As it can be seen in the above graph that it is an ideal square wave of amplitude equals 1, time period is 1mS and zero centric. In-built function of Matlab was used to generate square of frequency 1 kHz.
Task # 02: Represent the above generated square wave using compact Fourier series and show its amplitude and phase response.
A common engineering analysis technique is the partitioning of complex problems into simpler ones. The simpler problems are then solved, and the total solution becomes the sum of the simpler solutions Three requirements must be satisfied for such a solution to be both valid and useful. 1. First, we must be able to express the problem as a number of simpler problems. 2. Next, the problem must be linear, such that the solution for the sum of functions is equal to the sum of the solutions when we consider only one function at a time. 3. The third requirement is that the contributions of the simpler solutions to the total solution must become negligible after we consider only a few terms; otherwise, the advantage may be lost if a very large number of simple solutions is required. The Fourier analysis provides one of the most powerful and important set of tools and insights for analyzing, designing and understanding signals and LTI systems. In Fourier analysis a signal is decomposed into its constituent sinusoids, i.e. frequencies, the amplitudes of various frequencies form the frequency spectrum of the signal. In an inverse Fourier transform operation the signal can be synthesized by adding up its constituent frequencies. Fourier series is an approximation of a non-sinusoidal periodic function with a (sum of) sinusoids. With reasonable accuracy, a periodic signal can be expressed by a sinusoid. This approximate response would be the sinusoidal steady-state response of a linear system.
The following properties make the sinusoids the ideal choice as the elementary building block basis functions for signal analysis and synthesis: Orthogonality; two sinusoidal functions of different frequencies have the following orthogonal property
For harmonically related sinusoids the integration can be taken over one period. Similar equations can be derived for the product of cosines, or sine and cosine, of different frequencies. Orthogonality implies that the sinusoidal basis functions are independent and can be processed independently. 1. Sinusoidal functions are infinitely differentiable. This is important, as most signal analysis, synthesis and manipulation methods require the signals to be differentiable. 2. Sine and cosine signals of the same frequency have only a phase difference of π/2 or equivalently a relative time delay of a quarter of one period i.e. T0/4 . The Fourier expression of compact trigonometric fourier series is given by the following equations: Matlab code for magnitude and phase response: clear; % clear all variables clf; % clear all figures
% computing compact trigonometric fourier series components
syms Ck wave t k T=0.001; k=0:1:5 wo=(2*pi)/T; % fundamental frequency (rad/s) wave=exp (-1i*k*wo*t); Ck=(1/T)*(int (wave,t,0,0.0005) +int (-wave,t,0.0005,0.001)) simplify(Ck);
% Compute yt, the Fourier Series in Compact trigonometric form
N = 11; % summation limit (use N odd) c0 = 0; % dc bias t = 0:0.000001:0.0025; % declare time values
figure(1) yt = c0*ones(size(t)); % initialize yt to c0
for k = 1:2:N, % loop over series index n (odd)
ck = 2/(1i*k*pi); % Fourier Series Coefficient yt = yt +( 2*abs(ck)*sin(k*2*pi*1000*t)); % Fourier Series computation end
y = square(2*pi*1000*t); % plot original y(t)
plot(t,y,'-magenta'); hold on; % plot truncated Compact trigonometric FS plot(t,yt,'blue'); grid on; xlabel('t (seconds)'); ylabel('y(t)'); title(strcat('Compact Trigonometric Fourier Series with N = ',int2str(N),' harmonics'));
% Draw the magnitude spectrum of compact trigonometric Fourier Series
figure(2) % put next plots on figure 2 subplot(2,1,1) stem(0,c0); % plot c0 at kwo = 0 hold; for k = 1:2:N, % loop over series index k ck = 2/(1i*k*pi); % Fourier Series Coefficient stem(k*wo,abs(ck)) % plot |ck| vs kwo end hold on; for k = 0:2:N-1, % loop over even series index k ck = 0; % Fourier Series Coefficient stem(k*wo,abs(ck)); % plot |ck| vs kwo end
xlabel('w (rad/s)') ylabel('|ck|') title(strcat('Magnitude Spectrum with N = ',num2str(N),' harmonics')); grid on;
% Draw the phase spectrum of Compact Trigonometric Fourier Series
subplot(2,1,2) stem(0,angle(c0)*180/pi); % plot angle of c0 at kwo = 0 hold; for k = 1:2:N, % loop over odd series index k ck = 2/(1i*k*pi); % Fourier Series Coefficient stem(k*wo,angle(ck)*180/pi); % plot |ck| vs kwo end for k = 0:2:N-1, % loop over even series index k ck = 0; % Fourier Series Coefficient stem(k*wo,angle(ck)*180/pi); % plot |ck| vs kwo end xlabel('w (rad/s)') ylabel('angle(ck) (degrees)') title(strcat('Phase Spectrum with N = ',num2str(N),' harmonics')); grid on;
Task # 04: Suggest an appropriate value of N considering that the square waves generated in task 1 and 3 lie within 10 % difference.
The approach used to perform this task is calculation of total power of ideal square wave and then comparing it with the sum of power of individual harmonics. The first four harmonics of square wave contributes 90 % power of the total. So it has a 10% power less than the ideal square wave. The expression for Fourier series of the square wave is ∝ 4𝑉𝑠𝑖𝑛𝑤𝑜𝑡 𝑥(𝑡) = ∑ 𝑘=1 𝑘 𝑜𝑑𝑑 𝑘𝜋 The power of harmonics can be calculated as ∝ 4𝑉 𝑥(𝑡) = ∑ 𝑘=1 ([ ]^2)/2 𝑘 𝑜𝑑𝑑 𝑘𝜋 And the power of ideal square wave was calculated by formula Matlab code: % 1kHz square wave clc close all; clear all; t = -0.002:0.000001:0.002; y=0; % first signal for k=1:2:7; y=y+((1/k)*(4/pi)*sin(2*pi*k*1000*t));
end subplot(1,2,1); plot(t,y); x= square(2*pi*1000*t); % second signal subplot(1,2,2); plot(t,x) axis([-0.002 0.002 -1.5 1.5]); xlabel('Seconds'); ylabel('Amplitude') title('square wave') E1=sum(x.*conj(x))*0.000001; % computing signal energies E2=sum(y.*conj(y))*0.000001; C=sum(x.*conj(x))*0.000001/(sqrt(E1*E2)); C1=sum(x.*conj(y))*0.000001/(sqrt(E1*E2))
C1 =
0.9743
Figure: square wave 1.5 1.5
1 1
0.5 0.5
Amplitude 0 0
-0.5 -0.5
-1 -1
-1.5 -1.5 -2 -1 0 1 2 -2 -1 0 1 2 -3 Seconds -3 x 10 x 10 In above plot, it is showed that the wave synthesized from Fourier series have the same amplitude and frequency, the shape has a Gibb’s effect at the ends (turning point of wave) because the number of harmonics are less, if they are increased ideally infinite, the wave will look more ideal. If the number of harmonics is increased, the square wave generated will be much better than before but at the cost of time i.e. latency has increased due to increase in number of computations. It can be seen the above generated wave from its Fourier components (up to 4 harmonics) has 90 % resemblance to original square wave.
Question # 02:
Task # 01:Generate a voice signal using windows sound card utility and display it in the Matlab (both time and frequency domain response).
This task was performed by firstly recording of a 4 to 5 seconds sound by sound recorder of windows which was save with .wma extension. Matlab code:
% Task 1 (Generation of voice by windows Sound card utility
[m,fs]=audioread('C:\users\Soman\desktop\untitled.wma'); %read audio file (Fs=samples per second) %sound(m,fs); m=m(:,1); % message signal (voice) in time domain dt=1/fs; % seconds per sample t=0:dt:(length(m)*dt)-dt; % Time specifications in second N=length(m); f=(10*fs)/N.*(0:N-1); % Frequency specifications in hertz mF=fft(m,N); % Fourier Transform mF=abs(mF); % message signal in frequency domain
% ============================= % Plotting signal in time domain % =============================
% ================================ % Plotting Freq Response of Signal % ================================
figure(1); subplot(2,1,2); plot(f,mF,'b'); title('Freq Responce of Message Signal'); xlabel('f(Hz)'); ylabel('M(f)'); axis([0 12000 0 300]); grid on; Display in time and frequency domain: Message Signal
0.05 m(t)
-0.05
0.6 0.8 1 1.2 1.4 1.6 1.8
t(sec) Freq Responce of Message Signal 300
200 M(f)
100
0 0 2000 4000 6000 8000 10000 12000 f(Hz)
It can be seen that the components of audio signal are band limited to 15 kHz which is the maximum range of an audio signal. B (bandwidth) of this signal is 15 kHz but the significant information lies up to 6000 Hz. The spectra is showing that the frequency components below 200Hz have also less significance. The major information and power of voice lies within 500 to 6000 Hz.
Task # 02: Generate a DSB-SC waveform using the above signal and an appropriate bandpass carrier signal. Show the modulated signal both in time and frequency domains.
Matlab code: (Generation of voice by windows Sound card utility [m,fs]=audioread('C:\users\Soman\desktop\untitled.wma'); %read audio file (Fs=samples per second) %sound(m,fs); m=m(:,1); % message signal (voice) in time domain dt=1/fs; % seconds per sample t=0:dt:(length(m)*dt)-dt; % Time specifications in second N=length(m); f=(10*fs)/N.*(0:N-1); % Frequency specifications in hertz mF=fft(m,N); % Fourier Transform mF=abs(mF); % message signal in frequency domain
% Task 2 (Double side band suppress carrier amplitude modulation)
Fs=130000; % Sampling frequency of carrier and message signal Fc=59000; % Carrier frequency dsbsc=ammod(m,Fc,Fs); % Double side band suppressed-carrier modulation dsbsc_F=abs(fft(dsbsc)); % ============================= % Plotting signal in time domain % =============================
Time and frequency domain plot of DSB modulated voice signal:
DSBSC
0.05 DSBSC
-0.05
0.6 0.8 1 1.2 1.4 1.6 1.8
t(sec) Freq Responce of DSBSC 200
150 DSBSC(f)
100
50
0 0 1 2 3 4 5 6 f(Hz) 4 x 10 The above wave is a voice signal, so it is a mixture of frequencies with varying amplitude carrying the information but now it has shifted to 60 kHz because of modulation.No carrier is obtained in DSB-SC modulation as shown in the time plot and also spectra of signal.
Task # 03: Generate an Am waveform using the signal generated in task 1 and the carrier signal from task 2. Show the modulated signal both in time and frequency domains.
Matlab code:
(Generation of voice by windows Sound card utility
[m,fs]=audioread('C:\users\Soman\desktop\untitled.wma'); %read audio file (Fs=samples per second) %sound(m,fs); m=m(:,1); % message signal (voice) in time domain dt=1/fs; % seconds per sample t=0:dt:(length(m)*dt)-dt; % Time specifications in second N=length(m); f=(10*fs)/N.*(0:N-1); % Frequency specifications in hertz mF=fft(m,N); % Fourier Transform mF=abs(mF); % message signal in frequency domain
Fs=130000; % Sampling frequency of carrier and message signal
Fc=59000; % Carrier frequency
% Task 3 (Double side band amplitude modulation)
Am=ammod(m,Fc,Fs,0,0.05); % Transmitted-carrier modulation (amplitude modulation/Double side band) Am_F= abs(fft(Am)); % Frequency domain % ============================= % Plotting signal in time domain % =============================
% ================================ % Plotting Freq Response of Signals % ================================
subplot(212); plot(F,Am_F,'m'); title('Freq Responce of AM'); xlabel('f(Hz)'); ylabel('AM(f)'); axis([0 65000 0 600]); Time and frequency domain plot of amplitude modulated voice signal: AM 0.1
0.05 AM
-0.05
-0.1 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 t(sec) Freq Responce of AM 600
400 AM(f)
200
0 0 1 2 3 4 5 6 f(Hz) 4 x 10 Task # 04:Based on the results obtained from tasks 2 and 3 discuss if you achieved what was expected. In task 2, double sideband suppressed carrier was generated, so it contained less power because carrier was suppressed there while in task 3, AM was generated having carrier, as a result more energy contained in the signal and its spectrum. Amplitude modulation (AM) is similar to DSB but has the advantage of permitting a simpler demodulator, the envelope detector. In both tasks, after modulation the signal is not shifted to the specified carrier frequency, it may be because I have taken the frequency axis by my own.
Task # 05: Design a demodulator in Matlab (non-coherent) and show the demodulated signal in both time and frequency domains. Based on your results do you think you will need an amplification stage in the receiver?
Demodulator: Type: Rectifier detector The higher frequencies terms are eliminated by low pass filter and the output [A+m(t)]/π is obtained, then the signal is amplified actually attenuated by a scalar π.
Matlab code: (Generation of voice by windows Sound card utility [m,fs]=audioread('C:\users\Soman\desktop\untitled.wma'); %read audio file (Fs=samples per second) %sound(m,fs); m=m(:,1); % message signal (voice) in time domain dt=1/fs; % seconds per sample t=0:dt:(length(m)*dt)-dt; % Time specifications in second N=length(m); f=(10*fs)/N.*(0:N-1); % Frequency specifications in hertz mF=fft(m,N); % Fourier Transform mF=abs(mF); % message signal in frequency domain
Fs=130000; % Sampling frequency of carrier and message signal
Fc=59000; % Carrier frequency
% Task 3 (Double side band amplitude modulation)
Am=ammod(m,Fc,Fs,0,0.05); % Transmitted-carrier modulation (amplitude modulation/Double side band) Am_F= abs(fft(Am)); % Frequency domain % Task 5 (demodulation of Am signal (non-coherent)) r=Am.*(Am>=0); % Using rectifier technique r_F=abs(fft(r)); % Frequency domain
0 0.6 0.8 1 1.2 1.4 1.6 1.8 t(sec) Time domain plot of demodulated voice signal 0.04 Amplitude
0.02
0 0.6 0.8 1 1.2 1.4 1.6 1.8 time(sec) Freq Responce of Recovered signal 100 M(f)
50
0 0 2000 4000 6000 8000 10000 12000 f(Hz)
When a low pass filter is applied to the demodulator, the term (A+m(t))/π is obtained, as it has a dc offset of 1/π (amplitude of carrier divided by attenuation factor) and it is attenuated by a factor π, so an amplification is required to make it original voice signal. Also a high pass filter is required to get rid of dc offset. The cutoff frequency of the high pass filter should be approximately 100 to 200 Hz. We will lose these frequency components of our voice signal but that is not significant because a large number of information of voice signal lies within 500 to 3400 Hz range. . Lowpass filter magnitude plot: Magnitude Response (dB)
