DIGITAL TERRESTRIAL TELEVISION

Field evaluations in Taiwan of the

DVB-T COFDM
ATSC 8-VSB and
digital TV systems
Chi-Fang Huang, The Nan Chang and Chau-Yun Hsu
Graduate Institute of Communication Engineering, Tatung University, Taiwan

This article presents the results of field evaluations carried out in Taiwan on the two
principal digital TV broadcasting systems in use today ATSC 8-VSB (developed in
North America) and DVB-T COFDM (developed in Europe). The evaluations were
carried out in February and March 2001 in the Northern area and the Taipei
metropolis of Taiwan.
Based on the subsequent report (which forms the basis for this article), the Taiwan
telecommunications authority agreed to deregulate DTV broadcasting and adopt the
DVB-T transmission standard. DVB-T is now being used by the five terrestrial
television broadcasters in Taiwan.

With most segments of the electronics industry going digital in the late 1990s, the Taiwan government
became very aggressive in promoting the digitalization of terrestrial television broadcasting on the island. The
governments views were that the promotion of digital terrestrial television (DTT) would bring considerable
economic strength to Taiwan. Thus, in 1998, it announced that the North American ATSC transmission stand-
ard would be adopted, and it scheduled the start of engineering tests for July 1999. Furthermore, the govern-
ment declared that all DTT stations should be in service by December 2001. The existing analogue NTSC
transmissions should co-exist with the new digital transmissions for a further five years and, when the
expected digital coverage had reached 85% of the population, the analogue transmission licences should be
terminated.
At the time, the countries that were proposing to adopt the ATSC 8-VSB system were the USA, Canada, South
Korea and Argentina [1].
In 1999, Sinclair Broadcasting demonstrated COFDM and 8-VSB reception to broadcasters in Baltimore
(USA). These demonstrations raised deep concerns worldwide about the performance of the ATSC 8-VSB
system when compared with the European DVB-T COFDM system:
! COFDM was found to be more capable of handling the multipath signal propagation that often occurs in
city (and hilly or mountainous) areas;
! The DVB-T system worked very well with a simple indoor antenna, unlike the ATSC system which often
required an external directional yagi antenna.

EBU TECHNICAL REVIEW July 2003 1 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
! Mobile reception which had been seen by the broadcasting companies as a new business opportunity,
especially when data broadcasting is added was considered to be a viable option of the DVB-T system.
This point is very important for those countries, such as Taiwan, where cable TV penetration is dominant.

The Taiwanese evaluations of ATSC and DVB-T

A Taiwanese project sponsored by The Television Academy of Arts & Sciences of the Republic of China
was set up to carry out local evaluations of the ATSC 8-VSB and DVB-T COFDM systems [2]. The areas
chosen for these measurement tests were Northern Taiwan and the Taipei Metropolis. And the reception
modes to be considered were:
1) outdoor fixed-point reception (at 102 receiving locations);
2) outdoor mobile reception (along twelve highways and streets);
3) indoor fixed-point reception (at 103 receiving locations inside seven buildings).
The data collected from these locations and routes was considered to be adequate to satisfy the projects aims.
In accordance with the existing NTSC spectrum allocation in Taiwan, the channel width for the digital tests
was set at 6 MHz for both systems. The modulation parameters used for the evaluations are shown in Table 1:

Table 1
Modulation parameters to be measured

Reception mode ATSC DVB-T

Outdoor fixed-point reception 8-VSB QPSK:64-QAM / 8k FFT

Outdoor mobile reception 8-VSB QPSK:16-QAM / 2k FFT

Indoor fixed-point reception 8-VSB QPSK:64-QAM / 8k FFT

The transmitting antenna was installed at the CTS (Chinese Television System) station on the Yang-Ming-
Shan mountain (see Fig. 1 left). It had the following characteristics:
! RFS dipole array;
! Frequency = 593 MHz;

E / Emax
Station : Taiwan 0 Date : 5/26/00
1.0
0.9
0.8
0.7

0.6
0.5
0.4

0.3

0.2

0.1
270 90

Model P
: HP18(594) 180
Face XOffset YOffset Tilt Power Phase Polarization : Horizontal
A 0.00 0.28 0.0 0.0 0.0 Frequency (MHz) 533.00
:
B 0.00 0.28 0.0 252.0 0.0 Directivity (dB) : 5.46
C 0.00 0.28 0.0 796.0 0.0 Loaded Measured Unit Pattern
D 0.00 0.28 0.0 163.0 -30.0 File = php-570.hup
Pattern Tolerance +/- 5% of Emax
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Er

- -20 -14 -10.5 -8 -6 -4.4 -3.1 -1.9 -0.9 0 dB

Voltage and Power Ratios
0 dB = Max ERP
Directivity 5.46 dB

Figure 1
RFS dipole array for DTV transmitter: (left) location and (right) simulated radiation pattern

EBU TECHNICAL REVIEW July 2003 2 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
! Power at the antenna input = 5 KW;
! Polarization = horizontal (HP).

The simulated radiation pattern offered by the RFS antenna is shown in Fig. 1 right. The 00 axis is directed
to the north.
The results of these evaluations were then presented to the Directorate General of Telecommunications, Tai-
wan representing the broadcasters proposal for DTT in the spirit of telecommunication deregulation. The
proposal was accepted and, in mid-2001, the DVB-T transmission standard for terrestrial television was
adopted by the five terrestrial broadcasters in Taiwan.

Outdoor fixed-point reception

Many reports have been published [1] on
the results of evaluation tests carried out Log Periodic
on DVB-T COFDM and ATSC 8-VSB in Antenna ATSC ATSC
200 MHz ~ 1 GHz Demodulator Decoder Monitor
a number of countries or areas. As a Zenith Zenith
result, experts have gained some under-
Noise
standing of the characteristic differences Trilithic DVB-T DVB-T

splitter
Generator Demodulator Demodulator Monitor
BPF
between these two systems. However, the NC6109 BT BT

evaluation project described here was a

locality-oriented one to test the potential Stand-By

performance differences between these

two systems in parts of Taiwan. The test Spectrum
analyzer
areas used included a mountainous region Tek 2715

as well as an urban residential area con-

taining not only small houses but high
apartment blocks as well.
Since this was a comparison test, the out-
put power of the transmitter was set at
Figure 2
5 KW. The reception measuring system is Schematic of the reception measuring system
shown in Fig. 2. This equipment was
installed in a Harries vehicle (see Fig. 3).
The receiving antenna (see Fig. 4 left) was a log-periodic, horizontally polarized (HP), covering the range
from 200 MHz to 1 GHz. Its radiation pattern is shown in Fig. 4 right, which was measured in the anechoic
chamber of Tatung University. The pattern shown was measured at 593 MHz using horizontal polarization.

Figure 3 Figure 4
The measurement vehicle (left) the log-periodic antenna that was used and (right) its measured
radiation pattern at 593 MHz, HP

EBU TECHNICAL REVIEW July 2003 3 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
When receiving the DTT signals, the antenna was raised to a height of 8.1m above ground level and was
rotated to find the direction of best reception. The antenna direction was duly noted. The incoming signals
were routed through a band-pass filter of 593 MHz centre frequency and 78 MHz bandwidth. The band-pass
filter was included to reject ambient background noise from the considerations. A noise generator was also
included in the antenna feed to simulate multipath interference in the surrounding environment. A measure-
ment parameter that we used during the tests is the Threshold of Visibility (S/N@TOV) which is explained
later. At each measurement point, background noise from the generator was added at the weakest level and
was then progressively increased to determine the S/N@TOV value.
Finally, the composite RF signal was fed into a splitter and divided further into four paths one for the ATSC
8-VSB demodulator, one for the DVB-T COFDM demodulator, one for a spectrum analyzer, and a spare feed.
The professional demodulators and decoders used by the two systems are listed below:

ATSC receiver:
Demodulator: Zenith DTVDEMOD-S
Decoder: Zenith DTVDECODER-H
DVB-T receiver:
Demodulator: BT DTVM2000(T)
Decoder: BT DTVD-200 [BT: Broadcasting Technology Ltd.]

Video monitors were used for displaying the broadcast programmes and for evaluating the picture quality
using the CCIR 5-point grading scale (0 4). If the DTT signals were good, they produced traces on the spec-
trum analyzer as shown in Fig. 5.

(a) (b)
Figure 5
Spectrum trace under good reception conditions: (a) ATSC spectrum; (b) DVB-T spectrum

Figure 6 Figure 7
Distribution of the measurement points Distribution of the measurement points
near Tapei city

EBU TECHNICAL REVIEW July 2003 4 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
The spectrum analyzer was controlled by a notebook PC through a PCMCIA interface. At each outdoor fixed
location, the measurements were carried out separately on the two systems: via a radio link, staff at the trans-
mitting station were instructed to switch between the two systems. In addition to providing the visual traces,
the spectrum analyzer calculated the field strength (E) of the incoming RF signal and its average power within
the channel.
A total of 102 locations were used for the outdoor fixed-point measurements. Geographically, the most south-
erly point was at Hsin-Pu and the most northeasterly point was at Pa-Du. The location coordinates were
recorded by a GPS receiver. Furthermore, these coordinates were fed into a software called MapInfo [3] to be
sprinkled onto a map as shown in Fig. 6. An enlarged view showing the measurement points near Taipei
city is given in Fig. 7.

Measured point: Lin An-Tai Old House Best receiving 5 ETN

direction: (see Abbreviations)

DTT system ATSC DVB-T (mode 1: 8k) Spectrum

E (dBV/m) 42.2 42.1 Upper: ATSC
Lower: DVB-T

Average power (dBm) -77.1 -77.2

S/N (dB) 15.5 21.68

S/N@TOV (dB) 13.6 18.92

Site Margin (dB) 1.9 2.76

Tap energy (dB) -17.1

SER 0

MER 21.66

BER 1.67E-02

CCIR Grading 4 4

Figure 8
Record table for the measurements made at each outdoor point

Referring to Fig. 8, which is a record table of the measurements made at each outdoor point, the items on it can
be explained as follows (the second entry of the first row is the location name):
a) The best receiving direction the fourth entry of the first row found by rotating the receiving antenna.
b) E field the field strength of the RF signal, defined as mentioned above
c) Average Power calculated from the spectrum analyzer by a software application.
d) S/N (dB) after demodulation, this is the Signal-to-Noise ratio of the decoder before decoding [4], and it
may be viewed as the ideal signal compared to everything else. Taking the 8-VSB system as an exam-
ple, the S/N may be defined as:

EBU TECHNICAL REVIEW July 2003 5 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION

1 N 2

N
Ij
j = 1
S / N = 20 log ( dB )
N
1 I 2j
N j =1

where,

Ij is the ideal in-phase or real axis symbol value, transmitted during the jth interval,

I j is the difference along the real axis between the ideal signal value and the value actually received
during the jth interval.
In the 8-VSB system, there are seven kinds of source errors or noise which can affect the S/N value.
They are: Frequency response error, Group delay error, Amplitude error, Phase error, Phase noise, Broad-
band noise and Software (DSP) noise. As a result of using different demodulation and decoding mecha-
nisms, the S/N values shown on Fig. 8 are different, even if the RF field strength is more or less the same.
e) S/N@TOV (dB) after receiving the best picture quality on the video monitor, the noise generator begins
to increase the noise level until the picture cannot be displayed normally any more; for example, mosaic
patterning or frame stopping begins to happen. At this point, reception reaches the so-called TOV
(Threshold of Visibility). At this moment, the S/N value of the decoder is referred to as S/N@TOV.
f) Site Margin (dB) value of the d term above (i.e. S/N) minus the value of the e term (S/N@TOV). It
represents the dynamic range of the systems ability to withstand environmental noise.
g) Tap energy this term and the following one are special parameters used in the ATSC 8-VSB system. Its
value is obtained by dividing the amplitude of the interfering multipath signal with the amplitude of the
main wanted signal. It is read out from a chart that is displayed by the decoder software. Hence, the
larger the value of the Tap energy, the stronger is the interfering multipath signal.
h) SER Segment Error Ratio.
i) MER Modulus Error Ratio. This term and the following one are special parameters used in the DVB-T
COFDM system.

(a)

CCIR = 4 CCIR = 3 1 CCIR = 0

(b)

CCIR = 4 CCIR = 3 1 CCIR = 0

Figure 9
Received spectra of (a) DVB-T programmes (b) ATSC programmes

EBU TECHNICAL REVIEW July 2003 6 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
j) BER Bit Error Ratio
k) CCIR Grading by definition, the perfect picture is classified with a grade of < 5 >. It is classified as
grade < 0 > when no picture at all is displayed. In our measurements, grade < 5 > was not adopted; a nor-
mal received picture was classified as grade < 4 >. Other abnormal pictures were given a grade between
< 1 > and < 3 >, as judged by the tester. Fig. 9 shows the spectrum traces for different CCIR grades in the
case of the two DTT systems under test.
The detailed results obtained at each location have been well documented in the projects final report [5].
Table 2 compares the statistical results of the two systems, based on their CCIR gradings.
Table 2
Statistical results of outdoor receptions at 102 locations

CCIR Grade ATSC DVB-T

4 53 58
3 5 4
2 5 2
1 6 8
0 33 30

Measured point: Crossing of Fu-Hsin Best receiving 330 ETN

N. Rd. and Ming- direction: (see Abbreviations)
Sheng E. Rd.

DTT system ATSC DVB-T (mode 1: 8k) Spectrum

E (dBV/m) 70.6 70.5 Upper: ATSC
[Amp: +20dB] [Amp: +20dB] Lower: DVB-T

Average power (dBm) -48.7 -48.8

S/N (dB) 3.4 21.4765

S/N@TOV (dB) N/A N/A

Site Margin (dB) N/A N/A

Tap energy (dB) -2.6

SER 12852

MER 21.468

BER 1.67E-02

CCIR Grading 0 1

Figure 10
An example of no reception in a high field-strength area

EBU TECHNICAL REVIEW July 2003 7 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
Over the whole area, DVB-T was observed to provide better reception than ATSC. There were only two
points where reception of ATSC was graded as < 4 > while DVB-T only attained a < 3 >. This occurred in low
field-strength areas. The measured field strengths, ATSC vs. DVB-T, were < E = 46.1 vs. 46.7 dBV/m > and
< E = 46.6 vs. 46.5 dBV/m >.
When all 102 reception points are considered, the lowest field strengths which yielded CCIR grades of < 4 >
were at the following locations:
! ATSC Lin An-Tai Old House [N 250 04 24; E 1210 31 18], Emin = 42.2 dBV/m;
! DVB-T Hsin-Hai roadside gate [N 250 01 22; E 1210 32 10] of Taiwan University,
Emin = 41.9 dBV/m.

Consequently, it is not easy to find a simple explanation for obtaining such a high subjective grading despite
the low field strengths (close to the Emin value). On the other hand, there were some locations where the field
strength was well above the Emin value yet, even with the assistance of a 20 dB gain RF amplifier, the pic-
tures still failed (the channel spectrum may have been seriously impaired). The record shown in Fig. 10 is an
example. This evidence demonstrates that a high field strength is not always able to guarantee successful
reception, and this is in contrast to traditional analogue TV broadcasting.
For the purpose of analyzing the capabilities of these two different transmission systems in the presence of
multipath signals, the term Site Margin is used. As a whole, the value of this term in DVB-T reception is
larger than that in ATSC. For instance, at the Sung-Shan airport location [N 250 03 51; E 1210 32 39], the
Site Margin values were 8.47 dB vs. 1.7 dB. Obviously, the COFDM functions very well in this situation.
It must also be noted that many of the locations tested did not yield usable pictures, mainly due to building
clutter in the urban areas: these measurement points were graded < 0 >.

Outdoor mobile reception

A key factor we considered was the capabil- 90
30
ity of both systems in the mobile environ- 120 60

ment. For this part of the evaluation, the 20

150 30
vehicle shown in Fig. 3 was employed as the 10
mobile receiver. Since it was not possible
to record the real-time instant parameters of 180 0
the demodulator and decoder in a moving
vehicle, the original ports from the splitter
were respectively fed to a Panasonic ATSC 210 330

set-top box (TU-DST50W) and a Hitop [6]

DVB-T set-top box. The outputs were fur- 240
270
300

ther fed to a VHS tape recorder to provide

dynamic documentation of the results. Figure 11
(left) Maxviews omni-directional antenna and (right) its
Additionally, the port which was connected radiation pattern
to the spectrum analyzer was, for these tests,
linked to an FSS (Field Survey System) receiver the GPR4427A receiver from CHASE Communications
[7]. This system is assisted by a GPS receiver, such that the position data and the corresponding field strength
can be measured and recorded simultaneously. During mobile reception, it was impossible for the receiving
system to predict the incoming wave direction and, therefore, it was not possible to use the directional wide-
band log-periodic antenna. Instead, we used a Maxview [8] omni-directional antenna designed for horizontal
polarization (see Fig. 11 left), which we installed on the vehicle roof. Fig. 11 right shows its omni-direc-
tional radiation pattern at 593 MHz (again measured at Tatung University). This antenna is active in that it
provides 20 dB of RF gain.
In total, we measured reception along twelve highways and streets in this evaluation of mobile reception.
Some of these routes lie in a north-south direction and others lie in an east-west direction. By using MapInfos
pseudo-colour facility for field-strength illustration purposes, maps similar to Fig. 12 left could be created.

EBU TECHNICAL REVIEW July 2003 8 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
This map in fact shows the results of mobile reception along
the Cheng-Kuo highway which runs north-south and is the
main express highway through downtown Taipei.
The DVB-T modulation mode was set to 2k/16-QAM for
this session. However, for comparison purposes, further
tests were carried out on another Hsin-Shen express high-
way in downtown Taipei. This time, in addition to 2k/16-
QAM, we evaluated two other DVB-T modes: 2k/64-QAM
and 8k/64-QAM. It came as no surprise that the high data-
rate mode, 8k/64-QAM, is not so suitable for mobile recep-
tion. This judgment was based on real-time monitoring of
the picture quality. Fig. 12 right shows the pseudo-colour
results produced by Mapinfo for this highway.
All the TV programmes monitored along these twelve
highways and streets were video-taped. Furthermore, all
the GPS/E-field data for ATSC 8-VSB and DVB-T

Figure 12
Mobile reception tested on (left) the Cheng-
Kuo highway and (right) the Hsin-Shen high-
way

COFDM were recorded by the CHASE FSS.

By attaching a triggering sensor on one of the
vehicles wheels, it was possible to set a dis- Figure 13
tance interval say every 50m or 100m for E-field strength and path data were recorded by a
acquiring the real-time data on the vehicles CHASE FSS receiver

location and the received field

strength. The screenshot in
Fig. 13 shows an example of
these data in which the corre-
sponding E-field strength and the
path data were recorded simulta-
neously. Additionally, during the
reception procedure, the sur-
roundings and car speed (as wit-
nessed by the investigators) were
recorded orally (synchronized
with the picture video-taping).
Generally speaking, in these
measurement comparisons, it was
observed and may be concluded
Figure 14
that mobile reception cannot truly
Mobile reception tested on the No. 1 freeway passing through Taipei be associated with the ATSC 8-
city VSB system. During most of the

EBU TECHNICAL REVIEW July 2003 9 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
reception tests, the displayed pictures were graded as < 0 >. On the other hand, mobile reception of DVB-T
COFDM was very acceptable along large parts of the test routes but, on some sections, the pictures were inter-
rupted quite often. It is presumed that the deployment of an SFN (Single Frequency Network) is essential if a
reliable mobile service is to be achieved.

It was also observed with DVB-T that, if the high-speed data rate modulation is not applied, the influence of
vehicle speed on the picture grading can be ignored. Fig. 14 shows mobile reception along an east-west sec-
tion of the Taiwan No. 1 freeway through Taipei city. Note that the small dots are the instantaneously meas-
ured reception points.

Indoor fixed-point reception

Many high buildings are located in the urban areas of Taiwan, and new sky-scrapers are going up all the
time such that terrestrial TV reception is becoming ever harder to obtain. With most people residing in
apartments in these high buildings, the penetration of cable TV has almost reached 80% of the Taiwanese pop-
ulation. Any new terrestrial broadcasting service must therefore offer easy reception if it is to compete with
the cable TV services.

In the tests on indoor fixed-point reception, the equipment set-up was the same as that used for the outdoor
fixed-point reception tests, with the exception of the antenna. As during the mobile tests, the Maxview HP
omni-directional antenna was used during the indoor tests. Consequently, during these measurements, the
term best receiving direction was not taken into account. Based on the OFDM propagation characteristics,
we found that an omni-directional antenna was preferred over a directional one.

We evaluated reception at 103 locations inside the seven buildings listed in Table 3.
Table 3
Buildings used for the indoor reception tests

Organizations Measured locations Labelling for the

following text
Chinese Television System (CTS) 17 CTS
Taiwan Television Enterprise Ltd. (TTV) 14 TTV
China Television Company (CTV) 15 CTV
Formosa Television (FTV) 12 FTV
Sanlih E-Television Co., Ltd. (SET) 7 SET
The Directorate General of Telecommuni- 21 DGT
cations
Tatung Company 17 TATUNG

Table 4
Statistics of the measured results for indoor reception

CCIR Grade ATSC DVB-T

4 48 78
3 8 9
2 4 3
1 5 1
0 38 12

EBU TECHNICAL REVIEW July 2003 10 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
Table 4 shows the statistics of the measured results. In order to obtain satisfactory indoor reception, it seems
that an LoS (Line of Sight) is not necessary. As a matter of fact, most of the measured indoor locations had no
direct LoS towards the transmitter and, in many cases, the receiving antenna was not close to a window. That
is why it was mentioned above that a directional antenna is not suitable for receiving the OFDM signals
indoors.
It was conclusively shown that the DVB-T system is far superior to the ATSC system for indoor reception (see
Table 4). However, it is much harder, than in the case of outdoor fixed reception, to explain all the observed
phenomena. In particular, because of multipath factors, it is not easy to use the measured electric field strength
data to predict the graded picture quality of these two systems. Among the 103 measured indoor locations, the
lowest received field strength (Emin) for a picture grading of < 4 > was 47.5 dBV/m for ATSC (on the 3rd
floor of TTV) and 43.6 dBV/m for DVB-T (in the ground-floor lobby of CTV).
In the following, we try to explain certain anomalies we found at some indoor locations:
a) In the ground-floor lobby of CTV, the subjective picture grading was quite different between these
two systems
The received field strengths were almost the same < 44.2 (ATSC) vs. 43.6 dBV/m (DVB-T) > and their
received spectra looked quite healthy. But the CCIR grading for the ATSC pictures was < 1 > and for the
DVB-T pictures, < 4 >. With a great deal of RF clutter between the transmitting and the receiving
antennas, this was a very suitable location for demonstrating the differing characteristics of the two sys-
tems under test.

b) Marginal field strength, yet good reception on the 3rd Floor of TTV
As mentioned above, the field strengths at this location were among the weakest encountered during the
indoor tests. Nevertheless, both systems achieved a picture grading of < 4 >. It is worth noting that there
were indeed many other indoor locations where the field strength was greater than Emin, yet the reception
picture grades were less than < 4 >.
c) Very high field strength, yet bad reception
This happened on the 14th floor of the Tatung companys building where the window faces North, with an
LoS to the transmitter. The field strengths encountered here were of the order of 70 dBV/m. However,
reception at this location was impossible and it received a CCIR grading of < 0 > for both systems. This
is not easily explained. By examining the spectra recorded at this location (see Fig. 15), it can be seen
that the low-frequency section was seriously attenuated.

Figure 15
Received spectra of (left) ATSC and (right) DVB-T, showing considerable losses at low frequencies

d) Reception totally using a reflected signal (on the 9th floor of DGT)
The window where reception was evaluated faces south, and there is a large building just in front of it.
The transmitted signals were coming from the north, which is heavily shielded by internal office walls.
Consequently, reception at this location was totally based on the reflected waves from the facing wall of
the nearby building to the south. The results show that the ATSC system was not able to deal with such a
situation, but DVB-T produced a CCIR picture grading of < 4 >.

EBU TECHNICAL REVIEW July 2003 11 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
e) Relationship between spectrum flatness and the received vision quality (location 1 at the south-
west corner of the 6th floor of CTS; location 2 on the southeast side of the 9th floor of DGT)
Both these locations are not on the LoS to the transmitter and their spectra were seriously damaged by
multipath interference (see Fig. 16). The Site Margins associated with location 1 were not available for
ATSC and were 0 dB for DVB-T. The ATSC system failed at this location (grade < 0 >) but DVB-T
achieved a picture grading of < 4 >. These spectra show how the OFDM technique used in the DVB-T

Figure 16
Uneven received spectra:
location 1 (top left) ATSC; (top right) DVB-T
location 2 (bottom left) ATSC; (bottom right) DVB-T

Abbreviations
2k COFDM transmission mode with around 2000 ETN East of True North the direction or bearing, in
carriers a clockwise sense, relative to true north
8k COFDM transmission mode with around 8000 FFT Fast Fourier Transform
carriers GPS Global Positioning System
8-VSB 8-state Vestigial SideBand HP Horizontally Polarized
16-QAM 16-state Quadrature Amplitude Modulation ITU International Telecommunication Union
64-QAM 64-state Quadrature Amplitude Modulation LoS Line of Sight
ATSC Advanced Television Systems Committee
MER Modulus Error Ratio
(USA)
NTSC National Television System Committee (USA)
BER Bit-Error Ratio
OFDM Orthogonal Frequency Division Multiplex
CCIR (ITU) International Radio Consultative
Committee PCMCIA Personal Computer Memory Card International
Association
COFDM Coded Orthogonal Frequency Division
Multiplex QPSK Quadrature (Quaternary) Phase-Shift Keying
CW Carrier Wave RF Radio-Frequency
DTT Digital Terrestrial Television S/N Signal-to-Noise ratio
DTV Digital Television SER Segment Error Ratio
DVB Digital Video Broadcasting SFN Single-Frequency Network
DVB-T DVB - Terrestrial ToV Threshold of Visibility

EBU TECHNICAL REVIEW July 2003 12 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
system can take advantage of multipath signals to reconstruct the picture, even though the received DVB-
T spectra were obviously quite uneven. It is worth noting that the received field strengths were quite
strong at both locations around 51 dBV/m at location 1 and 57 dBV/m at location 2.
f) Role played by the pilot signal

Figure 17
Received spectra in which the ATSC pilot is destroyed:
location 1 (top left) ATSC; (top right) DVB-T
location 2 (bottom left) ATSC; (bottom right) DVB-T

At some locations, it was found that the lower end of the received spectra was extremely attenuated; for
example, as shown in Fig. 17.
The signals in the upper two images (ATSC on the left,
DVB-T on the right) were measured on the 16th floor of
the Tatung Company building (facing north). It can be
seen that the pilot signal of the ATSC spectrum is com-
pletely destroyed. The pilot signal which holds about
7% of the total energy of the channel spectrum plays
a very critical role in the demodulation of the ATSC
system [4][9] and is sometimes called simply the
ATSC Pilot (see Fig. 18). Hence, if the pilot signal is
considerably degraded, it is not possible to decode the
pictures. That is the reason why the ATSC system
obtained a picture grading of < 0 > at such locations.
However, there were no such problems with the DVB-
T system (grade < 4 > pictures) at these locations. Figure 18
ATSC Pilot
The two lower images in Fig. 17 (ATSC on the left)
show another similar case where the ATSC Pilot was
weakened and the pictures could not be decoded, hence a picture grading of < 0 >; on the other hand, the
DVB-T pictures were readily received (grade < 4 >). Again, it is worth noting that the electric field
strength of these received spectra were all above 60 dBV/m, and it is indeed improper to predict the pic-
ture quality simply based on the received electric field strength, especially for the ATSC system.

EBU TECHNICAL REVIEW July 2003 13 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION

Chi-Fang Huang received his B.Sc. degree in Electronic Engineering from National
Taiwan Ocean University (Keelung, Taiwan) in 1981. He obtained an M.Sc. and a
Ph.D. in electrical engineering from Tatung University (Taipei, Taiwan) in 1983 and
1990, respectively.

From 1990 to 1997, Dr Huang worked for Chunghwa Picture Tubes, Ltd., Taiwan, on
developing new high-precision magnetic deflection yokes for CRTs. From 1997 to
1999, as a Principal Engineer, he worked for Siemens Telecommunication Systems
Limited, Taiwan, on designing the high-speed backplane for the SONET Multiplexer.
Since 1999, he has been an Associate Professor of Tatung University. His current
interests are optical transparent antennas, small antenna designs, high-frequency
techniques and the modelling of high-speed interconnections.

The Nan Chang was born in Tainan, Taiwan, in 1953. He received a B.Sc. degree in
Physics from National Taiwan University (Taipei, Taiwan) in 1975, and an M.Sc. and
a Ph.D. in electrical engineering from Tatung Institute of Technology (Taipei, Taiwan)
in 1980 and 1987, respectively.

From 1980 to 1981, he was a Design Engineer in the Television Department at

Tatung Company (Taipei, Taiwan). From 1981 to 1984, he worked in the Antenna
Department of the same company, where he was responsible for antenna and RF cir-
cuit designs

In 1987, Dr Chang joined the faculty of Tatung Institute of Technology, where he

became a Professor in 1990. At present, he is with the Communication Institute of
Tatung University, where he is now a Professor and Director of the Microwave Labo-
ratory. He has published around thirty scientific papers. His current research inter-
ests are in reflectarray antennas and microwave circuit designs.

Chau-Yun Hsu received his B.Sc., M.Sc. and Ph.D. degrees in Electrical Engineer-
ing from Tatung Institute of Technology (Taipei, Taiwan) in 1981, 1983 and 1988,
respectively. He was a lecturer in the Department of Electrical Engineering at Tatung
University from 1983 to 1985, and served as the Associate Professor of Tatung Uni-
versity from 1988 to 1997.

Since 1998, Dr Hsu has been the Chair Professor of the Graduate Institute of Com-
munication Engineering at Tatung University. His current interests include wireless-
channel modelling and estimation, digital signal processing and image processing.

Conclusions
This article has described the results of comparative evaluations of the ATSC 8-VSB and DVB-T COFDM
digital broadcasting systems. The report which the article is based on was submitted in 2001 to the Taiwanese
government authority, for strategic information purposes.
The tests were carried out in the spring of 2001, in three different types of environment; namely, outdoor
fixed-point, outdoor mobile and indoor fixed-point. Fixed reception using an external directional antenna was
evaluated at 102 locations; mobile reception was evaluated along 12 highways or streets at fast/slow speeds
and indoor reception using an omni-directional set-top antenna was evaluated at 103 locations inside seven
large buildings. Based on the data gathered, the following conclusions may be drawn about these two different
digital TV transmission systems.
1) For outdoor fixed-point reception, the reliability of DVB-T was found to be statistically better than that
of the ATSC system but not obviously so.
Most Taiwanese people live in city areas where there are many apartments in tall concrete buildings. As a
result, the concept of traditional (analogue) TV broadcasting is basically unworkable in many parts of Tai-
wan and, hence, there is a high penetration of cable TV services. But now, with the arrival of digital broad-

EBU TECHNICAL REVIEW July 2003 14 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu
 DIGITAL TERRESTRIAL TELEVISION
casting systems, the European DVB-T system seems best able to cope with this situation particularly as it
can take advantage of the multipath signals caused by building clutter in the major towns and cities.
2) For indoor fixed-point reception, DVB-T showed its superiority over the ATSC system, mainly because
of its modulation technology, COFDM.
In this environment, DVB-T is fully capable of dealing with multipath signals. Moreover, the COFDM
technique can skilfully use reflected signals to enhance the reception quality even at locations where no
direct signal from the transmitter is present. At locations where the low end of the channel spectrum is
degraded seriously, the ATSC system readily fails because of its reliance on the low-frequency pilot sig-
nal for demodulation purposes. Technically, the DVB-T system is the more suitable candidate for provid-
ing indoor reception in the Taiwanese environment.
3) In the case of mobile reception, the ATSC system was unable to provide such a service during our tests.
When using the optional DVB low data-rate modulation and coding scheme 2k/16-QAM the DVB-T
system was found to be quite capable of providing a good mobile service. In addition, the COFDM mod-
ulation technique has the ability to enhance reception at locations where multipath is present and even
severe. This is a strong point and makes DVB-T COFDM a good solution for mobile TV reception.

Acknowledgements
This study was not possible to complete without help from The Television Academy of Arts & Sciences of the
Republic of China (the project sponsor). Special acknowledgement must also go to the engineering depart-
ment of Chinese Television System (CTS) and The Directorate General of Telecommunications, Ministry of
Transportation and Communications. Their assistance with this project is deeply appreciated.

Bibliography
[1] Yiyan Wu: Performance comparison of ATSC 8-VSB and VSB-T COFDM transmission systems for
digital television terrestrial broadcasting
IEEE Trans. Consumer Electronics, pp. 916 - 924, vol. 45, no. 3, August 1999.
[2] M. Massel: Digital Television DVB-T COFDM and ATSC 8-VSB
digitalTVbooks.com, 1999.
[3] MapInfo Corporation, USA: http://www.mapinfo.com/.
[4] Linley Gumm: Signal-to-noise relationships in 8-VSB
Tektronix Technical Brief, Sept. 1999.
[5] Tatung University: Field test on the terrestrial broadcasting transmission standards ATSC vs.
DVB-T
Final report (in Chinese), granted by The Television Academy of Arts & Sciences of the Republic of China.
[6] Hitop Communications Corp., Taiwan: http://www.hitopcomm.com/main.html.
[7] Chase Communications, http://www.chase-comms.co.uk/.
[8] Maxview limited, UK: http://www.maxview.ltd.uk/.
[9] http://www.tscm.com/TSCM101HDTV.html.

Contact
If any readers would like to discuss this article with Dr Huang, he can be contacted at: [email protected].
21 July 2003

EBU TECHNICAL REVIEW July 2003 15 / 15

C.-F. Huang, T.N. Chang and C.-Y. Hsu