Chapter one

Introduction to digital signal processing:

1
What is a Signal? What is a System?
• Signal:
 any physical quantity that varies with time, space, or any other independent variable or
variables
 Examples: pressure as a function of altitude, sound as a function of time, color as a function
of space, . . .
.
• System:
 a physical device that performs an operation on a signal
 Exemples: analog amplifier, noise canceler, communication channel, transistor, . .
.

•Independent Variable
• A signal can be represented as a function x(t) and consists of:
1.one or more dependent variable components (e.g., air pressure x, R-G-B color [x1 x2
x3]T );
2.one or more independent variables (e.g., time t, 3-D special location (s1; s2; s3)).
• Please note: in this course we will typically use time t to represent the independent
variable although in general it can correspond to any other type of independent
2
variable.
• Continuous-Time versus Discrete-Time Signals
• Continuous-Time Signals: signal is defied for every value of time in a given
interval (a; b) where a >= -∞1 and b<= ∞.
 Examples: voltage as a function of time, height as a function of pressure,
number of positron emissions as a function of time.

• Discrete-Time Signals: signal is defied only for certain specific values of

time; typically taken to be equally spaced points in an interval.
 Examples: number of stocks traded per day, average income per
province.

3
• Continuous-Amplitude versus Discrete-Amplitude
• Continuous-Amplitude Signals: signal amplitude takes on a spectrum of
values within one or more intervals
 Examples: color, temperature, pain-level

• Discrete-Amplitude Signals: signal amplitude takes on values from a finite

set
 Examples: digital image, population of a country

4
• Analog and Digital Signals
 analog signal = continuous-time + continuous amplitude
 digital signal = discrete-time + discrete amplitude

• Analog signals are fundamentally significant because we must interface with

the real world which is analog by nature.
• Digital signals are important because they facilitate the use of digital signal
processing (DSP) systems, which have practical and performance advantages
for several applications.
5
• Analog and Digital Systems
• analog system = analog signal input + analog signal output
 advantages: easy to interface to real world, do not need A/D or D/A converters, speed not
dependent on clock rate
• digital system = digital signal input + digital signal output
 advantages: re-configurability using software, greater control over accuracy/resolution,
predictable and reproducible behavior
• Deterministic vs. Random Signals
• Deterministic signal:
 any signal that can be uniquely described by an explicit mathematical expression, a table
of data, or a well-defined rule
 past, present and future values of the signal are known precisely without any uncertainty
 Random signal:
 any signal that lacks a unique and explicit mathematical expression and thus evolves in
time in an unpredictable manner
 it may not be possible to accurately describe the signal
 the deterministic model of the signal may be too complicated to be of use

6
• What is a “pure frequency” signal?

• What is a “pure frequency” signal?

• Xa(t) = A cos(Ωt + θ) = A cos(2πFt + θ)t t ∈ R
• analog signal, ∵ −A ≤ xa(t) ≤ A and −∞ < t < ∞
• A = amplitude
• Ω = frequency in rad/s
• F = frequency in Hz (or cycles/s); note: Ω = 2πF
• θ = phase in rad

7
• Continuous-time Sinusoids
• Xa(t) = A cos(Ωt + θ) = A cos(2πFt + θ), t ∈ R
1. for F ∈ R, xa(t) is periodic
 i.e., there exists Tp ∈ R+ such that Xa(t) = Xa(t + Tp)
2. distinct frequencies result in distinct sinusoids
 i.e., for F1 ≠ F2, A cos(2πF1t + θ) ≠ A cos(2πF2t + θ)
3. increasing frequency results in an increase in the rate of oscillation of the sinusoid
 i.e., for |F1| < |F2|, A cos(2πF1t + θ) has a lower rate of oscillation than A
cos(2πF2t + θ)
 Continuous-time Sinusoids: Frequency

8
• Discrete-time Sinusoids
• x(n) = A cos(ωn + θ) = A cos(2πfn + θ), n ∈ Z
• discrete-time signal (not digital), ∵ −A ≤ xa(t) ≤ A and n ∈ Z
• A = amplitude
• ω = frequency in rad/sample
• f = frequency in cycles/sample; note: ω = 2πf
• θ = phase in rad
• Discrete-time Sinusoids
• x(n) = A cos(ωn + θ) = A cos(2πfn + θ), n ∈ Z
1. x(n) is periodic only if its frequency f is a rational number
 Note: rational number is of the form k1/k2 for k1, k2 ∈ Z
 periodic discrete-time sinusoids: x(n) = 2 cos( (4/7) πn), x(n) = sin(−(π/5) n
+ √3)
 aperiodic discrete-time sinusoids: x(n) = 2 cos( 4/7 n), x(n) = sin((2πn + √3)

9
• x(n) = A cos(ωn + θ) = A cos(2πfn + θ), n ∈ Z
2. radian frequencies separated by an integer multiple of 2π are identical
• or cyclic frequencies separated by an integer multiple are identical
3. lowest rate of oscillation is achieved for ω = 2kπ and highest rate of oscillation
is achieved for ω = (2k + 1)π, for k ∈ Z
• subsequently, this corresponds to lowest rate for f = k (integer) and highest
rate for f = (2k+1)/2 (half integer), for k ∈ Z.

10
• Complex Exponentials

• Periodicity: Continuous-time Periodicity: Discrete-time

11
•.

12
•.

13
• Uniqueness: Continuous-time
• For F1 ≠ F2, A cos(2πF1t + θ) ≠ A cos(2πF2t + θ) except at discrete points in
time.
• Uniqueness: Discrete-time: Let f1 = f0 + k where k ∈ Z,

14
• Uniqueness: Discrete-time
• Therefore, dst-time sinusoids are unique for f ∈ [0, 1).
• For any sinusoid with f1 not∈ [0, 1), ∃ f0 ∈ [0, 1) such that

• Example: A dst-time sinusoid with frequency f1 = 4.56 is the same as a dst-

time sinusoid with frequency f0 = 4.56 − 4 = 0.56.
• Example: A dst-time sinusoid with frequency f1 = − 7/8 is the same as a dst-
time sinusoid with frequency f0 = − 7/8 + 1 = 1/8 .

15
• Harmonically Related Complex Exponentials

16
• Harmonically Related Complex Exponentials
• What does the family of harmonically related sinusoids sk (t) have in
common?

17
• Analog-to-Digital Conversion

• Sampling:
– conversion from cts-time to dst-time by taking “samples” at discrete time
instants
– E.g., uniform sampling: x(n) = xa(nT) where T is the sampling period and n ∈ Z
• Quantization:
– conversion from dst-time cts-valued signal to a dst-time dst-valued signal
– quantization error: eq(n) = xq(n) − x(n) for all n ∈ Z
• Coder

18
• Sampling Theorem
• If the highest frequency contained in an analog signal Xa(t) is Fmax = B and the
signal is sampled at a rate
• Fs > 2Fmax = 2B
• then Xa(t) can be exactly recovered from its sample values using the
interpolation function

• Note: FN = 2B = 2Fmax is called the Nyquist rate

• Therefore, given the interpolation relation, Xa(t) can be written as

19
• Bandlimited Interpolation

20
• Digital-to-Analog Conversion

• Common interpolation approaches: bandlimited interpolation, zero-order

hold, linear interpolation, higher-order interpolation techniques, e.g., using
splines
• In practice, “cheap” interpolation along with a smoothing filter is employed.

• Common interpolation approaches: bandlimited interpolation, zero-order

hold, linear interpolation, higher-order interpolation techniques, e.g., using
splines
• In practice, “cheap” interpolation along with a smoothing filter is employed.
21
• Digital-to-Analog Conversion

• Common interpolation approaches: bandlimited interpolation, zero-order

hold, linear interpolation, higher-order interpolation techniques, e.g., using
splines
• In practice, “cheap” interpolation along with a smoothing filter is employed.

22