3.

Time scaling: if x(t) is a signal (continuous/discrete) then x(αt) is the time scaled
version of x(t) where α is a constant.

4. Periodic signal: A periodic signal has a property that there is a positive value of T
such that x(t) = x(t+nT) for n = 0, 1, 2 … and T is a constant known as period.

5. Even or odd signal: From time reversal if the mirror image of the signal is same
as that of the image. That is if x(-t) = x(t) then the signal is even else if the signal
x(-t) = -x(t) then the signal is odd. Note that any signal can be broken to sum of
signal one which is even and other is odd as given below: -

Even {x(t)}=1/2[x(t)+x(-t)]

Odd {x(t)}=1/2[x(t)-x(-t)]

Exponential and sinusoidal signals:-

Continuous time complex exponential signal is of the form

x(t) = Ceat
where C and a are generally complex number.

1. Real and exponential signal that is if ‘C’ and ‘a’ are real then the signal is as
follows:
 2. Periodic complex exponential and sinusoidal signal: That is ‘a’ is imaginary

x(t) = Cejwot
x(t) = Cejwo(t+T) =C. ejwot.ejwoT but ejwoT=1
i.e. if w0=0 then x(t)=1 and if w0≠0 then T=2π/|w0|
Signal closely related is x(t)= a cos(w0t+φ)
Euler’s relation: ejwot=cosw0t+jsinw0t
Acos(w0t+φ)=A.Re{ej(wot+φ)} and Asin(w0t+φ)=A.Im{ej(wot+φ)}

3. Growing and decaying sinusoidal signal:

x(t)=Certcos(w0t+φ) if r>0 then growing signal and if r<0 then decaying signal

Sinusoidal signal multiplied by decaying exponential is referred as damped

exponential. Similarly for the discrete time characteristic where t becomes n.

Unit impulse and unit step function:

Important Properties of Signals:
 Linear time-invariant system and convolution integral:
LTI Systems: They are the systems that are linear and time invariant in nature.
Convolution: if h[n] is the impulse response of the system then if x[n] is the input
and y[n] is the output then,
y[n] = x[n] * h[n] // convolution

//convolution sum or superposition sum

Properties of LTI system:

1. Commutative: x1[n] * x2[n] = x2[n] * x1[n]
2. Distributive: x[n] * (h1[n] + h2[n]) = x[n] * h1[n] + x[n] * h2[n]
3. Associative: x[n] * (h1[n] * h2[n]) = (x[n] * h1[n]) * h2[n]
4. LTI system with and without memory
5. Invertibility and inverse system
6. Causality of LTI system
7. Stability of LTI system
8. Unit step response of LTI system: By convolution δ[n] = u[n] * h[n] there
for h[n] in discrete time LTI system is h[n] = δ[n]. δ[n-1]

Causal LTI system described by:

1. Linear constant coefficient differential equation: dy(t)/dt + 2y(t) = x(t)

2. Linear constant coefficient difference equation
 Fourier series representation of periodic signal
Fourier series representation of continuous time periodic signals can be given as a
linear combination of harmonically related complex exponentials:-

A signal is periodic if x(t) = x(t + nT) for all t where T is a constant known as
period and n = 0, 1, 2, … and T = ω0/2π.
E.g. x(t) = cosω0t, x(t) = ejω0t

Complex exponential

Фk(t) = eikω0t = ejk(2π/T)t k= 0, ±1, ±2…

∞ ∞
jkω0t
i.e. x(t) = Σ ak. e = Σ ak.ejk(2π/T)t
-∞ -∞

if k = 0, x (t) is constant
k = ±1 fundamental frequency ω0 is the fundamental components known as first
harmonic component.
k = ±2 second harmonic component.
k ± n is nth harmonic component.
This representation is known as Fourier series representation of periodic signal.

To determine ak multiply both with e-jkω0t

∞ ∞
i.e. x(t) e-jkω0t = Σ ak. ejkω0t e-jkω0t = Σ ak.ejk(2π/T)t e-jkω0t
-∞ -∞

Integrating from 0 to T
T T ∞ ∞ T
i.e. ∫x(t) e-jkω0t = ∫ Σ ak. ejkω0t e-jkω0t = Σ ∫ak[ejk(k-n)ω0tdt]
0 0 -∞ -∞ 0

Euler’s formula:
T T
∫ak[ejk(k-n)ω0tdt] = ∫ [cos(k-n)ω0t + j sin(k-n)ω0t]dt
0 0

For k ≠ n cos(k-n) and sin(k-n) are periodic and for k=n cos(k-n)=1 and ans is T.
There for
T
i.e. ∫x(t) e-j(k-n)ω0t = T, k=n
0
0, k≠n
T T
-jnω0t
Then for an = 1/T ∫x(t) e dt and for a0 = 1/T ∫x(t) dt
0 0
10. Integration: ∫x(t)dt (finite value) (its periodic only if a0 = 0) has the Fourier series
coefficient ak /jkw0.

11. Conjugate symmetry for real signals: i.e. if x(t) is real then ak = *ak,
re{ ak }=re{ a-k },im{ ak }= -im{ a-k }, | ak |=| a-k | and

12. Real and even: If x(t) is real and even then their coefficient is also real and even.

13. Real and odd: If x(t) is real and odd then the coefficient is purely imaginary and
odd.

14. Decomposition of real signal: xe(t) = Ev{x(t)} [x(t) is real] then coefficient
Re{ ak }and xo(t) = Od{x(t)} [x(t) is real] then coefficient is Im{ ak }

Fourier series representation of discrete time periodic signals

Linear combination of harmonically related complex exponentials:-
A signal is periodic if x[n] = x(n + mN) for all n where N is a constant known as
period and m = 0, 1, 2, … and N = ω0/2π. E.g. x[n] = cosω0t, x[n] = ejω0n
Complex exponential
Фk[n] = eikω0n = ejk(2π/N)n k= 0, ±1, ±2…
Фk[n] = Фk+N[n]
i.e. Ф0[n] = ФN[n], Ф1[n] = ФN+1[n]
x[0] = Σ ak., x[1] = Σ ak. ejk(2π/N)
k=<N> k=<N>

i.e. x[n] = Σ ak. ejkω0n = Σ ak.ejk(2π/N)n

k=<N> k=<N>

if k = 0, 1, …. N -1 or K=3, 4, …, N+2 etc…

K= N is Nth harmonic component or N successive integers.
This representation is known as Fourier series representation of periodic signal of
discrete type.
To determine ak multiply both with e-jrω0n
i.e. x[n] e-jrω0n = Σ ak. ejkω0n e-jrω0n = Σ ak.ejk(k-r)(2π/N)n
k<n> k=<n>

There for inner most sum if k = r is N and if K ≠ r is 0.

ar = 1/N Σ x[n] e-jrω0n where ω0 = 2π/N
k=<N>

ak = 1/N Σ x[n] e-jkω0n where ω0 = 2π/N

k=<N>
13. Real and odd: If x[n] is real and odd then the coefficient is purely imaginary and
odd.

14. Decomposition of real signal: xe[n] = Ev{x[n]} [x[n] is real] then coefficient
Re{ ak }and xo[n] = Od{x[n]} [x[n] is real] then coefficient is Im{ ak }

Paseval’s relation:

1. Continuous time periodic signal:

∞
1/T ∫ |x(t}|2 dt = Σ | ak |2 where ak is Fourier series coefficient and T is the time
T k=-∞

period of the signal.

I.e. average power or energy per unit time in one period.
Total paseval’s relation = sum of the average power in all harmonic components.

2. Discrete time periodic signal:

1/N Σ |x[n]|2 = Σ | ak |2 where ak is Fourier series coefficient and N is the time

n=<N> k=<N>

period of the signal.

I.e. | ak | = average power or energy per unit time in one period.
Total paseval’s relation = sum of the average power in all harmonic components.

Fourier transforms representation of periodic signal

Representation of a-periodic signal:

x(t) = 1 |t| <T

0, T1 < |t| < T/2 period T
Formula for Fourier transform

if x(t) is a signal:
∞
X(jω) = ∫ x(t). e-jωt dt
-∞

Formula for inverse Fourier transform

if X(jω) is the transformed signal:

∞
x(t) = 1/2π ∫ X(jω). e-jωt dω
-∞
 Fourier transforms representation of discrete time periodic signals
Representation of Fourier Transform: -
∞
X[ejω] = Σ x[n] e-jωn
N=-∞

Representation of inverse Fourier transform: -

x[n] = 1/2π ∫ X(ejω). ejωn dω
2π

Properties of continuous time Fourier transforms: -

1. Linearity: Let x[n] and y[n] has the same period N and let their transform be
x[n] ↔ X[ejω] and y[n] ↔ Y[ejω] then
z[n] = Ax[n] + By[n] ↔ A Y[ejω] + B Y[ejω] with period N
2. Time shifting:x[n] ↔ X[ejω] then x[t-t0] ↔ ejωno X[ejω]
3. Time scaling : x[n] ↔ X[ejω] then x[αt] ↔ 1/|α| X[ejω /α]
4. Multiplication: x[n] ↔ X[ejω] (with period N), y[n] ↔ Y[ejω] (with period N)
then
x[n].y[n] ↔ 1/2π ∫ X[ejθ].Y[ej(ω-θ)]
2π

5. Conjugate and conjugate symmetry: x[n] ↔ X[ejω] (with period N), then
x*[n] ↔ X*[e-jω] (with period T) where ‘*’ represents complex conjugate.
6. Frequency shifting: i.e. multiplication with e-jω0n => x[n]. e-jω0n the Fourier
transforms will be X[ej(ω-ω0)]
7. Convolution: If y[n] = x[n] * h[t] (where * represents convolution) then
Y[ejω]= X[ejω] ×H[ejω].
8. Difference in time: x[n]-x[n-1] ↔ (1- ejω)X(ejω)

9. Difference in frequency: nx[n] ↔ jdX[ejω]/dω

N ∞
-jω jω jθ
10. Accumulation: Σ x[k] ↔ 1/(1- e )X[e ] + πX[e ] Σ δ[ω-2πk]
k=-∞ k=-∞

11. Conjugate symmetry for real signals:

if x[n] is real then
Y[ejω] = X*[ e-jω], re{Y[ejω]}=re{ X[e-jω] },im{ Y[ejω] }= -im{ X[e-jω]},
| Y[ejω] |=| X[e-jω] | and
12. Real and even: If x[n] is real and even then X[ejω] is also real and even.
13. Real and odd: If x[n] is real and odd then X[ejω] is purely imaginary and odd.
14. Decomposition of real signal: xe[t] = Ev{x[n]} (x[n] is real) then
Fs=Re{ X[ejω] }and xo[t] = Od{x[n]} (x[n] is real) then Fs= Im{ Y[ejω] }

15. Time reversal: x[-t] ↔ X[e-jω]

Parseval’s relation:
If x[n] and X[ejω] are Fourier transform pair then ,
∞
Σ | x[n] |2 = 1/2π ∫ | X[ejω]|2 dω
n=-∞ 2π
Duality:
For discrete time signal the Fourier transform and its inverse are more or just same:
x[n] = 1/2π ∫ X[ejω] ejωn dω
2π
∞
X[e ] = Σ x[n] e-jωn
jω
n=-∞

Similarly for continuous time:

∞
X(jω) = ∫ x(t). e-jωt dt
-∞

∞
x(t) = 1/2π ∫ X(jω). e-jωt dω
-∞

Sampling theorem:
It states that the sampling frequency should be more than or equal to twice the
max frequency component of the message signal (base band signal)
fs ≥ 2 f m
System characterized by linear constant coefficient differential equation:
H[ejω] = Y[ejω]/ X[ejω]
where, x(t) is the input (X[ejω] its Fourier transform)
y(t) is the output (Y[ejω] its Fourier transform) and
h(t) is the impulse response of the system (H[ejω] its Fourier transform).

Fourier transform pairs:

x(t) X f (ω)
Z
x(t) x(t) e−jωt dt (def.)
Z
1
X f (ω) ejωt dω X f (ω)
2π

δ(t) 1

1 2 π δ(ω)

1
u(t) π δ(ω) +
jω

ejωo t 2 π δ(ω − ωo )

cos(ωo t) π δ(ω + ωo ) + π δ(ω − ωo )

sin(ωo t) j π δ(ω + ωo ) − j π δ(ω − ωo )

ωo “ω ”
o
sinc t ideal LPF
π π
cut-off frequency ωo
„ «
2 T
symmetric pulse sin ω
ω 2
width T , height 1

impulse train impulse train

2π
period T , height 1 period, height ωo =
T
 Laplace transform

Laplace transforms definition:

Laplace transforms:
∞
L{x(t)} = ∫ x(t). e-st dt = X(s) where s is complex variable and s = σ + jω
-∞

Inverse Laplace transforms:

t
L-1{X(s)} = 1/2π ∫ X(s). est ds = x(t)
0

Laplace transforms exist only if ∫ x(t). e-st dt exist. I.e. |∫ x(t). e-st dt| < ∞

Initial and final value theorem:

If x(t) = 0 for t < 0 and x(t) contains no impulse or higher order singularities at
t = 0 and let X(s) be the Laplace transform of x(t) then,
x(0+) = Lims->∞ SX(s) {Initial value theorem }
Limt->∞x(t) = Lims->0 SX(s) {Final value theorem }

ROC of Laplace transform:

ROC (region of convergence) of Laplace transform is the region in the x-y plane
where Laplace transform is ROC.
For e.g. X(s) = 1/s+a is Valid if s > -a in the x- y plane. If there are more than one root
the overlapping area is the ROC.
For e.g. X(s) = 1/((s+a)(s+b)) where |a| > |b| say then,
Poles and zeros:
Poles are the roots of the denominator of the fraction in the Laplace transform
and zeroes are the roots of numerator of the fraction. For example as follows: -
                   
  
                    
where a1, a2, …, an are known as zeroes and b1, b2, …. , bn are known as poles.
They are represented in a pole zero diagram as

Properties of Laplace transform:

If x(t), x1(t), and x2(t) are three signals and X(s), X1(s), X2(s) are their Laplace
transform respectively and a, b are some constant then,

1. Linearity: ax1(t) + bx2(t) = aX1(s) + bX2(s)

2. Time shifting: x(t-t0) = e-sto.X(s)
3. Shifting in s domain: estox(t) = X(s-s0)
4. Time scaling: x(at) = 1/|a| X(s/a)
5. Conjugation: x*(t) = X*(s*)
6. Convolution: x1(t)*x2(t) = X1(s).X2(s)
7. Differential in time domain: dx(t)/dt = sX(s)
8. Differential in frequency domain: -tx(t) = d/ds X(s)
9. Integration in time domain: Integration in time domain is division in frequency
domain.
t
i.e. ∫ x(τ)d(τ) = 1/s . X(s)
-∞

Applications of Laplace transform:

• For a system with a rational system function causality of the system is equivalent
to the ROC being to the right half plane to the right of the right most pole.
• An LTI system is stable if and only if the ROC of its system function H(s) include
the entire jω axis. (i.e. Re(s) = 0)
Laplace transform pairs:

x(t) X(s) ROC

Z
x(t) x(t) e−st dt (def.)

δ(t) 1 all s

1
u(t) Re(s) > 0
s
1
e−a t u(t) Re(s) > −a ; a is real.
s+a
s
cos(ωo t) u(t) Re(s) > 0
s2 + ωo2
ωo
sin(ωo t) u(t) Re(s) > 0
s2 + ωo2
s+a
e−a t cos(ωo t) u(t) Re(s) > −a
(s + a)2 + ωo2
ωo
e−a t sin(ωo t) u(t) Re(s) > −a
(s + a)2 + ωo2

Complex arithmetic operations:

 Z-transform and discrete Fourier transform
Z-transform definition:

If x(t) is the signal then its z transform X(z) is defined as follows: -

∞
X(z) = Σ x[n]. z-n
n=-∞

Note: Every transform is of the form ∫x(t).k(s,t)dt.

For example of discrete Fourier transforms
∞ ∞
X(rejω) = Σ x[n].(r.ejω)-n = Σ x[n].e-jωn.r-n
n=-∞ n=-∞
∞
If r = 1 then, X(rejω) = Σ x[n].e-jωn
n=-∞

ROC of Z transform:

ROC (region of convergence) of Z transform is the region in the x-y plane where Z
transform is valid. It dose not have any pole. It is a ring in z- plane centered about
origin. If x[n] is of finite duration then ROC in the entire Z plane except possibly z =
0 and/or z = ∞.
For e.g. X(z) = 1/z+a is Valid if z > -a and we draw a circle with center origin and
radius |a| in the z plane and if the transform is valid for values greater than a then the
ROC is exterior of the circle and if less than a then interior of the circle.. If there are
more than one root the overlapping area is the ROC.
For e.g. X(z) = 1/((z+a)(z+b)) where |a| < |b| say then,

Initial value theorem:

If x[n] = 0 for n < 0 then x[0] = Limz->∞ X(z)