Sampling Process

⦁ Sampling is defined as, “The process of measuring the instantaneous values of

continuous-time signal in a discrete form.”
⦁ Sample is a piece of data taken from the whole data which is continuous in the time
domain.

The sampling process:(a)Analog signal,(b) Instantaneously sampled version of the analog signal
1 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT
Sampling Process
Let 𝑔(𝑡) be an arbitrary signal of finite energy as shown in figure above, sampled at
uniform rate‘𝑇𝑠 ’seconds denoted by 𝑔(𝑛𝑇𝑠 ) where‘n’is an integer
Sampling Rate
⦁ To discretize the signals,the gap between the samples should be fixed.
⦁ That gap can be termed as a sampling periodTs.

Sampling frequency
⦁ Sampling frequency fs is the reciprocal of the sampling period.
⦁ This sampling frequency,can be simply called as Sampling rate.
⦁ The sampling rate denotes the number of samples taken per second, or for a finite set
of values.
2 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT
Sampling Process
Representation of 𝒈 𝒕 in terms of its samples:

⦁ This is done in 4 steps

1. Define the sampled version 𝑔𝛿 𝑡

2. Find the Fourier transform of 𝑔𝛿 𝑡 i.e., 𝐺𝛿 𝑓

3. Find the relationship between 𝐺 𝑓 and 𝐺𝛿 𝑓

4. Find the relationship between 𝑔 𝑡 and 𝑔(𝑛𝑇𝑠 )

3 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
1. Define the sampled version 𝒈𝜹 𝒕 Time domain
𝑔𝛿 𝑡 = 𝑔 𝑡 × 𝛿 𝑡
⦁ Impulse train is represented by 𝛿 𝑡 = ∑∞
𝑛 =−∞ 𝛿(𝑡 − 𝑛𝑇𝑠)
𝑔𝑡
⦁ Where 𝛿(𝑡 − 𝑛𝑇𝑠) represents a delta function positioned at time t = 𝑛𝑇𝑠

⦁ Signal obtained after sampling

∞
𝛿𝑡
𝑔𝛿 𝑡 = ∑ 𝑔(𝑡)𝛿(𝑡 − 𝑛𝑇𝑠)
𝑛=−∞
∞

𝑔𝛿 𝑡 = ∑ 𝑔(𝑛𝑇𝑠)𝛿(𝑡 − 𝑛𝑇𝑠) 𝑔𝛿 𝑡
𝑛=−∞

⦁ 𝑔(𝑛𝑇𝑠) is the instantaneous amplitude of 𝑔(𝑡) at instant t = 𝑇𝑠

4 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
2. Find the Fourier transform of 𝒈𝜹 𝒕 i.e., 𝑮𝜹 𝒇

⦁ Taking Fourier transform

∞

𝐺𝛿 𝑓 = 𝐹. 𝑇 ∑ 𝑔(𝑡)𝛿(𝑡 − 𝑛𝑇𝑠)
𝑛=−∞

⦁ Product in time domain = Convolution in frequency domain

∞

𝐺𝛿 𝑓 = 𝐹. 𝑇 𝑔(𝑡) ∗ 𝐹. 𝑇 ∑ 𝛿(𝑡 − 𝑛𝑇𝑠)

𝑛=−∞
∞

𝐺𝛿 𝑓 = 𝐺(𝑓) ∗ 𝐹. 𝑇 ∑ 𝛿(𝑡 − 𝑛𝑇𝑠)

𝑛=−∞

5 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
⦁ FourierTransform of the Impulse train 𝛿(𝑡 − 𝑛𝑇𝑠) is,
Frequency domain
∞ ∞

𝐹. 𝑇 ∑ 𝛿(𝑡 − 𝑛𝑇𝑠) = 𝑓𝑠 ∑ 𝐺(𝑓 − 𝑛𝑓𝑠) 𝐺𝛿 𝑓 = 𝐺(𝑓) ∗ 𝛿(𝑓)

𝑛=−∞ 𝑛=−∞
𝐺(𝑓)
1
⦁ Where 𝑓𝑠 =
𝑇𝑠

⦁ Therefore, Fourier Transform of the sampled version 𝑔𝛿 𝑡 is, 𝛿(𝑓)

∞ ∞

𝐺𝛿 𝑓 = 𝐺 𝑓 ∗ 𝑓𝑠 ∑ 𝐺 𝑓 − 𝑛𝑓𝑠 = 𝑓𝑠 ∑ 𝐺(𝑓) ∗ 𝐺(𝑓 − 𝑛𝑓𝑠)

𝑛=−∞ 𝑛=−∞
∞
𝐺𝛿 𝑓
𝑮𝜹 𝒇 = 𝒇𝒔 ∑ 𝑮(𝒇 − 𝒏𝒇𝒔)
𝒏=−∞

6 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
⦁ 𝐺 𝑓 is the Fourier transform of the original signal 𝑔 𝑡
⦁ 𝐺 𝑓 − 𝑛𝑓𝑠 = 𝐺(𝑓) at 𝑓 = 0, ±𝑓𝑠 , ±2𝑓𝑠 , ±3𝑓𝑠 …

⦁ That is same spectrum appears at 𝑓 = … , −3𝑓𝑠, −2𝑓𝑠 , −𝑓𝑠, 0, +𝑓𝑠, +2𝑓𝑠, +3𝑓𝑠…

⦁ Which means a periodic spectrum with time period equal to 𝑓𝑠 is generated in frequency
domain because of sampling 𝑔 𝑡 in time domain
𝐺𝛿 𝑓 =. . +𝑓𝑠 𝐺 𝑓 − 2𝑓𝑠 + 𝑓𝑠 𝐺 𝑓 − 𝑓𝑠 + 𝒇𝒔 𝑮 𝒇 + 𝑓𝑠 𝐺 𝑓 + 𝑓𝑠 + 𝑓𝑠 𝐺 𝑓 + 2𝑓𝑠 + 𝑓𝑠 𝐺 𝑓 + 3𝑓𝑠 +. .

𝐺𝛿 𝑓

7 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
⦁ The Fourier transformed version can be written as,
∞

𝑮𝜹 𝒇 = 𝒇𝒔𝑮(𝒇) +𝒇𝒔 ∑ 𝑮(𝒇 − 𝒏𝒇𝒔)

𝒏=−∞
𝒏≠𝟎

⦁ The first term 𝒇𝒔𝑮(𝒇) indicates the spectrum without sampling

⦁ The second term +𝒇𝒔 ∑𝒏=−∞
∞
𝑮(𝒇 − 𝒏𝒇𝒔) indicates the spectrum repeating at multiple
𝒏≠𝟎
frequencies of sampling frequency 𝒇𝒔

8 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process

⦁ Fourier transform of continuous time signal

∞

𝐹𝑇 𝑔(𝑡) = ∫ 𝑔(𝑡)𝑒−2𝜋𝑓𝑡 𝑑𝑡
−∞

⦁ Fourier transform of discontinuous time signal

∞

𝐹. 𝑇 𝑔𝛿 𝑡 = ∑ 𝑔(𝑛𝑇𝑠)𝑒−𝑗2𝜋𝑓𝑛𝑇𝑠
𝑛=−∞

9 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
⦁ If the signal g t is bandlimited such that there is no frequency components are available
greater than the available bandwidth “W ” Hz then
𝐺 𝑓 = 0 𝑓𝑜𝑟 𝑓 > 𝑊
1
⦁ let 𝑓𝑠 = 2𝑊 or 𝑇𝑠 = 2𝑊 ,whereW is the maximum frequency

Spectrum of a strictly
band-limited signal 𝑔(𝑡)

Spectrum of the sampled version of

𝑔 𝑡 for a sampling period T𝑠 = 1/2𝑊
10 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT
Sampling Process
3. Find the relationship between 𝑮 𝒇 and 𝑮𝜹 𝒇
1
⦁ Two assumptions are made: 𝐺 𝑓 = 0 𝑓𝑜𝑟 𝑓 > 𝑊 and 𝑓𝑠 = 2𝑊 or 𝑇𝑠 = 2𝑊

⦁ The equation:𝑮𝜹 𝒇 = 𝒇𝒔𝑮(𝒇) +𝒇𝒔 ∑∞

𝒏=−∞ 𝑮(𝒇 − 𝒏𝒇𝒔 ) can be written as
𝒏≠𝟎
∞

𝑓𝑠𝐺(𝑓) = 𝐺𝛿 𝑓 −𝑓𝑠 ∑ 𝐺(𝑓 − 𝑛𝑓𝑠)

𝑛=−∞
𝑛≠0
∞ ∞
1 1
𝐺 𝑓 = 𝐺𝛿 𝑓 − ∑ 𝐺 𝑓 − 𝑛𝑓𝑠 ⟹ (𝒇𝒔= 𝟐𝑾) ⟹ 𝐺𝛿 𝑓 − ∑ 𝐺 𝑓 − 𝑛𝑓𝑠
𝑓𝑠 𝑛=−∞
2𝑊 𝑛=−∞
𝑛≠0 𝑛≠0

𝟏
𝑮 𝒇 = 𝑮𝜹 𝒇 for −𝑊 ≤ 𝑓 ≤ 𝑊
𝟐𝑾

11 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
4. Find the relationship between 𝒈 𝒕 and 𝒈(𝒏𝑻𝒔)

⦁ We know that 𝐹. 𝑇 𝑔𝛿 𝑡 = 𝐺𝛿 𝑓
∞
1 1
𝐺 𝑓 = ∑ 𝑔(𝑛𝑇𝑠 )𝑒−𝑗2𝜋𝑓𝑛𝑇𝑠 we know that 𝑇𝑠 =
2𝑊 𝑛=−∞ 2𝑊

∞ ∞
1 𝑛 −𝑗2𝜋𝑓𝑛/2𝑊
1 𝑛
𝐺 𝑓 = ∑ 𝑔 𝑒 = ∑ 𝑔 𝑒 −𝑗𝜋𝑓𝑛/𝑊 𝑓𝑜𝑟 − 𝑊 ≤ 𝑓 ≤ 𝑊
2𝑊 𝑛=−∞ 2𝑊 2𝑊 𝑛=−∞ 2𝑊

⦁ 𝒈 𝒕 can be obtained by taking Inverse Fourier transform of G 𝒇

12 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling Process
⦁ We have:
∞
𝑛 1 𝑊 𝑛
𝑔 𝑡 = ∑ 𝑔( ) ∫ exp 𝑗2𝜋𝑓 𝑡 − 𝑑𝑓
2𝑊 2𝑊 −𝑊 2𝑊
𝑛=−∞

⦁ On further simplifying, we have:

∞
𝑛
𝑔 𝑡 = ∑ 𝑔 𝑠𝑖𝑛𝑐 2𝑊𝑡 − 𝑛 , −∞ < 𝑡 < ∞
2𝑊
𝑛=−∞

⦁ The above equation is also called as the desired reconstruction formula

13 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

The Sampling Theorem
⦁ The sampling theorem for strictly-band limited signals for finite energy that has no frequency
components higher than 𝑊 hertz can be stated in two parts, mentioned as:

1. It can be completely described by specifying the values of the signal instants of time
1
separated by seconds
2𝑊
1
𝑇𝑠 ≤
2𝑊
2. It can be completely recovered from a knowledge of its samples taken at the rate of 2𝑊
samples per second
𝑓𝑠≥ 2𝑊

14 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Nyquist Rate
⦁ Suppose that a signal is band-limited with no frequency components higher than W Hertz.

⦁ That means, W is the highest frequency.

⦁ For such asignal, for effective reproduction of the original signal, the sampling rate should be
twice the highest frequency.
𝑓𝑠= 2𝑊

⦁ This rate of sampling is called as Nyquist rate

LP filter
⦁ Nyquist interval
1
𝑇𝑠 =
2𝑊
Nyquist rate

15 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Limitations of Sampling Theorem
⦁ When 𝑓𝑠 < 2𝑊 in practice 𝑔 𝑡 is not strictly bandlimited so aliasing is produced.

⦁ When sampling rate is less than 2W Hz a high frequency component in the spectrum takes
the identity of a low frequency component in the spectrum.

⦁ Interference of high frequency components with that of low frequency components in the
spectrum is called Aliasing.

(a) Spectrum of a signal (b) Spectrum of an under-sampled version of the signal exhibiting the aliasing phenomenon
16 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT
Combating Aliasing
In order to combat aliasing, we may use two corrective measures:

⦁ A low-pass anti-aliasing filtercan be used prior to sampling in order to attenuate the high-
frequency components

⦁ The filtered signal is sampled at a rate slightly higher than the Nyquist rate

Note:

The use of a sampling rate higher than the Nyquist rate also has the beneficial effect of easing
the design of the reconstruction filter used to recover the original signal from its sampled version

17 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Sampling theorem and Aliasing

LP filter

Nyquist rate

aliasing

18 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling
There are three types of sampling techniques:

1. Ideal /Instantaneous /Impulse Sampling.

2. Natural /Chopper Sampling.

3. FlatTop /Rectangular Sampling.

19 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling
Ideal / Instantaneous / Impulse Sampling

⦁ Impulse sampling can be performed by multiplying input signal x(t) with impulse train
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

20 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling

⦁ To get the spectrum of sampled signal, consider Fourier transform of equation 1 on both
sides

⦁ 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.

21 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling
Natural / Chopper Sampling

⦁ Natural sampling is similar to impulse sampling,

⦁ except the impulse train is replaced by pulse train of periodT.

⦁ i.e.you multiply input signal x(t) to pulse train

22 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling

𝟏
= 𝑷(𝒏𝑚𝒔 )
𝑻

23 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling

24 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling
⦁ To get the spectrum of sampled signal, consider the Fourier transform on both sides.

25 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Types of Sampling
Flat Top / Rectangular 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.

26 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

 Types of Sampling
⦁ 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 figure below,

⦁ To get the sampled spectrum, consider Fourier transform on both sides for equation 1

27 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ Consider the analog signal 3𝑐𝑜𝑠50𝜋𝑡 + 10𝑠𝑖𝑛300𝜋𝑡 − 𝑐𝑜𝑠100𝜋𝑡 what is the Nyquist rate
for this signal?

Solution:

28 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ Consider the analog signal 3𝑐𝑜𝑠50𝜋𝑡 + 10𝑠𝑖𝑛300𝜋𝑡 − 𝑐𝑜𝑠100𝜋𝑡 what is the Nyquist rate
for this signal?

Solution:

⦁ 𝑥 𝑡 = 3𝑐𝑜𝑠50𝜋𝑡 + 10𝑠𝑖𝑛300𝜋𝑡 − 𝑐𝑜𝑠100𝜋𝑡

⦁ Nyquist rate = 2W

⦁ 2𝜋𝑓1 𝑡 = 50𝜋𝑡,𝑓1 = 25𝐻𝑧

⦁ 2𝜋𝑓2 𝑡 = 300𝜋𝑡,𝑓2 = 150𝐻𝑧 (BW of maximum frequency)

⦁ 2𝜋𝑓3 𝑡 = 100𝜋𝑡,𝑓3 = 50𝐻𝑧

⦁ Therefore, the Nyquist rate = 2W = 2 × 150𝐻𝑧 = 𝟑𝟎𝟎𝑯𝒛

29 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ 𝑚 𝑡 = 𝑠𝑖𝑛𝑐200𝜋𝑡 + 𝑠𝑖𝑛𝑐 2 200𝜋𝑡.Find the Nyquist rate.

Solution:

30 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ 𝑚 𝑡 = 𝑠𝑖𝑛𝑐200𝜋𝑡 + 𝑠𝑖𝑛𝑐 2 200𝜋𝑡.Find the Nyquist rate.

Solution:
𝑠𝑖𝑛𝜋200𝑡 𝑠𝑖𝑛 2 𝜋200𝑡
🞂 +
𝜋200𝑡 (𝜋200𝑡)2

⦁ 2𝜋𝑓1 𝑡 = 200𝜋𝑡,𝑓1 = 100𝐻𝑧

⦁ 2𝜋𝑓2 𝑡 = 400𝜋𝑡,𝑓2 = 200𝐻𝑧

⦁ Therefore, the Nyquist rate = 2W = 2 × 200𝐻𝑧 = 𝟒𝟎𝟎𝑯𝒛

31 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ Determine the sampling rate for a signal v t = 2 𝑐𝑜𝑠500𝜋𝑡 ∙ 𝑐𝑜𝑠1000𝜋𝑡

Solution:

32 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

N umerical
⦁ Determine the sampling rate for a signal v t = 2 𝑐𝑜𝑠500𝜋𝑡 ∙ 𝑐𝑜𝑠1000𝜋𝑡

Solution:

⦁ v t = 2 𝑐𝑜𝑠500𝜋𝑡 ∙ 𝑐𝑜𝑠1000𝜋𝑡
cos(500𝜋+1000𝜋)𝑡 + cos(1000𝜋−500𝜋)𝑡 cos(1500𝜋)𝑡 + cos(500𝜋)𝑡
🞂= 2 =2 =
2 2

🞂 = cos(1500𝜋)𝑡 + cos(500𝜋)𝑡

⦁ 𝑓1 = 750𝐻𝑧

⦁ 𝑓2 = 250𝐻𝑧

⦁ Therefore, the sampling rate 𝑓𝑠 ≥ 2𝑊 ≥ 2 × 750𝐻𝑧 ≥ 𝟏𝟓𝟎𝟎𝑯𝒛

33 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Quadrature Sampling of Bandpass Signals

34 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT https://pysdr.org/content/sampling.html

Quadrature Sampling of Bandpass Signals

35 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Spectrum

36 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Reconstruction of g(t)

37 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Direct sampling and Quadrature sampling

38 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT

Analog to Digital Conversion Blocks in D C S

Digital info. Bit stream Pulse waveforms

(Data bits) (baseband signals)
Textual Format
source info.
Pulse
Analog modulate
Sample Quantize Encode
info.

Sampling at rate
f s 1/T s
(sampling time=Ts)

39 Dr. T.Ilavarasan, Associate Professor,SENSE, VIT