of raster/bitmap
images, such as: GIMP, Paint Shop Pro, Adobe Photoshop, Pixel32,… Less suitable
are vector image editors, such as: Corel Draw!, Inkscape, etc.
There are described some basic principles and functions, which are used in most
common programs. Although some functions below are shown in GIMP, see sec. ??,
it should not be problem for savvy user to find same functions in his favourite editor.
Later, see section ??, there are specific tutorials only for GIMP.

13.1 Image resizing

The SSTV uses a relative small resolution in comparison with images you can get
from digital cameras, scanners or some internet galleries. So before transmission the
image should be resized to conventional resolution 320×240 pixels. You can achieve
this in almost all SSTV programs, but images sometimes haven’t aspect ratio 4 : 3
or we want use only part of image. So it is useful to prepare your images before
QSO.
The image resizing brings a little risk, see image 13.1, original image is test chart
on page 171.

160
Computer image processing 13

Figure 13.1: Results of image

resizing with two different algorithms.

There is a result of two image scaling algorithms. The left image was scaled down by
nearest-neighbor interpolation, when pixels in regular columns and rows are removed.
But the linear interpolation was used for the right image. The difference is visible on
the first sight, the nearest-neighbor method caused distortion and some details are
lost completely, e.g. thin lines are lost and on originally smooth curves are stair-like
lines more evident. The linear interpolation is more considerate to details, but it
depends on amount of decreasing and sometimes there should arise a moiré, here on
the raster of changing white and black stripes.
Almost every image editing software give an option for used scaling method, so
choose linear interpolation (or bilinear or bicubic). A raster image for interpolation
must have at least 16bit color depth, on indexed images with 256 or less colors it
doesn’t work, but you can convert them to more colors temporary and then back.

13.2 Color adjustment

Basic tools for the color adjustment can edit image contrast, brightness, color satu-
ration and hue. The advanced tools are curves adjustment and image histogram.
The histogram is a representation of the distributions of colors in an image (see
pic. 13.2). It is bar graph and there is representation of the tonal variations on the
horizontal axis and the vertical axis represents the number of pixel in the particular
tone.
In digital photography, you can easily review exposure by histogram. If the image
is underexposed the amount of darker pixels lays on left side of graph (more darker
pixels), vice-versa for overexposed images there is high number of lighter pixels on
right side. It is possible to check exposure in the digital camera menu and take
snapshot again, of course unless the image wasn’t dark or light in principle.
The ideal case is if the luminance is wide-spread inside the luminance range. For
low contrast the luminance values are spread only over smaller range of luminace.

161
13 Computer image processing

3500

3000

2500

Number of pixels
2000

1500

1000

500

0
0 50 100 150 200 250
Pixel intensity
Figure 13.2: The image histogram for 256 brigtness values.

0 50 100 150 200 250 0 50 100 150 200 250

Figure 13.3: Positive influence of

histogram equalization on image contrast.

The histogram equalization can help to adjust contrast of image – luminace is spread
over all range. For the result of histogram equalization see pic. 13.3.

162
Computer image processing 13

In the GIMP the tool for histogram adjustment is in menu Colors → Levels. In
case of low contrast image in pic. 13.3 the input levels were set on edges of growing
values, but the GIMP can make also automatic equalization with dialog 13.4.

Figure 13.4:

The second important tool is curve adjustment, its the one of most important tools
in professional graphics editors. It provides the complex settings of color tone,
brightness and contrast. The GIMP tool Adjust Color Curves (pic. 13.4) is in menu
Colors → Curves. Its functionality is very similar across image editing software.
The work with curves is quite simple. The tool window has on the 𝑥 axis displayed
input luminance values and on the 𝑦 axis output values, which are presented as
gray-scale. The curve self is the function 𝑦 = 𝑓(𝑥) and it transforms input values
to output and by its shaping the user sets the parameters of the transformation.
At the beginning the transformation function is 𝑦 = 𝑥 and therefore the output
level is equal to input. When you touch the curve by cursor a control point appears
on the curve. You can move this point in every direction and if the option Preview
is switched on, then the results will be shown immediately. It is possible to add
more control points for continuous transformation or you can set any curve shape
with discontinuous transformation.
What you can do with it? The results of transformations and the used curves is in
pic. 13.5. You can spread input values by histogram and improve contrast – 13.5A.
Another use is to darken or lighten the image, but there is many more options
how to do this then a simple setting of levels. With the curve you can focus on a

163
13 Computer image processing

range of colors, can edit only midrange portion of color scale and keep black and
light pixels, if you hold curve in the middle and move the control point upwards and
if you move the point more to the left or right you can change brightness of darker
or lighter tones. In example 13.5B are lighter tones made more brighter and the
lower end of curve is moved to the right, so darker pixel are more darker and the
image is more contrasted. For image darkening move the point downwards, it’s in
example 13.5C, also there is moved the right edge of the curve so lighter pixel are
more brighter now.
For changing contrast use the curve in 13.5D, there are two control points. The
curve with “S” shape makes dark tones darker and light tones brighter.
In addition to basic editing the curve can be used for a variety of effects, see
example 13.5. The shape is little crazy and result image lost real tones. Notice,
that there is inversion of dark tones in left part of the curve.
The curve and histogram adjustment is possible to do for image luminance and
for each color component too, so it is possible to change color perception of an image.

13.3 Filters
Images can be further improved using various filters. They can serve different pur-
poses, mainly for smoothing and noise reduction, increasing sharpness, edge detec-
tion, unsharp masking, etc.
Digital image itself is a discrete two-dimensional (2D) signal, which is character-
ized by its frequency spectrum, whose components similar to the 1D signal can be
determined using the discrete Fourier transform. So it’s possible to modify image
by using low or high pass filters or highlight certain components to affect the final
look.
For specific case, we can return to chapter 3.2 (page 15), there is described how
the SSTV transmission is affected by bandwidth. The limited bandwidth caused
the distortion of fine grid in resolution test pattern (pic. 3.3). Fine grid and sharp
gradients in the image represent a high frequency components and limited bandwidth
acted as a bandpass filter, which suppressed the higher frequencies. Also, an image
noise is represented by high frequency components and the low-pass filter can be
used for its reduction. Vice-versa, a high-pass filter is designed to highlight the edges
and sharpening.

13.3.1 Convolution matrix

As one of the methods of linear filters is used the higher mathematic procedure –
discrete convolution. A way the image is processed is given by convolution kernel,
which is given by matrix, typically with dimensions 3 × 3 or 5 × 5, but also others.

164
Computer image processing 13

A.

B.

C.

D.

E.

165

Figure 13.5:
13 Computer image processing

Source Image Pixel values Output pixel value

Convolution
kernel

Figure 13.6: Computation of discrete convolution.

The kernel moves pixel by pixel along lines and columns of the source image and
computes pixel values, see fig. 13.6. It moves from first pixel on the position 0,0,
then 0,1, etc. For each pixel is from neighborhood pixels (their numerical luminance
values) multiplied by kernel values, added together and resulting value is stored in
target image on same position as source. Then the matrix moves to next pixel and
computation is repeated.
The target image changes in way defined by values of convolution kernel, e.g.
matrix 𝐇𝟐 in fig. 13.6 use pixel and its neighborhood of 3 × 3 size and computes
weighted mean from these nine pixel with kernel:
1
1 2 1 (1 ⋅ 50 + 2 ⋅ 75 + 1 ⋅ 74+
𝟏 ⎛ 16
𝐇𝟐 = ⎜2 4 2⎞
⎟, = 𝟔𝟗
𝟏𝟔 +2 ⋅ 47 + 4 ⋅ 76 + 2 ⋅ 75+
⎝1 2 1⎠
+1 ⋅ 51 + 2 ⋅ 76 + 1 ⋅ 74)

The result is an image smoothing and noise reduction. Other matrices are used to
increase sharpness, edge detection or to create a relief.
In the GIMP the filter is in menu Filters → Generic → Convolution Matrix.
The important parameters are also divisor, which adjusts the values in the matrix
and offse is used to adjust the colors. See fig. 13.7, how to set matrix values
and parameters. Furthermore you can choose the color channels and behaviour of

166
Computer image processing 13

calculation on border pixels, it is also possible to determine the divisor and offset
automatically.

Figure 13.7: Convolution matrix parameters in GIMP.

13.3.2 Noise reduction

Thanks to noise reduction it is possible to improve partially received SSTV or WE-
FAX images by smoothing. It can be done several ways. Since SSTV transmission
takes a relatively long time the noise levels are often uneven, affected by random
interference and transmission conditions, so you can by tool Selection mark affected
areas, where apply the filter.

13.3.2.1 Spatial average filtering

The simplest way to reduce noise is average filter with convolution kernel 𝐇𝟏 .

1 1 1
𝟏⎛
𝐇𝟏 = ⎜ 1 1 1⎞
⎟
𝟗
⎝1 1 1⎠

167
13 Computer image processing

The next useful matrices for averaging are 𝐇𝟐 and 𝐇𝟑 , they aren’t using neighbor
pixel with same weight, but near pixels have greater coefficients then farther. The
values of the coefficients are based on binomial series.

1 4 6 4 1
1 2 1 ⎛ 4 16 24 16 4⎞
𝟏 ⎛ 𝟏 ⎜⎜ ⎟
⎟
𝐇𝟐 = ⎜2 4 2⎞
⎟ 𝐇𝟑 = ⎜ 6 24 36 24 6⎟
𝟏𝟔 𝟐𝟓𝟔 ⎜
⎜
⎜4 16
⎟
⎟
⎝1 2 1⎠ 24 16 4⎟
⎝1 4 6 4 1⎠

The disadvantage of averaging filters is that smoothing distorts sharp color gradients
and due to it thin lines and other details are distorted.

Received image Averaging 𝐇𝟏

Binom. filter 3 × 3𝐇𝟐 Binom. filter 5 × 5𝐇𝟑

Figure 13.8:

Next filters for noise reduction in GIMP are in menu Filters → Blur →…. E.g.
Gaussian blur is useful for SSTV noise reduction, the user can set a radius of a blur.
For image preparation can be used Selective Gaussian Blur, it can filter continuous
areas and edges leaves unaffected, but for received images is not suitable.

168
Computer image processing 13

13.3.2.2 Median filter

The term median is the middle value in list of numbers, here luminance values. The
image filter works with pixel neighborhood, e.g. with 3 × 3 size, then sort their
values and takes the median, i.e. middle value of the sequence:
21 22 20 Pixels: 21, 22, 20, 21, 19, 231, 19, 21, 20
21 19 231 Sorted: 19, 19, 20, 20, 21, 21, 21, 22, 231
19 21 20 Average ≐ 43; Median = 21
Table 13.1: none
We see that if we use the averaging, then significantly skewed value 231, which
can be caused by noise, greatly affects the result. In contrast, the median is not
influenced by extreme deviations. The disadvantage of median filtering is that it
sometimes distorts the thin lines and sharp gradient in the image.

Received image Binom. filter 𝐇𝟐

Median 3 × 3 Median 5 × 5
Figure 13.9: The results of noise reduction with median filter.
In GIMP the median filter is in Filters → Enhance → Despeckle. There are several
options for filter settings. The radius from 1 for 3 × 3 neighborhood to 20 for 41 × 41

169
13 Computer image processing

and the black level (−1–255) and white (0–256). The pixels darker or brighter than
these two levels will be removed, filter passes every level for the extreme values
−1 and 256. Moreover, it is still possible to select adaptive median, when GIMP
tries to determine optimal radius for given image location automatically (the user
parameters are ignored).

13.3.3 Sharpening
Often happens that images from cameras or scanners have reduced sharp edges. In
this case you can try to use sharpening filters, for edges highlighting. There are two
convolution kernels for sharpening 𝐇𝟒 and 𝐇𝟓 with an even stronger effect.

0 −1 0 −1 −1 −1
𝐇𝟒 = ⎛
⎜−1 5 − 1⎞
⎟ 𝐇𝟓 = ⎛
⎜−1 9 − 1⎞
⎟
⎝0 −1 0 ⎠ ⎝−1 −1 − 1⎠

In the GIMP the sharpening filters are in the menu Filters → Enhance →…. The
first one is Sharpen, and it has similar results as kernels above and it is possible to
set sharpening level and from preview select the best results.
The disadvantage of these filters is that, it also highlights the image noise and
some other unwanted details. It can be seen in fig. 13.10 when the filter 𝐇𝟓 is used,
so there appears teeth on vertical edges caused by camera interlacing. In this case it’s
possible to use filter called unsharp mask. The unsharp mask procedure in first step
apply Gaussian blur on image copy and then check the difference between blurred
and original image. When the difference is greater then a user-defined threshold,
then both images are subtracted and the result is added to original image. The
threshold limits the sharpness of output image, so small details of certain size is not
sharpened. Thanks to this the sharpening does not apply to noise and graininess.
The digital unsharp mask is great filter for enhancing sharpness. Some results are
influenced by these three parameters:

⊳ Amount indicates how much will increase the contrast of edges, by the edge
lightens or darkens, this setting will affect most the degree of sharpening.
⊳ Radius sets how much pixels around edges will be used. For smaller radius the
filter affect also smaller details, but for greater it may produce a bright rim
around edges.
⊳ Threshold controls the minimum brightness change that will be sharpened. Thanks
to threshold it is possible to enhance only stronger edges and finer leave un-
changed.

170
Computer image processing 13

No wonder, that unsharp mask is considered as king in improving the image. The
only drawback is that if you blow over it, then around edges occurs distinctive bright
rim and the image looks unnaturally.

Original image Sharpening 𝐇𝟒

from CCD camera

Sharpening 𝐇𝟓 radius: 2.5; amount:

1.5; threshold: 0
Figure 13.10: Results of image sharpening.

171
13 Computer image processing

Figure 13.11: none

172
 References
[1] Adámek, J.: Kódování. SNTL 1989.
[2] Abrams, Clay K6AEP; Taggart, Ralph WB8DQT: Color Computer SSTV. 73, Nov., Dec.
1984.
[3] American Radio Relay League: The ARRL Handbook for Radio Communications. ARRL
2011.
[4] Barber, James L. N7CXI: Proposal for SSTV Mode Specifications. Dayton SSTV forum,
2000.
[5] Bodson, Dennis W4PWF; Karty, Steven N5SK: FAX480 and SSTV Interfaces and Software.
QST, Jul. 1998.
[6] Bruchanov, Martin OK2MNM: Ještě k SSTV. AMA magazín, červenec 1997.
[7] Cameroni, Giuseppe I2CAB; Morellato, Giancarla I2AED: Get on SSTV – with the C-64.
Ham radio, Oct. 1986.
[8] Churchfield, Terry K3HKR: Amiga AVT System. 73 Amateur Radio, Jul. 1989.
[9] Cordesses, Lionel; Cordesses, Roland F2DC: Some Thoughts on “Real-Time” SSTV Process-
ing. QEX, No. 218, May, June 2003.
[10] Cook, S. A. G., G5XB: Hellscheiber, What is it and how it works. Radio Communication,
Apr. 1981.
[11] Dewitt, H. William, W2DD: The Robot Research Inc. Model 1200C Color Slow Scan Con-
verter. CQ, Jun. 1988.
[12] ETSI Standard: Digital Radio Mondiale (DRM); System Specification. 2005 [ETSI ES 201
980 V2.2.1].
[13] Dobeš, Josef; Žalud, Václav: Moderní radiotechnika. BEN technická literatura 2006.
[14] Evers, Hans PA0CX: The Hellscheiber Rediscovery. Ham Radio, Dec. 1979.
[15] Fanti, Franco: Facsimile. CQ-TV, vol. 88, 1974.
[16] Fingerhut, K., Půža, V.: Amatérská televize. Svazarm 1983.
[17] Frejlach, Karel: Příjem signálu SSTV mikropočítačem ZX Spectrum. Radioamatérský
zpravodaj, vol. 4, 9, 1989.
[18] Frejlach, Karel: Digitální radioamatérský provoz. Autor vlastním nákladem 1998.
[19] Ford, Steve WB8IMY: TV on 10. QST, Apr. 2001.
[20] Furukawa, Bruce T.; Murray, Thomas L.; McGee Kenneth A.: Video Surveillance of Active
Volcanoes Using Slow-Scan Television. USGS Bulletin, 1992.
[21] Geier, Michael Jay KB1UM: The SSTV Explorer. Radio Fun, vol. 6, 1995a.
[22] Geier, Michael Jay KB1UM: Pasokon TV Slow-Scan TV Interface. 73 Amateur Radio
Today, July 1995b.
[23] Glanc, Antonín OK1GW: Amatérská televize. Amatérské rádio, vol. 6, 7, 8, 1971.
[24] Glidden, Ramon L. W5NOO: Getting Started with Slow Scan Television. QST, Sep. 1997.
[25] GNU Image Manipulation Program – User Manual. .
[26] Gola, Miroslav ing. OK2UGS: Přijímač faksimile v pásmu KV (40). Praktická elektronika
ARadio, vol. 4, 5, 1999.
[27] Goodman, Dick WA3USG: SSTV with the Robot 1200C Scan Converter and the Martin
Emmerson EPROM Version 4.0. 73 Amateur Radio Today, Jul. 1991.
[28] Harlan, Gene WB9MMM: Slow-Scan TV with the Sound Blaster. QEX, May 1993.
[29] Hioki, Takashi JF1GUQ: Pasokon TV. Mobile Ham Amateur Radio Magazine, Feb. 1993.

173
[30] Hubeňák, Jiří RNDr.: Zařízení pro příjem faksimile počítačem PC AT. Amatérské Rádio,
vol. 6, 1994.
[31] Jordan, Karel Ing.: Radioamatérské družice. Ústřední rada radioklubu Svazarmu 1983.
[32] Karmasin, Karel OK2FD: SSTV a PC. AMA magazín, duben 1997.
[33] Langner, John WB2OSZ: SSTV – The AVT System Secrets Revealed. CQ-TV, vol. 149,
1990.
[34] Langner, John WB2OSZ: Slow-Scan Television – It isn’t expensive anymore. QST, Jan.
1993.
[35] Langner, John WB2OSZ: Color SSTV for the Atari ST. 73 Amateur Radio, Dec., Jan. 1989,
1990.
[36] Langner, John WB2OSZ: Slow-Scan TV Questions and Answers. Radio Fun, Feb. 1995.
[37] Montalbano, John KA2PYJ: The ViewPort VGA Color SSTV System. 73, Aug. 1992.
[38] Nozdrovicky L.: Základy televízie. Slovenské vydavateľstvo technickej literatúry 1962.
[39] Null Linda, Lobur Julia: The Essentials of Computer Organization and Architecture. Jones
and Bartlett Publishers 2003.
[40] : Nové směry v SSTV. Amatérské rádio, vol. 7, 8, 1986.
[41] Pagel, Paul N1FB: Radioware SSTV Explorer. QST, Apr. 1994.
[42] Pitt, Doug: Basic NBTV Techniques. 1998.
[43] Prinz, Richard OE1RIB: Kamera-Shield auf Arduino-Basis für ein SSTV-Sendemodul. Funka-
mateur, vol. 10, 2012.
[44] Prosise, Jeff: Bitmapové soubory – pohled dovnitř. PC Magazine, únor 1997.
[45] Ramon, Carine ON7LX: HF Managers Handbook. IARU 2003.
[46] Rensen, Marius: HF FAX Station List. , 2002.
[47] Schick, Martin K. KA4IWG: Color SSTV and the Atari Computer. QST, Aug. 1985.
[48] Seger, Jiří: Jak se lidé dorozumívali. Albatros 1987.
[49] Taggart, Dr. Ralph E. WB8DQT: A New Standard for Amateur Radio Fascimile. QST,
Feb. 1993.
[50] Taggart, Dr. Ralph E. WB8DQT: Digital Slow-Scan Television. QST, Feb. 2004.
[51] Taggart, Dr. Ralph E. WB8DQT: The ARRL Image Communications Handbook. ARRL
2002.
[52] Taggart, Dr. Ralph E. WB8DQT: The Romscanner. QST, Mar. 1986.
[53] U.S. Department Of Commerce: Worldwide Marine Radiofacsimile Broadcast Schedules.
NOAA 2010.
[54] Vester, Ben K3BC: An Inexpensive SSTV System. QST, Jan. 1994a.
[55] Vester, Ben K3BC: An Inexpensive SSTV System continues to Grow. QST, Dec. 1994b.
[56] Vít, V.: Televizní technika – barevné přenosové soustavy. BEN technická literatura 1997.
[57] Vlček, K.: Komprese a kódová zabezpečení v mutimediálních komunikacích. BEN technická
literatura 2004.
[58] Vyhláška FMS č. 390/1992 Sb. o povolováníamatérských vysílacích stanic. Ministerstvo
spojů 1992.
[59] Wolfgang, Larry WR1B: TASCO Electronics TSC-70U Slow-Scan TV System. QST, Apr.
1997.
[61] Žára Jiří, Bedřich Beneš, Felkel Petr: Moderní počítačová grafika. Computer Press 1998.

174