UNIVERSIDAD DE LA SALLE

Signals analysis – Assignment # 2: Fourier Transform

1. SIMULINK: this problem must be solved using Matlab’s Simulink toolbox.

The files Assignment2_System*.mdl contain three different system implementations (System I, System II and System III) .
Making use of Simulink determine, for each system, if it corresponds to:
a) A linear or a non-linear system
b) A variant or a time-invariant system
In addition, determine (also using Simulink) the impulse response of each of these systems.

2. Consider the following discrete-time sequences:

⎧1 0≤ n≤ 4
x[n] = ⎨
⎩0 otherwise
⎧ 2 n=0
⎪ −1 n =1
⎪
h[n] = ⎨
⎪ 3 n=2
⎪⎩ 0 otherwise

a) Compute the convolution of both sequences (x[n]*h[n]) using the mathematical definition (that is using the
convolution sum)
b) Verify that convolution is commutative: obtain analytically the result of convolving h[n]*x[n], and verify that the
result is the same as that obtained in (a)
c) Repeat (a) and (b) using Matlab’s conv function, and verify that the results are the same.

3. For the continuous and discrete-time signals given below:

a) Obtain their period and frequency (both in rad/sec and in Hz)

b) Obtain their corresponding mathematical expression (as functions of t and n respectively)
c) Make a detailed plot of their frequency spectrum, without obtaining a mathematical expression of their Fourier
transforms. For the continuous-time signal plot its spectrum both in rad/sec and in Hz; for the discrete-time signal plot
its spectrum in the range -3π (rad/s) < ω < 3π (rad/s).
d) Based on the previous plots, obtain a mathematical expression of the corresponding Fourier transforms.
e) Apply the inverse Fourier transform to the expressions obtained in (d), and verify that the result corresponds to the
signals already obtained in (b).

4. Consider the discrete-time signal:

⎛ 17π ⎞
x[n] = 2Cos⎜ n⎟
⎝ 8 ⎠

a) Is this a periodic signal? If so, which is its period?

b) Make a plot of its frequency spectrum, for the range -3π (rad/sec) < ω < 3π (rad/sec)
c) Compare this signal with the discrete-time signal of exercise (3) and explain the differences both in time and in
frequency.

5. The figure below shows four signals, from xa(t) to xd(t), with all plots having the same time and amplitude scale:
 UNIVERSIDAD DE LA SALLE

(a) Which signal has the higher frequency? Do not make any computation
(b) Make a plot of the frequency spectrum (both in rad/s and in Hz) of each of these signals

6. The figure below contains the signals x1(t) y x2(t) (left) and |X1(jω)|, the magnitude of the Fourier transform of x1(t) (right):

Make a plot of |X2(jω)|, the magnitude of the frequency spectrum of x2(t).

7. The figure below is a plot of the signal cos(ωon):

Which of the signals plotted below corresponds to cos((ωo+2π)n)? For each of these signals obtain a detailed plot of its
Fourier transform.

8. MATLAB: all of the items in this problem must be solved using Matlab; when required also obtain analytical expressions.
The smoothness of a signal determines the frequency content of its spectrum. Consider the signals:
x(t) = u(t+0.5)-u(t-0.5)
 UNIVERSIDAD DE LA SALLE
y(t) = (1+cos(πt))[u(t+0.5)-u(t-0.5)]
a) Plot these signals. Can you tell which one is smoother?
b) Find and plot the magnitude and phase of the Fourier transform of x(t)
c) Find and plot (both using Matlab and obtaining an analytical expression) the magnitude and phase of the Fourier
transform of y(t)
d) Which one of these two signals has higher frequencies? Can you now tell which of the signals is smoother? Use
Matlab to decide. Make x(t) and y(t) have unit energy and plot (in Matlab) 20log10|Y(jω)| and 20log10|X(jω)| and see
which of the spectra shows lower frequencies.

9. MATLAB: all of the items in this problem must be solved using Matlab.
Let x(t) = r(t+1)-2r(t)+r(t-1) (where r(t) is the ramp signal) and y(t) = dx(t)/dt.
a) Plot x(t) and y(t)
b) Find X(jω) and plot its magnitude and phase spectra.
c) Find Y(jω) and plot its magnitude and phase spectra.
d) Determine from the above spectra which of these two signals is smoother. Use Matlab to plot 20log10|Y(jω)| and
20log10|X(jω)| and decide. Would you say in general that computing the derivative of a signal generates high
frequencies or possible discontinuities?

10. MATLAB: all of the items in this problem must be solved using Matlab.
Repeat items (a) to (c) of exercise (7) but assuming that:
x(t) = u(t + 1) − 2u(t) + u(t − 1)
t
y(t) = ∫ x(τ )dτ
−∞
From the spectra of x(t) and y(t) would you say that in general by integrating a signal you get rid of higher frequencies, or
smooth out a signal?

11. The continuous-time signal x(t) = sin(6πt) is sampled each Ts secs, and the corresponding discrete-time signal obtained,
x[n], is shown below:

¿Which is the value of Ts required to obtain such signal? Make a plot of the spectra of both signals (x(t) and x[n]).

12. A continuous-time signal x(t) is sampled with a frequency of 5Hz, and the discrete-time signal x[n] is obtained. Which of
the signals shown below can be sampled with this frequency so as to satisfy the Nyquist criterion? For each signal make a
plot of the original spectrum (corresponding to the continuous-time signals) and the resulting spectrum (corresponding to
the discrete-time signals) after sampling.

t (secs)

13. Consider the LTI continuous-time system, with input x(t) and output y(t), shown below:
 UNIVERSIDAD DE LA SALLE

The magnitude and angle (|H(jω)| y ∠H(jω)) of the frequency response are shown below:

(a) Suppose the input is given by the expression x(t) = cos(50t). Make a plot of the frequency spectrum (both magnitude
and angle) of the output signal, and obtain its corresponding mathematical expression y(t).
(b) Repeat (a) assuming that the input is sin(25t)
(c) Repeat (a) assuming that the input is 2cos(80t+π/3) + 5sen(220t+π/4)
(d) Repeat (a) assuming that the frequency spectrum of the input is:

(e) Repeat (a) assuming that the frequency spectrum of the input is:

(f) Make a plot of Y(jω) (the frequency spectrum of the output signal), if the input signal has the frequency spectrum
shown below:

14. Consider the LTI continuous-time system, with input x(t) and output y(t), shown below:

(a) If the output signal, y(t), is the one shown below, obtain a plot of x(t).

(b) Obtain the impulse response of this system.

 UNIVERSIDAD DE LA SALLE
(c) What kind of filter is this system (low-pass, high pass, all-pass, notch)?
(d) Assume that you make a cascade connection of the systems H(jω) and “?”, as shown below:

Determine the frequency response and the impulse response of the system “?” so as to ensure that the output y(t) of the
combined system equals x(t).
(e) Based on the result obtained in (b) answer the question: in the real world, is it possible to implement the system “?”?

15. Let y(t) = x1(t)x2(t), with x1(t) and x2(t) being both real signals. If the frequency spectra of y(t) and x1(t) are:

(a) Make a plot of the frequency spectrum of x2(t)

(b) Find a mathematical expression for the signal x2(t)
(c) Determine the bandwidth of the signals x1(t), x2(t) and y(t)

16. Consider the LTI discrete-time system shown below, with input x[n] and output y[n]:

If the input signal and the impulse response of the system have the frequency spectra shown in the figure below:

(a) Make a plot of the frequency spectrum of the output signal

(b) Make a plot of the frequency response a the filter defined as: H2(ejω) = 10 – H(ejω)
(c) Make a plot of the frequency spectrum of the signal obtained after filtering x(t) with the system defined in (b)
(d) Now suppose that H(ejω) and X(ejω) are defined as:
 UNIVERSIDAD DE LA SALLE

Obtain an expression of y[n] in terms of x[n]. In addition, discuss the range of valid values of α to ensure that the
system is physically realizable.

17. A continuous-time signal x(t) is made up of two narrowband pulses. The signal and its frequency spectrum are shown
below:

The previous signal is used as the input to an LTI continuous-time system, whose frequency response (magnitude) is shown
below:

(a) Make a plot of y(t), the output of the system

(b) Now suppose that the signal y(t) obtained in (a) is used as the input to the system shown below:

where the A/D converter samples the signal y(t) with a sampling frequency fs, and each sample of y[n], the resulting
discrete-time signal, is codified using M bits; the resulting bit stream is stored in a memory.

(i) Define the minimal sampling frequency required in the A/D conversion.
(ii) Suppose that you sample y(t) with twice the frequency obtained in (i) and that this signal lasts for 10 sec;
assume also that the corresponding discrete sequence is codified with 8 bits. Which is the required memory
size (in bytes)?
(iii) Discuss the effect that the change in the sampling frequency and/or in the number of bits has on the required
memory size.
(iv) Make a plot of the frequency spectrum of the signal at the output of the A/D converter, if the sampling
frequency is 800 rad/s and the frequency spectrum of the signal at the input of the converter is the one shown
below: