Signals Sampling Theorem
Statement: A continuous time signal can be represented in its samples and can be recovered back
when sampling frequency fs is greater than or equal to the twice the highest frequency component
of message signal. i. e.
f
fs ≥ 2f m ..
s ≥ 2fm
Proof: Consider a continuous time signal x(t). The spectrum of x(t) is a band limited to fm Hz i.e.
the spectrum of x(t) is zero for |ω|>ωm.
Sampling of input signal x(t) can be obtained by multiplying x(t) with an impulse train δ(t) of period
Ts. The output of multiplier is a discrete signal called sampled signal which is represented with y(t)
in the following diagrams:
Here, you can observe that the sampled signal takes the period of impulse. The process of
sampling can be explained by the following mathematical expression:
Sampled
Sampled signal
signal y
y((t
t)) =
= x
x((t
t)).. δ
δ((t
t)) .. .. .. .. .. .. (
(11)
)
The trigonometric Fourier series representation of δ
δ (t) is given by
∞
∞
δ
δ((t
t)) =
= a
a0 + Σ
0 + Σ
(
(aan cos nω s t
n cos nωs t++b
bn sin nω s t
n sin nωs t)) .. .. .. .. .. .. (
(22)
)
n
n==1
1
T
T
Where
1
1 2
2 1
1 1
1
a
a0 =
0 =
∫
∫− δ
δ((t
t))d
dtt =
= δ
δ((0
0)) =
=
T −T
T T T
Tss Tss Tss
2
2
T
T
2
2 2
2 2
2 2
2
a
an =
n =
∫
∫− δ
δ((t
t)) cos
cos n
nωωs dt =
s dt =
δ
δ((0
0)) cos
cos n
nωωs 0 =
s0 =
T −T
T T T
Tss T22 T
2
2
T
T
2
2 2
2 2
2
b
bn =
= ∫
∫− δ
δ((t
t)) sin
sin n
nωωs t
tddt
t =
= δ
δ((0
0)) sin
sin n
nωωs 0
0 =
= 0
0
n T −T
T s T s
Tss Tss
2
2
Substitute above values in equation 2.
1
1 ∞
∞ 2
2
∴
∴ δ
δ((t
t)) =
= +
+ΣΣ (
( cos
cos n
nωωs t + 0)
s t + 0)
T
Ts n
n==1
1 T
Ts
s s
Substitute δ(t) in equation 1.
→
→ y
y((t
t)) =
= x
x((t
t)).. δ
δ((t
t))
1
1 ∞
∞ 2
2
=
= x
x((t
t))[[ +
+ΣΣ (
( cos
cos n
nωωs t)]
s t)]
T
Ts n
n==1
1 T
Ts
s s
1
1 ∞
∞
=
= [[x
x((t
t))+
+22Σ
Σ (
(cos
cos n
nωωs t)x(t)]
s t)x(t)]
T
Ts n
n==1
1
s
1
1
y
y((t
t)) =
= [[x
x((t
t))+
+22 cos
cos ω
ωs t. x(t) + 2 cos 2ω s t
s t. x(t) + 2 cos 2ωs t.. x
x((t
t))+
+22 cos
cos 3
3ωωs t. x(t) . . . . . . ]
s t. x(t) . . . . . . ]
T
Tss
Take Fourier transform on both sides.
1
1
Y
Y((ω
ω)) =
= [[X
X((ω
ω))+
+XX(
(ωω−
−ωωs ) + X(ω + ω s )
s ) + X(ω + ωs )++X
X((ω
ω−−2
2ωωs ) + X(ω + 2ω s )
s ) + X(ω + 2ωs )++ .. .. .. ]]
T
Tss
1
1 ∞
∞
∴
∴ Y
Y((ω
ω)) =
= Σ
Σn=−∞ X
X((ω
ω−−n
nωωs
s)
) w
whhe
erre
e n
n =
= 0
0,, ±
±11,, ±
±22,, .. .. ..
T
Ts n=−∞
s
To reconstruct x(t), you must recover input signal spectrum X(ω) from sampled signal spectrum
Y(ω), which is possible when there is no overlapping between the cycles of Y(ω).
Possibility of sampled frequency spectrum with different conditions is given by the following
diagrams:
Aliasing Effect
The overlapped region in case of under sampling represents aliasing effect, which can be removed
by
considering fs >2fm
By using anti aliasing filters.
Signals Sampling Techniques
There are three types of sampling techniques:
Impulse sampling.
Natural sampling.
Flat Top sampling.
Impulse Sampling
Impulse sampling can be performed by multiplying input signal x(t) with impulse train
Σ
∞
∞
Σn=−∞ δ
n=−∞
δ((t
t−−n
nTT)
) of period 'T'. Here, the amplitude of impulse changes with respect to
amplitude of input signal x(t). The output of sampler is given by
y
y((t
t)) =
= x
x((t
t))×
× impulse train
∞
∞
=
= x
x((t
t))×
×ΣΣn=−∞ δ
δ((t
t−−n
nTT)
)
n=−∞
∞
∞
y
y((t
t)) =
= y
yδ (t) = Σ x(nt)δ(t − nT ) . . . . . . 1
δ (t) = Σn=−∞ x(nt)δ(t − nT ) . . . . . . 1
n=−∞
To get the spectrum of sampled signal, consider Fourier transform of equation 1 on both sides
1
1 ∞
∞
Y
Y((ω
ω)) =
= Σ
Σn=−∞ X
X((ω
ω−−n
nωωs
s)
)
T
T n=−∞
This is called ideal sampling or impulse sampling. You cannot use this practically because pulse
width cannot be zero and the generation of impulse train is not possible practically.
Natural Sampling
Natural sampling is similar to impulse sampling, except the impulse train is replaced by pulse train
of period T. i.e. you multiply input signal x(t) to pulse train Σ
∞
∞
Σn=−∞ P
n=−∞
P((t
t−−n
nTT)
) as shown below
The output of sampler is
y
y((t
t)) =
= x
x((t
t))×
× pulse
pulse train
train
=
= x
x((t
t))×
×pp(
(tt)
)
∞
∞
=
= x
x((t
t))×
×ΣΣn=−∞ P
P((t
t−−n
nTT)
) .. .. .. .. .. .. (
(11)
)
n=−∞
The exponential Fourier series representation of p(t) can be given as
∞
∞ j
jnnω
ωs t
st
p
p((t
t)) =
= Σ
Σn=−∞ F
Fn e
ne
.. .. .. .. .. .. (
(22)
)
n=−∞
∞
∞ j
j22π
πnnf
fs tt
=
= Σ
Σn=−∞ F
Fn e
ne
s
n=−∞
T
T
1
1 −
−jjn
nω t
Where Fn
n
=
T
∫−
−TT
2
2
p(t)e
ωs
st
dt
T
2
2
1
1
=
= (
(nnω
ωs )
s)
T
TPP
Substitute Fn value in equation 2
∞
∞ 1
1 j
jnnω
ωs t
st
∴
∴ p
p((t
t)) =
= Σ
Σn=−∞ P
P((n
nωωs )e
s )e
n=−∞ T
T
1
1 ∞
∞ j
jnnω
ωs t
st
=
= Σ
Σn=−∞ P
P((n
nωωs )e
s )e
T
T n=−∞
Substitute p(t) in equation 1
y
y((t
t)) =
= x
x((t
t))×
×pp(
(tt)
)
1
1 ∞
∞ j
jnnω
ωs t
st
=
= x
x((t
t))×
× Σ
Σn=−∞ P
P((n
nωωs )e
s) e
T
T n=−∞
1
1 ∞
∞ j
jnnω
ωs t
st
y
y((t
t)) =
= Σn=−∞ P
Σ P((n
nωωs )
)xx(
(tt)
)ee
T n=−∞ s
T
To get the spectrum of sampled signal, consider the Fourier transform on both sides.
1
1 ∞
∞ j
jnnω
ωs t
st
F
F .. T
T [[y
y((t
t))]] =
= F
F .. T
T [[ Σ
Σn=−∞ P
P((n
nωωs ) x(t) e
s ) x(t) e
]]
T
T n=−∞
1
1 ∞
∞ j
jnnω
ωs t
t
=
= Σ
Σn=−∞ P
P((n
nωωs ) F . T [x(t) e
s ) F . T [x(t) e
s
]]
T
T n=−∞
According to frequency shifting property
j
jnnω
ωs t
st
F
F .. T
T [[x
x((t
t))e
e ]] =
= X
X[[ω
ω−−n
nωωs
s]
]
1
1 ∞
∞
∴
∴ Y
Y [[ω
ω]] =
= Σ
Σn=−∞ P
P((n
nωωs ) X[ω − nω s ]]
s ) X[ω − nωs
T
T n=−∞
Flat Top Sampling
During transmission, noise is introduced at top of the transmission pulse which can be easily
removed if the pulse is in the form of flat top. Here, the top of the samples are flat i.e. they have
constant amplitude. Hence, it is called as flat top sampling or practical sampling. Flat top sampling
makes use of sample and hold circuit.
Theoretically, the sampled signal can be obtained by convolution of rectangular pulse p(t) with
ideally sampled signal say yδ(t) as shown in the diagram:
i.e. y
y((t
t)) =
= p
p((t
t))×
×yyδ
δ
(
(tt)
) .. .. .. .. .. .. (
(11)
)
To get the sampled spectrum, consider Fourier transform on both sides for equation 1
Y
Y [[ω
ω]] =
= F
F .. T
T [[P
P((t
t))×
×yyδ (t)]
δ (t)]
By the knowledge of convolution property,
Y
Y [[ω
ω]] =
= P
P((ω
ω))Y
Yδ (ω)
δ (ω)
Here P
P((ω
ω)) =
= T
TSSa
a((
ω
ωT
2
2
T
)
) =
= 2
2 sin
sin ω
ωTT/
/ωω
Nyquist Rate
It is the minimum sampling rate at which signal can be converted into samples and can be
recovered back without distortion.
Nyquist rate fN = 2fm hz
Nyquist interval = f
1
1
= 1
1
seconds.
fNN 2
2ffm
m
Samplings of Band Pass Signals
In case of band pass signals, the spectrum of band pass signal X[ω] = 0 for the frequencies outside
the range f1 ≤ f ≤ f2. The frequency f1 is always greater than zero. Plus, there is no aliasing effect
when fs > 2f2. But it has two disadvantages:
The sampling rate is large in proportion with f2. This has practical limitations.
The sampled signal spectrum has spectral gaps.
To overcome this, the band pass theorem states that the input signal x(t) can be converted into its
samples and can be recovered back without distortion when sampling frequency fs < 2f2.
Also,
1
1 2
2ff2
2
f
fs =
s =
=
=
T
T m
m
f
f2
Where m is the largest integer < B
B
2
and B is the bandwidth of the signal. If f2=KB, then
1
1 2
2KKB
B
f
fs =
s =
=
=
T
T m
m
For band pass signals of bandwidth 2fm and the minimum sampling rate fs= 2 B = 4fm,
the spectrum of sampled signal is given by Y
Y [[ω
ω]] =
=
1
1
T
Σ
∞
∞
Σn=−∞ X
n=−∞
X[[ω
ω−−2
2nnB
B]]
T
