Simulation and Evaluation oI a DVB System using

Simulink (Vol. II)

Bachelor thesis perIormed in Electronics Systems
by
Alberto Prieto Pastor

LITH-ISY-EX-ET--05/0309--SE
Linkping 2005-04-18














Simulation and Evaluation of a DVB System Using
Simulink (Vol. II)

Bachelor thesis in Department oI Electronic Systems at Linkping Institute
oI Technology
by

Alberto Prieto Pastor
LITH-ISY-EX-ET--05/0309--SE






Supervisor: Kent Palmkvist
Examiner: Kent Palmkvist

Linkping 2005-04-18





Avdc!nIng, InstItutInn
Division, Depailnenl


Inslilulionen fi syslenleknik
581 83 LINKIINC

Datum
Dale
2OO5-O4-18


5prak
Language

Rappnrttyp
Repoil calegoiy

I5BN

Svenska/Svedish
X LngeIska/LngIish
LicenlialavhandIing
X Lxanensailele

I5RN LITH-ISY-LX-LT--O5/O3O9--SL


C-uppsals
D-uppsals

5crIctItc! nch scrIcnummcr
TilIe of seiies, nunleiing

I55N

viig iappoil
____





URL fr c!cktrnnIsk vcrsInn
hllp://vvv.ep.Iiu.se/exjoll/isy/2OO5/3O9/




TItc!
TilIe


SinuIalion and evaIualion of a DV Syslen using SinuIink (VoI. II)



Frfattarc
Aulhoi

AIleilo Iiielo Iasloi



5ammanfattnIng
DV (DigilaI Video ioadcasling) is lhe leIevision digilaI syslen. Il's hovevei nuch noie lhan a
sinpIe iepIacenenl foi exisling anaIogue leIevision liansnission, lhis syslen has nany
advanlages such as picluie quaIily and aIIovs you a iange of nev fealuies and seivices incIuding
sullilIing, nuIlipIe audio liacks, inleiaclive conlenl, nuIlinedia conlenl... The syslen is lased in
lhe OIDM lechnoIogy, vhich aIIovs DV syslen lo expIoil lhe specliun fiequencies in a lellei
vay, saving specliun, lul OIDM has an inpoilanl diavlack lhal is lhe peak lo aveiage povei
ialio piolIen. OIDM is lased on lhe IIT aIgoiilhns geneialing oilhogonaI sulcaiiieis. This
lhesis laIks aloul lhe inpiovenenl of lhe IAIR piolIen using lhe sofl conpiession nelhod. The
sinuIalion has leen deveIoped in SinuIink and MalIal. Wilh aII lhe infoinalion piesenled in
lhis lhesis, any usei can sinuIale lhe syslen. Thus, lhis lhesis can le inpioved using olhei
lechniques lo soIve lhe IAIR piolIen.



Nyckc!nrd
Keyvoid
OIDM, DV, IIIT, QAM, SinuIink, IAIR, CIipping, Conpiession, peak lo aveiage povei ialio


Simulation and Evaluation oI a DVB system using Simulink (Vol II) Abstract
i
Abstract

DVB (Digital Video Broadcasting) is the television digital system. It's however
much more than a simple replacement Ior existing analogue television transmission,
this system has many advantages such as picture quality and allows you a range oI
new Ieatures and services including subtitling, multiple audio tracks, interactive
content, multimedia content...
The system is based in the OFDM technology, which allows DVB system to exploit
the spectrum Irequencies in a better way, saving spectrum, but OFDM has an
important drawback that is the peak to average power ratio problem. OFDM is based
on the FFT algorithms generating orthogonal subcarriers.
This thesis talks about the improvement oI the PAPR problem using the soIt
compression method.
The simulation has been developed in Simulink and Matlab. With all the
inIormation presented in this thesis, any user can simulate the system. Thus, this
thesis can be improved using other techniques to solve the PAPR problem.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Abstract
ii
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PreIace
iii
Preface

Here I`m in Iront oI the computer writing the last lines oI one cycle which I`d
never imagined how it went to go, some years ago when I started with the career in
Gandia.

There were some hard work years with good and bad moments. This is one oI
these great moments, the end oI my career with this thesis, and Ior this reason I would
like to thank both people and institutions the support obtained.

First oI all I would like to thank the 'Linkpings University and 'Universidad
Politecnica de Valencia because they allowed me the development oI this thesis here
in Sweden.

To my coordinator Kent Palmkvist because oI the help and the patient that he had
with me.
To my coordinator in Gandia Jaime Lloret.

To my parents and sister, and the rest oI the Iamily because they have been a big
support in Gandia and now here in Sweden.

To my Iiends, all the marvellous people that I have met in Gandia, due to the nice
moments that we spent there, and people in Linkping as: Nando, Ramon, Juanjo,
Sergio, Carlos, Edu and Lopez; the best part oI the workday was the lunchtime with
you.

And specially to Sergio Isla because he pushed me to came here to Sweden, and
without him, I`d never have this thesis Iinished.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) PreIace
iv
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Index
v
Index

1. The Thesis......................................................................................................... 1
1.1. Overview................................................................................................... 1
1.2. Purpose...................................................................................................... 1
1.3. Planning..................................................................................................... 1
2. OFDM............................................................................................................... 3
2.1. Origin ........................................................................................................ 3
2.2. Description ................................................................................................ 3
2.3. Orthogonality. Subcarriers generation using the IFFT ................................ 4
2.3.1. Orthogonality..................................................................................... 4
2.3.2. Subcarriers generation using the IFFT................................................ 5
2.4. The PAPR.................................................................................................. 6
2.5. Other concepts about OFDM modulation................................................... 6
2.5.1. Attenuation ........................................................................................ 6
2.5.2. Delay Spread...................................................................................... 6
2.5.3. Cyclic PreIix ...................................................................................... 6
2.5.4. Applications....................................................................................... 7
2.5.5. Transmitter & Receiver ...................................................................... 8
3. Digital Video Broadcasting (DVB) .................................................................... 9
3.1. History....................................................................................................... 9
3.2. DeIinition and operation ............................................................................ 9
3.3. BeneIits ................................................................................................... 10
4. Simulink.......................................................................................................... 11
4.1. BrieI description ...................................................................................... 11
4.2. Some Ieatures .......................................................................................... 11
5. Simulation Model ............................................................................................ 13
5.1. Main Menu .............................................................................................. 13
5.2. Transmitter .............................................................................................. 14
5.2.1. Bernoulli .......................................................................................... 14
5.2.2. FEC Coder ....................................................................................... 14
5.2.3. Interleaver ........................................................................................ 14
5.2.4. QAM Mapping................................................................................. 15
5.2.5. IFFT................................................................................................. 15
5.3. Channel ................................................................................................... 16
5.4. Receiver................................................................................................... 16
5.4.1. FFT.................................................................................................. 17
5.4.2. QAM Demapping............................................................................. 17
5.4.3. Deinterleaver.................................................................................... 18
5.4.4. Decoder............................................................................................ 19
5.5. Error Rate Calculation ............................................................................. 20
5.6. Simulation Result..................................................................................... 21
6. PAPR Problem ................................................................................................ 23
6.1. Compression method ............................................................................... 23
6.2. Implementation oI the Compression method using Simulink.................... 24
6.3. Graphical soIt reduction results................................................................ 27
6.3.1. 2k 16QAM....................................................................................... 27
6.3.2. 8k 16QAM....................................................................................... 29
6.3.3. 2k 64QAM....................................................................................... 30
6.3.4. 8k 64QAM....................................................................................... 32
6.4. Other possible solutions ........................................................................... 33
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Index
vi
7. Comparison ..................................................................................................... 35
7.1. Advantages and Drawbacks Clipping Method.......................................... 35
7.2. Advantages and drawbacks oI SoIt Reduction Method............................. 35
7.3. BER results Ior Clipping and Compression methods ................................ 35
7.4. Conclusions ............................................................................................. 37
8. Possible Improvements.................................................................................... 39
9. ReIerences....................................................................................................... 41
10. Appendix..................................................................................................... 43
10.1. Model Pre-load Function...................................................................... 43
10.2. Model Post-load Function .................................................................... 43
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Index
vii
Table of Figures

Figure 1: Conventional multicarrier technique. .......................................................... 4
Figure 2: Orthogonal multicarrier modulation technique............................................ 5
Figure 3: Spectra oI an OFDM subchannel and OFDM signal. .................................. 5
Figure 4: This is one time-domain symbol with the cyclic preIix ............................... 7
Figure 5: Receiver OFDM. ........................................................................................ 8
Figure 6: Transmitter OFDM..................................................................................... 8
Figure 7: Example oI subsystem using Simulink................................................... 12
Figure 8: Main Menu............................................................................................... 13
Table 1: Operation Modes Parameters. .................................................................... 13
Figure 9: Transmitter Blocks. .................................................................................. 14
Figure 10: Inside the FEC Coder Block. .................................................................. 14
Figure 11: Inside the Interleaver Block. ................................................................... 15
Figure 12: Inside the QAM mapping Block. ............................................................ 15
Figure 13: Inside the IFFT Block............................................................................. 15
Figure 14: AWNG Channel. .................................................................................... 16
Figure 15: Receiver Blocks...................................................................................... 16
Figure 16: Inside FFT Block.................................................................................... 17
Figure 17: Inside QAM Demapping Block. ............................................................. 17
Figure 18: 16 QAM Constellation............................................................................ 18
Figure 19: Inside the Deinterleaver Block................................................................ 18
Figure 20: Inside Decoder Block. ............................................................................ 19
Figure 21: Viterbi decoder Parameters..................................................................... 20
Figure 22: BER Block ............................................................................................. 20
Figure 23: BER Ior 2k 16QAM and 2k 64QAM...................................................... 21
Figure 24: BER Ior 8k 16QAM and 8k 64QAM...................................................... 21
Figure 25: Time Domain signal. .............................................................................. 21
Figure 26: Frequency domain signal. ....................................................................... 22
Figure 27: Rectangular window. .............................................................................. 23
Figure 28: Non-Rectangular Window. ..................................................................... 24
Figure 29: SoItreduction Block................................................................................ 24
Figure 30: Inside SoItreduction Block. .................................................................... 25
Figure 31: Inside the Real Reduce Block. ................................................................ 26
Figure 32: 2k 16QAM Real signals in time domain. ................................................ 27
Figure 33: 2k 16QAM Imaginary signals in time domain......................................... 28
Figure 34: 2k 16QAM signals in Irequency domain (a)(b). ...................................... 28
Figure 35: 8k 16QAM Real signals in time domain. ................................................ 29
Figure 36: 8k 16QAM Imaginary signals in time domain......................................... 29
Figure 37: 8k 16QAM signals in Irequency domain (a)(b). ...................................... 30
Figure 38: 2k 64QAM Real signals in time domain. ................................................ 30
Figure 39: 2k 64QAM Imaginary signals in time domain......................................... 31
Figure 40: 2k 64QAM signals in Irequency domain(a)(b). ....................................... 31
Figure 41: 8k 64QAM Real signals in time domain. ................................................ 32
Figure 42: 8k 64QAM Imaginary signals in time domain......................................... 32
Figure 43: 8k 64QAM signals in Irequency domain(a)(b). ....................................... 33
Figure 44: BER Ior 2k 16QAM. (a)Clipping Method (b) Compression Method....... 36
Figure 45: BER Ior 2k 64QAM. (a)Clipping Method (b) Compression Method....... 36
Figure 46: BER Ior 8k 16QAM. (a)Clipping Method (b) Compression Method....... 36
Figure 47: BER Ior 8k 64QAM. (a)Clipping Method (b) Compression Method....... 37
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Index
viii

Simulation and Evaluation oI a DVB system using Simulink (Vol II) The Thesis
1
1. The Thesis
1.1. Overview
This report has been divided into eight chapters, where in chapter 2 a basic
understanding oI OFDM modulation that is the modulation that DVB System uses Ior
its implementation is intended to give. A brieI description about DVB will be
described in the chapter 3.
In chapter 4 there are any inIormation about Simulink that is the program which
allows us to design the system. The next chapter, number 5, will explain in detail each
block oI the model and aIter we will simulate the model and we could observe the
graphic results. The chapter 6 is the most important in this thesis because is when it
will tackle the principal task oI the thesis, the peak to average power problem, this
report will propose solutions and will design the soIt compression method using
Simulink.
The chapter 7 is a comparative between 2 methods to solve the PAPR problem, the
clipping method and the soIt compression method and in the chapter 8 possible
improvements about the development oI my thesis will be said.

Finally chapters 9 and 10 will show the reIerences and the appendix respectively.

1.2. Purpose

The principal purpose in this thesis is to solve the peak to average power problem,
which appears when OFDM modulation is used. DVB system uses this kind oI
modulation.

1.3. Planning

First oI all, is important to remark that this thesis is the second part oI a global
thesis, my mate Sergio Isla Hernandez has written the Iirst part, which has the same
name.
Is mandatory to say that we have worked together in the investigation to solve the
PAPR problem but aIter I have only explained in detail in this thesis about the
receiver and the second solution while my mate has developed the transmitter and the
Iirst solution.
In the chapter 7 we can do the comparison, because naturally we have the results
about the clipping method and soIt compression method.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) The Thesis
2
Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
3
2. OFDM
2.1. Origin
Orthogonal Frequencv Division Multiplexing (OFDM), a special Iorm oI Multi-
Carrier Modulation, (MCM), with densely spaced subcarriers and overlapping spectra
was patented in the US in 1966 by Chang, Bell Labs. Soon aIter in 1971, Weinstein
and Ebert proposed the use oI Fast Fourier Transform (FFT) and guard interval.

In the 1985, OFDM was described Ior mobile communication, and in 1987 Alar
and Lasalle suggested the use oI OFDM Ior digital broadcasting.

It is also important to emphasize the date oI 1997, because Digital Jideo
Broadcast (DVB) was created and only one year later, Magic WAND project,
demonstrated OFDM modems Ior wireless LAN.

Finally in the 2000s there have been three important events about OFDM:
IEEE 802.11a and HIPERLAN/2 standards Ior wireless LAN.
V-OFDM Ior Fixed Wireless Access.
OFDM considered Ior new IEEE 802.11 and 802.16 standards.

2.2. Description
In recent years Wireless communication has seen a spectacular boom. One main
reason is that many committees are working on standards to ensure that end products
Ior customers conIorm to global speciIications and rules established by network
standards. Some committees working on wireless standardization are IEEE 802.11
(wireless local area networks), IEEE 802.15 (wireless personal area networks), IEEE
802.16 (wireless broadband area networks), HiperLAN2, and Bluetooth.
OFDM technology is behind many oI these standards. With all these standards
addressing OFDM, newcomers to wireless might wonder what it is about OFDM
technology that makes it an obvious choice Ior these high-rate wireless standards. |1|

There are many reasons which OFDM standard has been chosen, but one oI the
most important is to increase the robustness against Irequency selective Iading or
narrowband interIerence.

In a single carrier modulation data is sent sequentially over the channel by
modulating one single carrier, in this kind oI modulation we can have problems such
as a single Iade or interIerence can cause the losses oI all the inIormation. This event,
is not possible in a multi-carrier system (only a Iew sub-carriers are damaged)
because it uses an error correcting code, which is able to correct damaged carriers.

OFDM, as a MCM, consist oI dividing all the available bandwidth into several
sub-channels. MCM is accomplished by transmitting data stream divided into several
parallel bit streams, each oI which has a much lower bit rate. The lower bit rate
implies that the symbol period is long. As the symbol period increases the ISI
reduces, thereby improving the channel perIormance so that at any instant time many
data symbols are being transmitted. |2|


Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
4
OFDM as other systems has its own advantages and drawbacks which are the
Iollowing.

EIIicient way oI dealing with multipath delay spread.
EIIicient use oI spectrum by allowing overlap.
By dividing the channel into narrowband Ilat Iading subchannels, OFDM is
more resistant to Irequency selective Iading than single carrier system are.
In relatively slow time varying channels, it is possible to signiIicantly enhance
the capacity by adapting the data rate per subcarrier according to the SNR oI
that particular subcarrier.
Eliminates ISI and IFI through use oI a cyclic preIix.
Using adequate channel coding and interleaving one can recover symbols lost
due to the Irequency selectivity oI the channel.
OFDM makes single Irequency networks possible, which is specially
attractive Ior broadcasting applications.

- OFDM is more sensitive to Irequency oIIset and phase noise.
- OFDM has a relatively large Peak to Average power ratio, (PAPR) which
tends to reduce the power eIIiciency oI the RF ampliIier. (The objective oI our
thesis is to reduce this drawback).

2.3. Orthogonality. Subcarriers generation using the IFFT

2.3.1. Orthogonality
At Iirst in a classical parallel data system, the total signal Irequency band is divided
into N nonoverlaping Irequency subchannels.

Each subchannel is modulated with a separate symbol, and then the N subchannels
are Irequency-multiplexed. It seems good to avoid spectral overlap oI channels to
eliminate interchannel interIerences, but this kind oI modulation, has the problem oI
ineIIicient use oI the available spectrum. (Figure 1).


Figure 1: Conventional multicarrier technique.

To solve this ineIIiciency the idea proposed were to use parallel data and FDM,
with overlapping subchannels. Using the overlapping multicarrier modulation we can
save almost 50 oI bandwidth. (Figure 2)

To realize the overlapping multicarrier technique, however we need to reduce
crosstalk between subcarriers, which means that we want orthogonality between the
diIIerent modulated carriers. |1|

Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
5
Figure 2: Orthogonal multicarrier modulation technique.

We`ve just Ieatured an important term about OFDM, the orthogonality, and now
this concept will be explained careIully.

The orthogonality between subcarriers should IulIill these two properties:

Each subcarrier has an exactly an integer number oI cycles in the T interval.
The number oI adjacent subcarriers diIIers by exactly one.

II the system carry out the mentioned properties, it is possible that the sidebands oI
each subcarrier can be overlapped, and aIter, we can receive the total signal without
adjacent carrier interIerence

2.3.2. Subcarriers generation using the IFFT
An OFDM signal, consist oI a sum oI subcarriers that are modulated by using
Phase Shift Keving (PSK), or Quadrature Amplitude Modulation (QAM), iI the
complex are QAM symbols, Ns is the number oI subcarriers, T the symbol duration,
and I
c
the carrier Irequency, then one OFDM symbol starting at t T
s
can be written
as :



t
s
_ t _ t
s
T



s(t) 0, t t
s
` t~ t
s
T

In this representation the real and imaginary parts correspond to the in-phase and
quadrature parts oI the OFDM signals, which have to be multiplied by a cosine and
sine, oI the desired carrier Irequency to produce the Iinal OFDM signal.


Figure 3: Spectra of an OFDM subchannel and OFDM signal.
s ( ) t

Ns
2
Ns
2
1
exp f 2
i
T
( ) t ts
d
iNs/2
Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
6
2.4. The PAPR
An OFDM signal consists oI a number oI independently modulated subcarriers,
which can give a large peak to average power ratio (PAPR) when added up
coherently. Large PAPR means large ampliIier backoIIs, which reduces power
eIIiciency oI RF ampliIier, also another drawback oI PAPR is that brings an increase
complexity oI Analog to Digital and Digital to Analog converters.

There are several solutions to solve the PAPR problem such as clipping and peak
windowing, and compression techniques. In this thesis I will talk about the
compression method and my mate oI project about the clipping method.

Reducing the Peak to Average Power Ratio is the main task oI this report.

2.5. Other concepts about OFDM modulation
In this section we are just going to deIine some concepts which are not the
objectives oI this thesis, but we consider that it is important to mention.

2.5.1. Attenuation
It is possible to deIine attenuation as the loss oI power when the signal is
transmitted Irom the transmitter to the receiver. It can be caused Ior many reasons, but
the most important are:
The path length.
Path obstructions or shadowing produced by buildings and hills.
Multipath.

2.5.2. Delay Spread
Delay spread is the time spent between the arrival oI the Iirst and the last multipath
signal.
The received signal is the sum oI the direct signal plus all signals that are reIlected
by building, hills, . which arrive later than direct signal as a consequence oI the
extra path covered.
In a digital system delay spread can cause Intersvmbol Interference (ISI).

2.5.3. Cyclic Prefix
By adding a guard time, called cyclic preIix, the channel can be made to behave as
iI the transmitted waveIorms were Irom time minus inIinite, and thus ensure
orthogonality, which essentially prevents one subcarrier Irom interIering with another
(called intercarrier interference, or ICI).

The cyclic preIix is actually a copy oI the last portion oI the data symbol appended
to the Iront oI the symbol during the guard interval. |3|
Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
7


Figure 4: 1his is one time-domain symbol with the cyclic prefix

2.5.4. Applications
Although OFDM was born in the 60`s, it has been in the last years when it has
started to be used in commercial systems. The systems in which OFDM is used are
the Iollowing:

(DAB), Digital Audio Broadcasting uses (QPSK) Quadrature Phase Shift
Keving. Its principal advantage is that it is able to establish single
Irequency network, this means that transmitters which are separated at a
distance, can send the same signal simultaneously. This event produces
that the receiver receives the transmitted signal, each one with a diIIerent
propagation delay. It`s important to underline that propagation delay is
smaller than the Guard interval.

(DVB), Digital Jideo Broadcasting is quite similar to DAB system, the
main diIIerence between them is that the Iirst one is used Ior broadcasting
oI digital television signals using MPEG-2 source compression, that
reduces a standard television channel to 3Mbps, also reduces the HDTV
Channel to 20Mbps.

(WLAN), Wireless Local Area Network. The main objective oI this
system, which standardization is taken place in Europe, USA, and Japan at
the same time, is to get high data rates in the 5 GHz Irequency band.

Finally the standard that intends to provide high speed Internet access and
multimedia services to Iixed users receives the name oI (WLL), Wireless
Local Loops.


Simulation and Evaluation oI a DVB system using Simulink (Vol II) OFDM
8
2.5.5. Transmitter & Receiver
The Iigure shows the basic diagram block about an OFDM receiver in which we
can see a group oI blocks that they are going to be explained in detail later.
The shaded blocks are considered the most important blocks Ior the development
oI the thesis, and this is the reason that they will be simulated in advance with
simulink .




Figure 5: Receiver OFDM.

The Iigure 6 shows the transmitter block diagram that is exactly the same than the
receiver one but in the other direction.

This report does not have to explain the transmitter`s block in detail; this is only
mention here because the explanation oI all these blocks will be carried out in the
other part oI the thesis.



Figure : 1ransmitter OFDM.

Ciclic PreIix
Removal
Seriall to
Parallel
QAM
demapping
Parallel to
Seriall
FFT Deinterleav
er
Fec
Decoder
De-
Scrambler
Scrambler Fec Coder Seriall to
Parallel
QAM
Mapping
Interleaver
IFFT
Parallel to
Seriall
Ciclic
PreIix
Simulation and Evaluation oI a DVB system using Simulink (Vol II) DVB
9
3. Digital Video Broadcasting (DVB)

3.1. History
Founded in September 1993, the DVB Project is a market-led consortium oI public
and private sector organizations in the television industry. Its aim is to establish the
Iramework Ior the introduction oI MPEG-2 based digital television services. Now
comprising over 200 organizations Irom more than 25 countries around the world,
DVB Iosters market-led systems, which meet the real needs, and economic
circumstances, oI the consumer electronics and the broadcast industry.

DVB is an European initiative. Equipment conIorming to DVB standard is now in
use on six continents and is DVB rapidly becoming the world-wide standard Ior
digital TV. At the time DVB was being developed in Europe, a parallel program oI
standards and equipment development was also going on in the USA by the Advanced
Television Svstem Committee (ATSC) |4|.

3.2. Definition and operation
Digital Video Broadcasting is a transmission scheme based on the moving pictures
expert group MPEG-2 video compression. It's however much more than a simple
replacement Ior existing analogue television transmission.

DVB provides superior picture quality with the opportunity to view pictures in
standard Iormat or wide screen (16:9) Iormat, along with mono, stereo or surround
sound.
It also allows a range oI new Ieatures and services including subtitling, multiple
audio tracks, interactive content, multimedia content...

The radioelectric channels oI Digital Television take up the same bandwidth
(8MHz) than analogue television channels but due to the use oI compression
techniques oI sound and image signals (MPEG) can hold up a diIIerent number oI
channels depending on the bit rate, being able to oscillate between one high
deIinition television program to Iive programs with a similar quality than the actual
television.

The DVB-T system is characterized by two transmission modes: the 2k mode and
the 8k mode.
We have. 6 817 carriers per OFDM symbol Ior the 8k-mode (6 048 useIul, the
others Ior synchronization and signaling) and 1 705 carriers per OFDM symbol Ior
the 2k-mode (1 512 useIul carriers) are speciIied in the DVB-T system. Virtual
carriers are inserted in such a way that the total number oI carriers becomes a power
oI two, so that the Iaster algorithm oI the Inverse Fast Fourier Transform (IFFT) can
be used. At the receiving side, the corresponding signals can be easily recovered using
the respective 2k-FFT or 8k-FFT.

In order to ensure robust transmission oI the OFDM signal, an error protection
code is applied. In addition to the Iixed algorithm oI energy dispersal, block coding,
outer and inner interleaving, a Rate Compatible Punctured Convolutional (RCPC).
The mother code has a constraint length oI 7 bits and works with a code rate oI 1/2.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) DVB
10
The two generator polynomials oI the convolutional encoder are 171 and 133 in octal
notation.
To adapt the error protection to the actual transmitting conditions, several code
rates can be chosen (1/2 to 7/8). The code rate 1/2 has the highest redundancy, but the
highest transmission saIety. This mode should be applied to strongly disturbed
channels. On the other hand a code rate oI 7/8 has a low redundancy but a very weak
error protection. ThereIore, it should be used Ior channels with only low interIerence.
As mentioned above, every carrier is modulated by a modulation symbol. 16-QAM
and 64-QAM are used as modulation methods, e.g. 4 or 6 bits per modulation symbol.

3.3. Benefits
Using DVB such as broadcasting method provides many beneIits.

Soon reach almost everybody, thus allowing to receive signals with the
current analog antenna domestic system, moreover the Irequencies that the
DVB-T system will use are the same as those oI the analog TV service.

Allows the portable and movement reception.

Require less transmission power than other systems.

The behavior oI digital signal due to interIerences is better than the
analogue signal one.

Carry out the regional/ local broadcasting.

To reduce the network implementation costs, thanks to the re-use oI the
existing analog inIrastructure.

Allows the convergence TV-PC.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulink
11
4. Simulink

4.1. Brief description
Simulink is a platIorm Ior multidomain simulation and model-based design oI
dynamic systems. It provides an interactive graphical environment and a customizable
set oI block libraries that let you accurately design, simulate, implement, and test
control, signal processing, communications, and other time-varying systems. Add-on
products extend the Simulink environment with tools Ior speciIic modeling and
design tasks and Ior code generation, algorithm implementation, test, and veriIication.

Simulink is integrated with MATLAB, providing immediate access to an
extensive range oI tools Ior algorithm development, data visualization, data analysis
and access, and numerical computation.

With Simulink, you can quickly create, model, and maintain a detailed block
diagram oI yours system using a comprehensive set oI predeIined blocks.
Simulink provides tools Ior hierarchical modeling, data management, and
subsystem customization, making it easy to create concise, accurate representations,
regardless oI your system`s complexity.
Simulink includes more than 1000 blocks that implement Iunctions commonly
used in modeling a system.

With Simulink, you build models by dragging and dropping blocks Irom the
library browser onto the graphical editor and connecting them with lines that establish
mathematical relationships between the blocks. You have immediate access to
common graphical editing Iunctions, such as copy, paste, and undo.

Simulink is also practical. With thousands oI engineers around the world using it
to model and solve real problems, knowledge oI this tool will serve you well
throughout your proIessional career. |5|

4.2. Some features
Simulink has many interesting characteristics and now this section is going to
explain some oI them that have been used in the development oI this thesis:
Using oI workspace allows you to store all the values in several variables,
and later you can see all these values just by writing the variable's name
that you've deIined beIore.
Simulink lets you organize your model into clear, manageable levels oI
hierarchy by using subsystems. Subsystems encapsulate a group oI blocks
and signals in a single block. They are the primary method Ior including
hierarchy in your model.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulink
12


Figure 7: Example of subsystem using Simulink.

The Callbacks pane lets you speciIy Iunctions to be invoked by Simulink
at speciIic points in the simulation oI the model. You can enter the names
oI any callback Iunctions you want to be invoked in the appropriate Iields.
In DVB there are several modes oI operation, my mate oI thesis and me
have developed only Iour oI them. The values oI the diIIerent models have
been loaded in the callbacks pane and they can be watched in the appendix
at the end oI this thesis. We`ve created a menu where users can choose the
option that they want.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
13
5. Simulation Model
As I have said beIore in the point 2.5.5, only the receiver blocks will be explained,
and not the transmitter blocks, because it is not this thesis task, anyway transmitter
blocks will be mentioned because these blocks will be necessary these to run the
simulation model. Apart oI the transmitter and receiver blocks, this section gives
inIormation about the main menu oI the Simulation Model and the channel used in the
simulation.

5.1. Main Menu
In the thesis Iour possible cases have been developed, where you can choose one oI
them depending oI the kind oI modulation and the number oI subcarriers that you
want to use. For this reason one menu has been created, which will ask you at the
beginning, what option you want to run.




Figure 8: Main Menu.


Each case has diIIerent values Ior its variables; you can see in the table 1 the most
important parameters Ior each one.


2k 16 QAM 8k 16 QAM 2k 64 QAM 8k 64 QAM
Subcarriers
2048 8192 2048 8192
Samples per
Frame
5112 10224 5112 10224
Puncture Vector
|1 1 0 1 1 0| |1 1 0 1 1 0| |1 1 0 1 1 0| |1 1 0 1 1 0|
M-ary number
4 4 6 6
QAM symbols
1704 6816 1704 6816
Zeros added
344 1376 344 1376

1able 1: Operation Modes Parameters.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
14

5.2. Transmitter
The Iigure 9 shows the global transmitter blocks.

Figure 9: 1ransmitter Blocks.

5.2.1. Bernoulli
First oI all we should generate a random sequence oI binary numbers. Simulink
allows us to choose between diIIerent blocks to carry out this task, like random
integer, Poisson integer and Iinally Bernoulli binary generator, but Ior the
development oI this thesis the best option is the last one because the system must
work with binary numbers.
5.2.2. FEC Coder
The purpose oI using the Forward Error Correction (FEC) is to add enough
redundancy and protection to the signal so it can be robustness opposite to the noise
oI the channel, being possible to correct errors later in reception.



Figure 1: Inside the FEC Coder Block.

5.2.3. Interleaver
To avoid the burst errors, which can produce a lot oI consecutive errors causing a
high BER in reception, we must use the interleaver block we can see in the Iigure 9.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
15

Figure 11: Inside the Interleaver Block.

5.2.4. QAM Mapping
Quadrature amplitude modulation (QAM) is the most popular type oI modulation
in combination with OFDM. Especially rectangular constellations are easy to
implement.

Figure 12: Inside the QAM mapping Block.


5.2.5. IFFT
Probably this is the most important block Ior all OFDM systems. This block
generates the orthogonal subcarriers. The orthogonality between subcarriers as well as
saving bandwidth, it is also important Ior the right operation oI the system.

Figure 13: Inside the IFF1 Block.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
16
5.3. Channel
The channel is the block that allows us to transmit the signal Irom the transmitter
to the receiver.


Figure 14: AWAC Channel.

The AWGN Channel block adds white Gaussian noise to a real or complex input
signal. When the input signal is real, this block adds real Gaussian noise and produces
a real output signal. When the input signal is complex, this block adds complex
Gaussian noise and produces a complex output signal.

Typical values oI the SNR Ior an AWGN channel are between 5 and 30. When the
value is closer to zero this produces a high BER, on the other hand, iI the value oI
SNR increases we will get a better BER. We can see this Iact in the Iigure 23.
In this report we always simulate the system with a value oI 18.5 dB. Later in the
point 5.5 and in the comparison we can see what will happen when we try with
diIIerent values.

The reason that this channel has been chosen instead oI other channels, is because
this thesis is only going to evaluate the PAPR problem, with this kind oI channel is
enough to carry out this task.
AWGN block does not introduce problems like Iading and multipath, which would
be more diIIicult to the development oI this thesis.


5.4. Receiver
The Iigure 17 shows the global receiver blocks.

Figure 15: Receiver Blocks.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
17
5.4.1. FFT
The purpose oI this block is to get QAM symbols at the output, Irom the
subcarriers created in the IFFT block beIore the channel.


Figure 1: Inside FF1 Block.

We can see three diIIerent blocks.
The Iirst one is called FFT block, and its task is to perIorm the Fast Fourier
TransIorm.
The Selector1 is the second block. BeIore in the transmitter, we had to add zeros,
because the IFFT block was Ieed up with power oI two numbers. Naturally iI in the
transmitter we have added zeros, now in the receiver we have to remove them and the
block that carries out this task is the Selector1. For instance iI in the transmitter in the
2k (16 or 64 QAM) case we have 1704 QAM symbols and we add 344 zeros with the
zero pad, the resulting selector element is |1:852 1705:2048 853:1704|; where the
added zeros are between 1705 and 2048. Now in the receiver the selector element is
|1:852 1197:2048|; this means that we remove all the values between 853 and 1196
that are the zeros added in the transmitter block.
Finally the Frame conversion block, takes the output values oI the selector block
and converts them into Irames. At the output oI this block we get the QAM symbols.
5.4.2. QAM Demapping
The complex QAM symbols at the output oI the FFT block have to be demapped.
This is the principal task oI the QAM Demapping block.


Figure 17: Inside QAM Demapping Block.

As we can see, only two blocks compose the global block, the Iirst one whose
name is Rectangular QAM demodulates the input signal (which are complex number)
using the rectangular quadrature amplitude modulation method. The outputs oI this
block are integer numbers which belong each one to the Gray constellation points that
you can see in the Iigure:
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
18


Figure 18: 1 QAM Constellation.
This kind oI constellation is called Gray.
II we look at two nearby points (horizontally or vertically), we realize that they
diIIer in only one bit. We can see this in the Iigure18. That is a good thing when we
work on data transmission because iI the channel damages the signal, we could
receive 1,j1/3 instead oI 1,j, Ior instance. That means that we haven`t got the right
signal but this is not bad because we have received just one wrong bit. It is much
more possible to receive a nearby ground point than a distant one. |6|
The second block is called Integer to Bit Converter, whose task is to convert the
integer numbers in blocks oI N bits depending oI the kind oI case that we have
chosen, Ior instance iI we choose 16QAM this block rearranges in groups oI 4 bits,
because ( 2`416)

5.4.3. Deinterleaver
The interleaver block rearranges the bits. We have to remember that in the
transmitter the interleaver avoided possible burst and Iading errors by changing the
input bits order.



Figure 19: Inside the Deinterleaver Block.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
19
We can see in the Iigure, that the design oI this block is not very complicated.
Initially we put the buIIer which Iunction is only to store the input bits. The second
block called Matrix Deinterleaver as we have said beIore is to rearrange these bits.

5.4.4. Decoder
This is the last receiver block. Here all the redundant bits which were added in the
transmitter must be removed; it is mandatory to remember that we have used a code
rate oI 3/4.


Figure 2: Inside Decoder Block.

The Unipolar to Bipolar Converter takes the input bits and converts them; iI the
input is "1" the output will be "-1" and iI the input is "0" the output will be "1". This is
necessary because in this system a Viterbi decoder, which only distinguishes between
1 and -1, is used.

The second block (Insert Zero) adds redundant bits (only zeros) with the insert zero
vector. Inside oI this block my insert zero vector is |110110|. This block constructs an
output vector by inserting zeros among the elements oI the input vector:
Each 1 indicates that the block should place the next element oI the input in the
output vector

Each 0 indicates that the block should place a 0 in the output vector
The reason oI doing this is because in the transmitter we removed bits using the
puncture vector, and now in the receiver we must add them.

The Viterbi decoder is the last block and decodes input symbols to produce binary
output symbols. We use the same trellis structure than in the transmitter
(poly2trellis(7, |171 133|)) The parameters oI this Trellis Structure are: 7, because we
want to work in octal, and 171 and 133 because this Trellis structure are the speciIied
in DVB standard.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
20


Figure 21: Jiterbi decoder Parameters


5.5. Error Rate Calculation
The Error Rate Calculation block compares input data Irom a transmitter with
output data Irom the receiver. It calculates the error rate as a running statistic, by
dividing the total number oI unequal pairs oI data elements by the total number oI
input data elements Irom one source. |6| As in our case the input are bits, then the
block computes the bit error rate. In the Iigure we can see one example where the Iirst
box oI the display means the Error rate, the second is the total number oI error bits,
and the third one is the total number oI bits that we sent.





Figure 22: BER Block


II we want to design a good system is mandatory that the BER would be small.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
21


Figure 23: BER for 2k 1QAM and 2k 4QAM


Figure 24: BER for 8k 1QAM and 8k 4QAM

In the Iigures 23 and 24 we show how the signal is damaged when the SNR is
decreasing.

5.6. Simulation Result
When we run the simulation in our particular case oI 2k 16QAM we can see as
result in the time domain the signal oI the Iigure 25 and in the Irequency domain in
the Iigure26. The Iigure 25 where we can observe several peaks allows us to get a
graphical idea oI the PAPR problem.



Figure 25: 1ime Domain signal.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Simulation Model
22



Figure 2: Frequency domain signal.
The PAPR is the main problem oI the OFDM signal, and the objective oI this
thesis is to evaluate this PAPR and try to get solutions to solve it. We want to remark
that we will always work in the time domain to reduce the PAPR.


Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
23
6. PAPR Problem
A large PAP ratio brings disadvantages like an increased complexity oI the analog-
to-digital and digital-to-analog converters, which should have enough levels to
represent all the inIormation, and the reduction oI the eIIiciency oI the RF power
ampliIier. To reduce the PAPR, several techniques have been proposed, and can be
divided in diIIerent categories.
Clipping reduction.
SoIt reduction (Compression Techniques).
Coding techniques: Use a special Iorward-error correcting code set that
excludes OFDM symbols with a large PAP ratio.
Techniques based on Scrambling: Scramble each OFDM symbol with
diIIerent scrambling sequences and select that sequence that gives the
smallest PAP ratio.

In this thesis only the second possible solution will be carried out. We will design
the blocks in Simulink to reduce the PAPR, and we could observe the diIIerence
between the original signal and the reduced signal.

The place where we can observe easier the peak to average power ratio is at the
end oI the transmitter and just beIore the channel. This is the reason that we will place
the block to solve the PAPR problem in this part oI the model.

6.1. Compression method
With clipping solutions we have problems such as: the non-linear distortion oI the
OFDM signal signiIicantly increases the level oI the out-oI band radiation. This eIIect
can be understood easily because with this method we multiply the signal by a
rectangular window Iunction as you can see in the Iigure 27.



Figure 27: Rectangular window.
To remedy the out oI band problem oI clipping in this system the soIt compression
method is used, where we multiply the signal by a nonrectangular window. (Iigure28)

Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
24


Figure 28: Aon-Rectangular Window.

While in the clipping method only change the values oI the peaks that are over the
threshold, the soIt compression method will change all the values oI the signal but not
in the same proportion. The peaks which values are larger in comparison to the
threshold will suIIer more attenuation than the values that are below the threshold.
We can understand it watching the Iigure 30. Ideally the window should be as
narrowband as possible to minimize the out oI band interIerence and on the other
hand the window should not be too long in the time domain, because data implies that
many signal samples are aIIected, which increases the BER.

To get the window oI the Iigure 28 is necessary to use the Iunction:
g(A) A/ (1A
2p
)
1/2p


As a conclusion this method is very similar to the clipping method with the
advantage oI solving the problem oI out-oI band but with the drawback that the BER
will be higher because all the samples will be aIIected while in clipping only the
peaks over the threshold will be changed.

6.2. Implementation of the Compression method using
Simulink
The Iigure 29 shows the subsystem "SoItreduction" where are implemented many
blocks which allow us to solve the problem oI the peak to average power.
We must emphasize that to see the signal in the time domain is necessary to divide
the signal into Real and Imaginary part. ThereIore iI we want to do the soIt
compression oI the signal we must do it discerning between the real and imaginary
part.


Figure 29: Softreduction Block.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
25

The block called SoItreduction is the most important block in this report, because
inside this block the PAPR problem will be solved. We divide this block into two
subblocks, as we can see in the Iigure 30.


Figure 3: Inside Softreduction Block.


The Iirst subblock reduces the real part, and the second block reduces the
imaginary part, this report only will explain the 'Real Reduces block because the
'Imaginary Reduces is the same than the Iirst.

BeIore reducing the signal we have converted the complex signal into real and
imaginary, and when we would have reduced the signal we will convert the real and
imaginary part into a complex signal which will be transmitted by the channel.

Now in the Iigure 31 are shown all the blocks inside the Real Reduces block.

At the beginning oI the design oI the block we have to distinguish between positive
or negative input values; Ior this reason we have used the relational operator, which
task is only to compare the input value with the constant which value is 0. We only
explain the case oI positive values because the operations Ior the negative values are
very similar and you can observe all the math operations in the bottom oI the Iigure
31.
We have designed this block according to the Iunction that we have deIined in the
top oI this point. At the output oI the relational operator we will get in the case oI
positive values: "1" iI the input is bigger than 0 and "0" iI the input is smaller than
"0". AIter that the input signal will be multiplied by the sequence oI 1 and 0 that we
have obtained aIter the comparison.
The next step is to elevate the obtained signal to 2p where the p is a diIIerent
constant Ior each case, later we have to add one to the signal and another time to
elevate all this signal to another constant that in this case will be 1/2p. Finally we
have to divide the input signal by this. As we can see we only Iollow the operations
maths oI the Iunction explained in page 24.

For the negative values we only calculate the absolute value oI the signal and we
work with the same Iunction Ior positive values.

The last step is to add the two signals, positive and negative, and we will get the
soIt compressed signal.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
26


Figure 31: Inside the Real Reduce Block.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
27
Probably there is a block in some library oI Simulink that carries out this task,
but we think that with this implementation we can realize how this compression
method is working only using math operations.


6.3. Graphical soft reduction results
Now the design oI the DVB system, included the block to solve the peak to
average power problem is Iinished, thereIore the next step is to simulate it and to
observe all the results.
First we will show the real signal graphic without been reduced and later the
graphic oI the reduced signal, where we can see the diIIerence between them.
We will do this Ior each case that is 2 k 16QAM, 8 k 16QAM, 2 k 64QAM, and 8
k 64QAM. We have to remark that in each case a diIIerent threshold is chosen
depending oI the obtained average, and this value is arbitrary, the important is that we
can choose where the signal will suIIer an important compression.

6.3.1. 2k 16QAM



Figure 32: 2k 1QAM Real signals in time domain.


Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
28
The Iirst case is the 2k 16QAM. It is important to remark in this case, that 2k
means that we work with 2048 subcarrier, and we use a 16QAM modulator. As we
can see in the Iollowing Iigures all the values will be modiIied, but not in the same
way. All the values that are close to the average oI the signal will be aIIected only a
little bit (the Iinal value oI these peaks will have around a 90 oI the initial value)
but the peaks which values are larger compared to the average will be modiIied
considerably (30 oI the initial value).

We can see in the top oI the Iigure30 the real part oI the signal with several peaks
which are over the average. AIter using the Clipping Block we realize in the bottom
oI the Figure 32 that there are not too much peaks which values are over ,0.1,.



Figure 33: 2k 1QAM Imaginary signals in time domain.
Opposite to the real part here in the imaginary part the average is the same value
but in negative, this means that most oI the peaks are below 0 value. In the Iigure 33
we can see the imaginary part in the top and the reduced imaginary part in the bottom.



Figure 34: 2k 1QAM signals in frequency domain (a)(b).
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
29
The Iigure 34 shows all signal in the Irequency domain at the output oI the
channel. This is no possible in the time domain because to see the signal in this
domain is mandatory to separate the signal between Real and Imaginary part. We can
see that in the Iigure (b) the peaks are not as high as in the Iigure (a).
6.3.2. 8k 16QAM
8k 16 QAM is the second case. In this option, 8k means that we work with 8192
subcarriers, and a 16QAM modulator is used. For this case we have the limit on 0.04.



Figure 35: 8k 1QAM Real signals in time domain.

The Iigure 36 only shows the imaginary signal with or without being reduced.



Figure 3: 8k 1QAM Imaginary signals in time domain.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
30
In the Iigure 37 we show the entire signal in the Irequency domain at the output oI
the channel. The Iigure (a) shows us the original signal and the Iigure (b) the signal
aIter clipping.




Figure 37: 8k 1QAM signals in frequency domain (a)(b).
6.3.3. 2k 64QAM

The third case is the 2k 64QAM. It is important to remark in this case, that 2k
means that we work with 2048 subcarriers, and we use a 64QAM modulator. The
limit now is ,0.2,.



Figure 38: 2k 4QAM Real signals in time domain.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
31

We can see in the top oI the Iigure 38 the real part oI the signal with several peaks
which are over the average. II we realize, iI the peak value is much higher, even
though being reduced is still being high.



Figure 39: 2k 4QAM Imaginary signals in time domain.




Figure 4: 2k 4QAM signals in frequency domain(a)(b).


Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
32
6.3.4. 8k 64QAM

The last case is the 8k 64QAM. This option works with 8192 subcarriers and uses
a 64QAM modulator. The threshold that we have chosen is ,0.1,.



Figure 41: 8k 4QAM Real signals in time domain.


Figure 42: 8k 4QAM Imaginary signals in time domain.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
33


Figure 43: 8k 4QAM signals in frequency domain(a)(b).

6.4. Other possible solutions
As we have said beIore in the point 5, there are several options to solve the peak to
average power problem. Now some oI them are going to be explained brieIly, and we
will say the advantages and drawbacks oI each one oI the cases.

Clipping Method: It is the simplest way to reduce the peak to average
power ratio. With this method the peak amplitude becomes limited to some
threshold

Coding and Scrambling methods: A disadvantage oI distortion techniques
(Clipping and SoIt Compression) with a large PAP ratio, suIIer more
degradation, so they are more vulnerable to errors. To reduce this eIIect,
Iorward-error correction coding can be applied across several OFDM
symbols. By doing so, errors caused by symbols with a large degradation
can be corrected by the surrounding symbol. In a coded OFDM system, the
error probability is no longer dependent on the power oI individual
symbols, but rather on the power oI a number oI consecutive symbols.
Although such a low symbol error probability may be good enough Ior
real-time circuit-switched traIIic, such as voice, it may still cause problems
Ior packet data. To solve this problem standard scrambling techniques can
be used to ensure that the transmitter data between initial transmission and
retransmissions are uncorrelated. To achieve this, the scrambler has to use
a diIIerent seed Ior every transmission, which can be realized Ior instance
by simply adding one to the seed aIter Ior every transmission |1|

Simulation and Evaluation oI a DVB system using Simulink (Vol II) PAPR
34
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Comparison
35
7. Comparison
As I have said in the point 1.3, this is the second part oI the thesis, but there are
another part developed Irom my mate Sergio Isla Hernandez.
This thesis only evaluates the results about the soIt compression method but the
comparative between Clipping and soIt compression methods can be explained
because we have researched together and later we have divided all the inIormation in
two parts.
7.1. Advantages and Drawbacks Clipping Method
It is very easy to understand, peaks over the threshold are removed, and below
the threshold have the same value than at the input.
It is possible to implement this block, using only math operations.
Is not important the number oI bits that you introduce, only mind the peaks
that are over the threshold
We have to choose a threshold, depending oI the value oI the threshold, the
BER will be high or not, iI the threshold is not high there will be several peaks
and this Iact will introduce a high BER, otherwise iI we choose a lower
threshold, probably there will be not peaks and the BER will be almost cero.
The nonlinear distortions oI the OFDM signal signiIicantly increase the level
oI the out-oI-band radiation
7.2. Advantages and drawbacks of Soft Reduction Method
It is very easy to understand too.
This method like the clipping method can be implemented with math
operations.
The principal advantage respect to the Iirst method is its behavior isn't so
drastic. Actually is a soIt compression.
With this method we remedy the out oI band problem oI clipping method.

We change all the values, independently iI they are over or below the
threshold, but the compression is not the same Ior all the values. ( Figure 28)
In this method is important the number oI bits that we introduce, because we
change all the values, needless to say that Ior lower input values, the output
value is almost the same.

7.3. BER results for Clipping and Compression methods
The next Iigures are going to show how aIIects the clipping method and the soIt
compression method to the BER.

The Iollowing graphics show us the relation between SNR and BER Ior each one
oI the diIIerent cases that we have developed beIore.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Comparison
36


Figure 44: BER for 2k 1QAM. (a)Clipping Method (b) Compression Method




Figure 45: BER for 2k 4QAM. (a)Clipping Method (b) Compression Method




Figure 4: BER for 8k 1QAM. (a)Clipping Method (b) Compression Method
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Comparison
37


Figure 47: BER for 8k 4QAM. (a)Clipping Method (b) Compression Method

We can see in the Iigures Irom 44 to 47 that the BER is higher in the soIt
compression method than in the clipping method when the SNR value decreases.

7.4. Conclusions
The main objective oI this thesis is to remove the Peak to Average Power Ratio
problem, (PAPR), which appears in systems like DVB when the OFDM modulation is
used. In order to solve this problem the SoIt Compression Method has been proposed.

Once the system is designed and run is possible to see that the goal oI this thesis
has been gotten, in spite oI the model is still having other problems like the increase
oI BER because we change all oI the signal values.

When we simulate the model without the compression solution, we can observe
several peaks which will be reduced proportionally once introduced the compression
block.

In a Iuture, probably somebody can improve this task using another diIIerent
system with new Ieatures.

Finally all the investigation has been carried out by two students (diIIering
between transmitter and receiver) being diIIicult to separate the inIormation in two
diIIerent reports because both are necessary to understand how the system is working.
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Comparison
38
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Improvements
39
8. Possible Improvements
Once Iinished the development oI this thesis we will be able to say possible
improvements which will be:

We have chosen a channel which only introduces Gaussian noise, but
actually when we send one signal though the channel, it usually has
problems such as Iading or multipath. One improvement would be to
design the same system that we have designed but choosing other channel.

We did not worry about the synchronism in the system, but this parameter
is important iI we want to design a real DVB system.

In this system exists only one available sort oI modulation technique, but iI
we have a look at the diIIerent applications in which OFDM is used you
will realize that some oI them use other techniques, Ior instance QPSK. (In
this case QAM modulation was used).

To eliminate intercarrier interIerence (ICI) is important to design the cyclic
preIix block, this block has been not designed in this report because we
have not this problem, but actually when we work with OFDM the ICI is a
big problem.

We have only researched about two methods to solve the peak to average
power ratio problem, but there are other techniques to solve this problem
like coding and scrambling.



Simulation and Evaluation oI a DVB system using Simulink (Vol II) Improvements
40
Simulation and Evaluation oI a DVB system using Simulink (Vol II) ReIerences
41
9. References
|1| Richard Van Nee, Ramjee Prasad, "OFDM For Wireless Multimedia
Comunications" Artech House 2000.

|2| Engels,Mark OFDM Systems, How to make them work? Kluwer Academic
Publishers, 2002.

|3| IEC: OFDM Ior mobile Data Communications: http://www.iec.org (2005-2-17)

|4| DVB Digital Video Broadcasting: http://erg.abdn.ac.uk/research/ (2005-2-21)

|5|Mathworks: http://www.Mathworks.com (2005-3-7)

|6| Simulink 6.2 help.

Simulation and Evaluation oI a DVB system using Simulink (Vol II) ReIerences
42
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Appendix
43
10. Appendix
In this section we include Matlab`s Iunctions to set the value oI the variables in the
system.
10.1. Model Pre-load Function

nsubcarriers=1704;
outputrows=2048; %IFFT/zeropad and IFFT/UEselector
selectorelements1=[1:852 1705:2048 853:1704]; %IFFT/UEselector
selectorelements2=[1:852 1197:2048]; %FFT/UEselector
numbitsperinteger=4; %QAMmapping/bit to integer converter
mnumber=16; %QAMmapping/rectangular QAM
buffersize=6816;
interrow=96; % Interleaver/ matrix interleaver
intercolumn=71; % Interleaver/ matrix interleaver
bitsperblock=5112; % Bernoulli block
constant=2048;
threshold= 0.1;
threshold1= -0.1;
threshold2= 0.1;
threshold3= -0.1;
exponente=0.6;
exponente2=1.666667;

10.2. Model Post-load Function


Mode=menu('Choose a modelbetwen the 4 different options ' ,'2k16QAM','2k64QAM','8k16QAM','8k64QAM');
switch Mode
case 1
nsubcarriers=1704;
outputrows=2048; % IFFT/zeropad and IFFT/UEselector
selectorelements1=[1:852 1705:2048 853:1704]; % IFFT/UEselector
selectorelements2=[1:852 1197:2048]; % FFT/UEselector
numbitsperinteger=4; % QAMmapping/bit to integer converter
mnumber=16; % QAMmapping/rectangular QAM
buffersize=6816;
interrow=96; % Interleaver/ matrix interleaver
intercolumn=71; % Interleaver/ matrix interleaver
bitsperblock=5112; % Bernoulli block
constant=2048;
threshold= 0.1;
threshold1= -0.1;
threshold2= 0.1;
threshold3= -0.1;
exponente=0.6;
exponente2=1.666667;
case 2
nsubcarriers=1704;
outputrows=2048; % IFFT/zeropad and IFFT/UEselector
selectorelements1=[1:852 1705:2048 853:1704]; % IFFT/UEselector
selectorelements2=[1:852 1197:2048]; % FFT/UEselector
numbitsperinteger=6; % QAMmapping/bit to integer converter
mnumber=64; % QAMmapping/rectangular QAM
buffersize=10224;
interrow=144; % Interleaver/ matrix interleaver
intercolumn=71; % Interleaver/ matrix interleaver
bitsperblock=7668; % Bernoulli block
constant=2048;
threshold= 0.2;
threshold1= -0.2;
threshold2= 0.2;
threshold3= -0.2;
exponente=1;
exponente2=1;
case 3
nsubcarriers=6816;
outputrows=8192; % IFFT/zeropad and IFFT/UEselector
selectorelements1=[1:3408 6817:8192 3409:6816]; % IFFT/UEselector
Simulation and Evaluation oI a DVB system using Simulink (Vol II) Appendix
44
selectorelements2=[1:3408 4785:8192]; % FFT/UEselector
numbitsperinteger=4; % QAMmapping/bit to integer converter
mnumber=16; % QAMmapping/rectangular QAM
buffersize=13632;
interrow=142; % Interleaver/ matrix interleaver
intercolumn=96; % Interleaver/ matrix interleaver
bitsperblock=10224; % Bernoulli block
constant=8192;
threshold= 0.04;
threshold1= -0.04;
threshold2=0.04;
threshold3=-0.04;
exponente=0.6;
exponente2=1.666667;
case 4
nsubcarriers=6816;
outputrows=8192; % IFFT/zeropad and IFFT/UEselector
selectorelements1=[1:3418 6837:8192 3419:6836]; % IFFT/UEselector
selectorelements2=[1:3418 4775:8192]; % FFT/UEselector
numbitsperinteger=6; % QAMmapping/bit to integer converter
mnumber=64; % QAMmapping/rectangular QAM
buffersize=41016;
interrow=1709; % Interleaver/ matrix interleaver
intercolumn=24; % Interleaver/ matrix interleaver
bitsperblock=30762; % Bernoulli block
constant=8192;
threshold= 0.1;
threshold1= -0.1;
threshold2=0.1;
threshold3=-0.1;
exponente=0.9;
exponente2=1.11111111;
end

45
Pa svenska

Detta dokument hlls tillgngligt p Internet eller dess Iramtida ersttare under en lngre
tid Irn publiceringsdatum under Irutsttning att inga extra-ordinra omstndigheter uppstr.
Tillgng till dokumentet innebr tillstnd Ir var och en att lsa, ladda ner, skriva ut enstaka
kopior Ir enskilt bruk och att anvnda det oIrndrat Ir ickekommersiell Iorskning och Ir
undervisning. verIring av upphovsrtten vid en senare tidpunkt kan inte upphva detta
tillstnd. All annan anvndning av dokumentet krver upphovsmannens medgivande. Fr att
garantera ktheten, skerheten och tillgngligheten Iinns det lsningar av teknisk och
administrativ art.
Upphovsmannens ideella rtt inneIattar rtt att bli nmnd som upphovsman i den omIattning
som god sed krver vid anvndning av dokumentet p ovan beskrivna stt samt skydd mot att
dokumentet ndras eller presenteras i sdan Iorm eller i sdant sammanhang som r
krnkande Ir upphovsmannens litterra eller konstnrliga anseende eller egenart.
Fr ytterligare inIormation om Linkping University Electronic Press se Irlagets hemsida
http://www.ep.liu.se/


In English

The publishers will keep this document online on the Internet - or its possible replacement -
Ior a considerable time Irom the date oI publication barring exceptional circumstances.
The online availability oI the document implies a permanent permission Ior anyone to read, to
download, to print out single copies Ior your own use and to use it unchanged Ior any non-
commercial research and educational purpose. Subsequent transIers oI copyright cannot
revoke this permission. All other uses oI the document are conditional on the consent oI the
copyright owner. The publisher has taken technical and administrative measures to assure
authenticity, security and accessibility.
According to intellectual property law the author has the right to be mentioned when his/her
work is accessed as described above and to be protected against inIringement.
For additional inIormation about the Linkping University Electronic Press and its procedures
Ior publication and Ior assurance oI document integrity, please reIer to its WWW home page:
http://www.ep.liu.se/

Alberto Prieto Pastor