ECE216 Signals and Systems

2010/11
Mid-Term Review
Signals Versus Systems
A signal is a function of time.
Time variable may be a real value, or an integer value
Continuous-time signal x(t);
Discrete-time signal x[n].
A system operates on an “input” signal and produces an
“output” signal.
CT or DT systems, depending on nature of input and
output signals.
Basic Operations on Signals
Time shifting: means delay by , =
delay by N samples.
Reflection: is the reflection of in the vertical
axis; similarly for a DT signal.
Time scaling: expands by a factor of T, if T>1,
or compresses it by 1/T, if T < 1.
Amplitude scaling: scales signal vertically by factor
of ; applies also to DT signals.
Types of Signals
Periodic
Even or Odd
Complex Exponential – very important class of signals,
includes exponentially damped sinusoids and
pure sinusoids .
Unit step and unit impulse
DT: for n = 0, 1, 2,...; for n = 0, all
other samples = 0.
CT Impulse and Step
CT unit step is defined as

CT unit impulse is a bit complicated. Visualized as a signal

that is zero everywhere except at t = 0. At t = 0, its value is
not defined but it satisfies

Both these functions play critical roles in system analysis.

Systems
A system produces a response to an input signal.
Systems may be natural e.g. the atmosphere acts on
sunlight and changes it; or they may be man-made e.g. an
audio recording system.
Man-made systems are usually designed to accomplish
certain desirable tasks.
Complex systems usually made up of sub-systems, and
parallel and serial inter-connections of these.
Important System Properties
We describe a system using words with particular
meanings, that define the properties a system possesses.
Memory: A system is memory-less if its output at any time
is a function only of the input signal at that time.
Invertible: A system is invertible if there exists another
system that can recover its input from its output.
Causality: A system is causal if its output at any time
depends only on input signal values in the past.
Important System Properties
Stability: A system is stable if all bounded input signals
result in bounded output signals.
Time Invariance: A system is time-invariant if a time shift
of the input results in the same time shift of the output.
Linearity: A system is linear if a linear combination of any
set of input signals gives the same linear combination of
output signals.
LTI Systems

The class of systems that are both linear and time-invariant

is particularly important.
Input and output signals are related by convolution in the
time domain.
Input and output signals are related by multiplication in the
frequency domain.
Leads to frequency-response description of filters.
DT Convolution
If input signal is x[n], impulse response of LTI system is
h[n] and output is y[n], then

If system is causal then

CT Convolution
Similarly for CT systems,

In both CT and DT convolution, we can use the famous

“flip and shift” method to solve simple problems.
A practical CT system is made up of physical elements –
they do not actually compute the system output from the
input using convolution!
Convolution
DT systems are mostly (all?) man-made. Many of these
(FIR systems) do compute the output from the input
through the convolution operation.
Generally, convolution is hard to visualize and hence hard
to get any intuition from.
This is an excellent reason for resorting to frequency-
domain methods!
Convolution Is...
Commutative:
Distributive over addition:

Implies that impulse response of parallel inter-

connection of two LTI systems is the sum of the two
impulse responses.
Associative:
Impulse response of cascade is convolution of the two.
LTI System Properties
From an LTI system’s impulse response, we can deduce all
of its properties.
Memory-less if and only if
Invertible if and only if there exists s.t.

Causal if and only if

Stable if and only if
ODEs and Difference
Equations
Linear constant-coefficient differential equations that link
system output with input + system is at “initial rest” =>
system is LTI.
Ditto for DT systems – replace “differential” with
“difference”.
Difference equations are easy to solve, differential
equations relatively harder.
DT system’s impulse response is easy to find;
CT impulse response not so easy to find.
LTI System Response to
Complex Exp. Inputs
Important observation that for any LTI system,

We call the complex exponential functions and

eigenfunctions of LTI systems.
The “eigenvalues” are
Periodic CT Signals
A periodic signal x(t) with fundamental period is said to
have “harmonic” frequencies where k is an
integer.
Sinusoids at all harmonic frequencies repeat every .
The k-th harmonic sinusoid has k complete cycles in that
time.
Can we decompose all periodic signals into
CT Fourier Series
It turns out the answer is “yes”, for all practical signals (i.e.
those that don’t “blow up”, or have an infinite number of
maxima/minima in one period, etc.)
The Fourier coefficients are obtained as

By plotting the magnitude of against we obtain the

magnitude spectrum of x(t); plotting the angle of gives
us the phase spectrum.
CTFS Properties
Linearity:

Time shifting:

Frequency shifting:

Differentiation in time:
CTFS Properties
Parseval’s theorem:

Parseval’s theorem shows that the k-th Fourier

coefficient’s squared magnitude is the power of the k-th
harmonic sinusoid in x(t).
These and other properties can be used to find Fourier
coefficients of signals that can be written in terms of
signals with known Fourier series.
CTFS and LTI Systems
The response of an LTI system to the complex sinusoidal
input is

is evaluated at :

Since the system is linear, a linear combination of complex

sinusoids will result in the same linear combination of
responses.
CTFS and LTI Systems
Therefore, when a periodic input x(t) with Fourier
representation is supplied to an LTI
system, its response would be

So the output is periodic with the same fundamental

frequency as the input, having Fourier coefficients
Frequency Response
Should be clear that the function determines the
action of an LTI system on its inputs –
if has larger magnitude over a frequency range
compared to another , then the spectral
content of the input in the latter range will be reduced
relative to that in the former.
This means that any LTI system can be viewed as a
frequency-selective filter, with its frequency
response.
Simple Circuit Examples
In an RC circuit, the voltage across the capacitor has a
“lowpass” relationship with the input i.e. the system acts as
a lowpass filter.
On the other hand, the voltage across the resistor has a
“highpass” response to the input voltage – the system acts
as a highpass filter.
These filters are very crude approximations to the ideal
lowpass and highpass responses.