EE483 Intro to Digital Signal Processing
Instructor: Richard Leahy, Fall 2016
Homework #1, Due Monday, September 5th.
To be considered on-time, all homeworks must be received by 5.00pm on the dute date. All homework must be submitted electronically as a single PDF file. The file must be clearly readible. I
will email you separately in the next couple of days with instructions on how to submit.
[1] Find the continuous-time Fourier transform of
x t = rect t sin 0 t
where 0 is a real positive constant and rect t = 1 t 0.5 .
0 otherwise
[2] Find the continuous-time Fourier transform of:
t t 0
x t = te
0 t0
where is a real positive constant.
2
d x t - as a function of X j , the
[3] (a) Write the continous time Fourier transform of y t = -------------2
dt
Fourier transform of x t .
(b) We can also write y t as the output of a continuous time LTI system with input x t . What is
the frequency response of this sytem?
[4] Find the inverse Fourier transform of
X j = e
where is a real positive constant.
[5] Given that the following Fourier transform pair:
2
2
x t = exp t and X j = exp
------- , find the Fourier transform of the zero-mean
4
t2
1
Gaussian distribution: p t = -------------- exp --------.
2
2 2
[6] For each of the following with real input x n and real output y n , determine whether the
systems are (a) linear (b) time invariant, (c) BIBO stable, (d) causal.
(a) y n = 2x n + 2
(b) y n = 4x n + 4x n
(c) y n = x n n 5 + u n where u n is the unit step sequence.
4
bk x n k + x n
(d) y n =
where b k
k=0
In each case, use the definition of the condition to examine whether each condition is satisfied for
ALL bounded input sequences.
[7] Prove that associative, distributive and commutative properties of the discrete convolution
sum (see lecture notes #3).
[8] Suppose you cascade two systems with the following impulse responses:
1
n=0
h 1 n = a u n and h 2 n = a
n=1
otherwise
0
Determine their combined impulse response using the properties from problem [7] and the discrete time convolution sum? (note: even if you are familiar with discrete Fourier and z transforms,
you cannot use them to answer this (or any other question in this homework).
n
[9] For what values of a (if any) will the system with impulse response below be stable?
n
(a) h n = na u n
n
(b) h n = na u n
(c) h n = na
[10] The cross-correlation operator is closely related to the convolution sum, and for real
sequences the correlation of two sequences x n and h n is defined as
r xh n =
x n + k h k
k =
(a) Is the cross-correlation operator commutative, ie does r xh n = r hx n ? If not, what is the
relationship between r xh n and r hx n .
(a) Under what conditions on h n will the cross-correlation r xh n be identical to the convolution of x n and h n . (Later well see that cross-correlations are very useful operators for target
detection (.e.g in radar and sonar) and also for characterizing transfer funtions (e.g. finding the
acoustical impulse response of a room).