CAN207 Continuous and Discrete Time

Signals and Systems

Lecture 24
Final Revision_Part 2

Zhao Wang & Qing Liu

[email protected]
Room SC340

1
Content

• Again: Important changes of exam rules

• Office hour arrangement
• Walk through every lecture (L14-L22)

• Please fill Module Questionnaire! Please :)

2
1. Information of final exam

Equation List on LMO 3

4
5
3. Office hour arrangement

6
4. Walk through every lecture

• The following slides are the “outline” of every

lecture;
• We will work through some after-class quiz on
LMO for each one.

7
Lecture 14 DTFS and DTFT_1
• 1. Discrete-Time Fourier Series (for periodic sequences)
– Review the concepts of eigenfunction for LTI systems
– Definition of DTFS
– Examples

• 2. Discrete-Time Fourier Transform (for general sequences)

– From DTFS to DTFT
– Definition of DTFT
– Examples (Calculation of DTFT)
– DTFT of periodic signals (optional)

8
Synthesis & Analysis Equation for DTFS and DTFT
• DTFS for a periodic discrete-time signal 𝑥[𝑛]:
2𝜋
– period: N; fundamental frequency: 𝜔0 =
𝑁
𝑁−1

𝑥[𝑛] = ෍ 𝑎𝑘 𝑒 𝑗𝑘𝜔0𝑛 = ෍ 𝑎𝑘 𝑒 𝑗𝑘𝜔0𝑛 Synthesis equation of DTFS

𝑘= 𝑁 𝑘=0
𝑁−1
1
𝑎𝑘 = ෍ 𝑥[𝑛]𝑒 −𝑗𝑘𝜔0 𝑛 = ෍ 𝑥[𝑛]𝑒 −𝑗𝑘𝜔0 𝑛 Analysis equation of DTFS
𝑁
𝑛= 𝑁 𝑛=0
𝑎𝑘 only has N distinct values, and is periodic, i.e. 𝑎𝑘 = 𝑎𝑘+𝑁
• DTFT for a general aperiodic discrete-time signal 𝑥[𝑛]:
1
𝑥[𝑛] = න 𝑋(𝑒 𝑗𝜔 )𝑒 𝑗𝜔𝑛 𝑑𝜔 Synthesis equation of DTFT
2𝜋 2𝜋
∞

𝑋(𝑒 𝑗𝜔 ) = ෍ 𝑥[𝑛]𝑒 −𝑗𝜔𝑛 Analysis equation of DTFT

𝑛=−∞
𝑋 𝑒 𝑗𝜔 , or 𝑋 𝜔 is continuous and periodic in 𝜔 (period = 2𝜋),
i.e. 𝑋 𝜔 = 𝑋 𝜔 + 2𝜋 .
9
Quiz

10
Quiz

11
Lecture 15 DTFT_2
• 1. Definition revisit
– From CTFT to DTFT
– Existence of DTFT (convergence)

• 2. DTFT properties (for easy calculation of DTFT)

– Periodicity, linearity, time-reversal, shifting, differencing in TD,
differentiation in FD...
– Convolution property
– Modulation property
– Duality

12
DTFT Properties

13
Quiz

14
Lecture 16 DTFT_3
• 1. Commonly used DTFT pairs
– Use DTFT properties when needed (Equation list provided)
• 2. Inverse DTFT
– Definition and calculation
– Partial Fraction Expansion
• 3. DTFT in LTID Systems
– Relationship between impulse response ℎ[𝑛], LCCDE and frequency
response (or transfer function, system function) H(𝜔)
– Magnitude and phase spectrum

15
Quiz

16
Inverse DTFT by PFE

• Consider a rational transform in the form

σ 𝑀 −𝑗𝑘𝜔
𝑗𝜔 𝑘=0 𝑏𝑘 𝑒
𝑋(𝑒 ) =
(1 − 𝛼1 𝑒 −𝑗𝜔 )(1 − 𝛼2 𝑒 −𝑗𝜔 ). . . (1 − 𝛼𝑁 𝑒 −𝑗𝜔 )
– where the poles α1, α2, ... , αN are distinct.
– the order of the numerator polynomial of 𝑒 −𝑗𝜔 is less than the order of the denominator
polynomial.
• The transform 𝑋(𝑒 𝑗𝜔 ) can be expanded into partial fractions in the form
𝑘1 𝑘2 𝑘𝑁
𝑋(𝑒 𝑗𝜔 ) = + +. . . +
1 − 𝛼1 𝑒 −𝑗𝜔 1 − 𝛼2 𝑒 −𝑗𝜔 1 − 𝛼𝑁 𝑒 −𝑗𝜔
– the coefficients k1, k2,...,kN can be computed by
𝑘𝑖 = (1 − 𝛼𝑖 𝑒 −𝑗𝜔 ) 𝑋(𝑒 𝑗𝜔 )ቚ 𝑗𝜔 𝑖 = 1, 2, . . . , 𝑁
𝑒 =𝛼𝑖

• 𝑥[𝑛] can be obtained:

𝑥 𝑛 = 𝑘1 ∙ 𝛼1 𝑛 𝑢 𝑛 + 𝑘2 ∙ 𝛼2 𝑛 𝑢 𝑛 + ⋯ + 𝑘𝑁 ∙ 𝛼𝑁 𝑛 𝑢 𝑛
17
Quiz

18
Lecture 17 Filtering
• 1. Continuous-Time (analogue) vs. Discrete-Time (digital)
filtering
– Frequency mapping (analogue frequency vs. digital frequency)
• 2. Ideal and Practical filters
– Four types: lowpass, highpass, bandpass and bandstop
• 3. CT Filters examples
– Simple electrical circuit filters
– Butterworth filters
• 4. DT Filters examples
– Recursive and non-recursive filters
– FIR vs. IIR filters.
19
Analogue and Digital Filters
Note: in discussing discrete-time signals
and systems, Ω is used to denote the
analog frequency of the Continuous-Time
Fourier Transform (CTFT), and 𝜔 is used
to denote the digital frequency of the
Discrete-Time Fourier Transform (DTFT)
for clarity.

1 2𝜋
• Frequency Mapping: 𝐹𝑠 = ; 𝛺𝑠 =
𝑇𝑠 𝑇𝑠
= 2𝜋𝐹𝑠 .
𝛺 𝜔 𝐹 𝜔
= , or 𝜔 = 𝛺𝑇𝑠 ; or = , or 𝜔 = 2𝜋𝐹/𝐹𝑠 = 2𝜋𝐹 ∙ 𝑇𝑠
𝛺𝑠 2𝜋 𝐹𝑠 2𝜋
- 𝑇𝑠 : sampling interval or sampling period (sec, or sec/sample);
- 𝛺: analogue angular frequency (rad/s); 𝛺𝑠 : angular sampling frequency (rad/s);
- 𝐹: analogue linear frequency (Hz); 𝐹𝑠 : linear sampling frequency (Hz);
- 𝜔: angular digital frequency (rad/sample).

20
Quiz

21
Filter Types

CT: focus on 0, ∞

DT: focus on 0, 𝜋

22
Lecture 18 Sampling
• 1. Sampling
– 1.0. What and Why?
– 1.1. Sampling in Time domain (TD)
– 1.2. Sampling in Frequency domain (FD)
– 1.3. Nyquist Theorem

• 2. Reconstruction
– 2.1 Interpolation
– 2.2 Reconstruction theory
– 2.3 Reconstruction in FD – filtering
– 2.4 Reconstruction in TD – interpolation
– 2.5 Realisation
23
 𝑥𝑎 (𝑡) 𝑥𝑠 (𝑡) 𝑥[𝑛]
Sampling in TD and FD Pulse train
𝑥𝑎 (𝑛𝑇)
𝛿𝑇𝑠 (𝑡) Sequence
Time-Domain:
𝑥𝑠 (𝑡) 𝑥𝑎 (𝑡) 𝑥[𝑛]

𝑇𝑠 2𝑇𝑠 3𝑇𝑠 4𝑇𝑠 5𝑇𝑠 6𝑇𝑠 7𝑇𝑠 𝑛

Frequency-Domain:

copy, shift

scaling
𝜔 = 𝛺𝑇𝑠

24
Nyquist-Shannon Theorem and Ideal Reconstruction

• The Nyquist-Shannon theorem!

Ω0 𝑇𝑠 ≤ 𝜋 Ω𝑠 ≥ 2Ω0
where Ω0 is the maximum (highest) frequency in analogue signal.

𝛺 = 𝜔/𝑇𝑠

𝑋𝑎 (𝑗Ω) 𝑥𝑎 (𝑡)
25
Quiz

26
Quiz

27
Quiz

28
Lecture 19 Z-Transform_1
• 1. Definition of z-transform
– Eigenfunctions
– Relationship between DTFT and z-transform
– Visualisation of DTFT and z-transform
– s-plane to z-plane
• 2. Region of Convergence (ROC)
– Definition and graphical depiction
– Zeros and Poles (Zero-pole plot)
– ROC properties
• 3. Commonly use z-transform pairs (Equation list provided)

29
z-Transform vs. DTFT
• Definition equation of DTFT:
∞ ∞

𝐻 𝑒 𝑗𝜔 = ෍ ℎ 𝑛 𝑒 −𝑗𝜔𝑛 = ෍ ℎ[𝑛]𝑧 −𝑛 ቮ = 𝐻(𝑧)ቚ

𝑧=𝑒 𝑗𝜔
𝑛=−∞ 𝑛=−∞
𝑧=𝑒 𝑗𝜔

• Definition equation of z-transform (𝑧 = 𝑟𝑒 𝑗𝜔 ):

∞ ∞

𝐻(𝑧) = ෍ ℎ[𝑛]𝑧 −𝑛 = ෍ ℎ[𝑛]𝑟 −𝑛 𝑒 −𝑗𝜔𝑛 = 𝐷𝑇𝐹𝑇{ℎ[𝑛]𝑟 −𝑛 }

𝑛=−∞ 𝑛=−∞

30
ROC and Properties
• For a given sequence, the set R of values of z for which its z-transform
converges is called the region of convergence (ROC).
– annular region of the z-plane
𝑅− < |𝑧| < 𝑅+ , where 0 ≤ 𝑅− < 𝑅+ ≤ ∞
– The z-transform must always be specified with its ROC !
• DTFT exists ROC include |z|=1 stable
• Right-sided sequence ROC is outside the outermost pole circle;
- causal (right sided and equal to 0 for n < 0): ROC =includes infinity
(𝑋 ∞ < ∞) O(N(z)) ≤ O(D(z));
• Left-sided sequence ROC is inside the innermost pole circle;
- anti-causal (left sided and equal to 0 for n > 0): ROC also includes 0
(𝑋 0 < ∞) O(N(z)) > O(D(z));
• Double-sided sequence ROC will be the intersection of the two ROC
areas, i.e., a ring shape.
31
Lecture 20 Z-Transform_2
• 4. Properties of z-transform
– Linearity, time-shifting, time-reversal, time-scaling, z-domain
scaling, z-domain differentiation, time-difference, time-
accumulation, conjugation, time convolution.
– Comparing with DTFT and Laplace transform
• 5. Inverse z-Transform
– Table Look-up
– Long Division (Power series expansion)
– Partial Fraction Expansion
• ROC determination

32
Z-Transform Properties

Always consider ROC!

33
Inverse z-Transform
• Long-division:

• PFE:
Consider a proper fraction transform 𝑋(𝑧) given with its denominator factored out as
𝑁(𝑧)
𝑋(𝑧) =
(𝑧 − 𝑧1 )(𝑧 − 𝑧2 ). . . (𝑧 − 𝑧𝑁 )
Expanding the transform into partial fractions in the form
𝑘1 𝑧 𝑘2 𝑧 𝑘 𝑧 𝑋(𝑧) 𝑘1 𝑘2 𝑘
𝑋(𝑧) = + +. . . + 𝑁 or equivalently, = + +. . . + 𝑁
𝑧−𝑧1 𝑧−𝑧2 𝑧−𝑧𝑁 𝑧 𝑧−𝑧1 𝑧−𝑧2 𝑧−𝑧𝑁
Consider simple poles, the coefficients 𝑘𝑖 could be obtained from:
𝑋(𝑧)
𝑘𝑖 = (𝑧 − 𝑧𝑖 )
𝑧 𝑧=𝑧𝑖

Each one has its inverse transform like:

−1 −1
𝑘𝑖 𝑧 𝑎𝑛 𝑢[𝑛], 𝑅𝑂𝐶: |𝑧| > |𝑎|
𝑥𝑖 [𝑛] = 𝒵 {𝑋𝑖 (𝑧)} = 𝒵 =ቊ 𝑛
𝑧 − 𝑧𝑖 −𝑎 𝑢[−𝑛 − 1], 𝑅𝑂𝐶: |𝑧| < |𝑎|
34
Quiz

35
Quiz

36
Lecture 21 Z-Transform_3
• 6. Analysis of LTID systems using z-transform
– Impulse response ℎ[𝑛], LCCDE and system transfer function 𝐻(𝑧)
– Zero-pole plot
– Causality and stability determination from ROC
– Geometric Evaluation of DTFT based on zero-pole locations
– System behavior
• 7. Block diagram representation
– Direct form I and II (canonic form)
– Cascade and parallel form
• 8. Unilateral z-transform (optional)

37
LTID System Analysis

𝑥[𝑛] 𝑦[𝑛] 𝑋(𝑧) 𝑌(𝑧)

ℎ[𝑛] 𝐻(𝑧)

• LCCDE (Linear Constant Coefficient Difference Equation):

𝑁 𝑀

෍ 𝑎𝑘 𝑦[𝑛 − 𝑘] = ෍ 𝑏𝑘 𝑥[𝑛 − 𝑘]
𝑘=0 𝑘=0

z-transform
• Transfer function:
𝑌(𝑧) σ𝑀
𝑘=0 𝑏𝑘 𝑧
−𝑘
𝐻 𝑧 = =
𝑋(𝑧) σ𝑁
𝑘=0 𝑎𝑘 𝑧
−𝑘

• Impulse response z-transform

+∞

𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛] = ෍ 𝑥[𝑘]ℎ[𝑛 − 𝑘]

𝑘=−∞

38
Quiz

LCCDE?

39
LTID System Properties
• Causal: ℎ[𝑛] = 0 for 𝑛 < 0
– (a) the ROC is the exterior of a circle outside the outermost pole;
– (b) with 𝐻(𝑧) expressed as a ratio of polynomials in z, the order of
the numerator cannot be greater than the order of the denominator; or
equivalently, ROC includes infinity.

• Stable impulse response being absolutely summable:

– ROC include unit circle |z|=1

• Causal & Stable:

– all of the poles of 𝐻(𝑧) lie inside the unit circle, i.e., they must all
have magnitude smaller than 1.

40
Quiz

Note:
in option c), although infinity is not
explicitly included in ROC, for this
specific X(z), note the order of numerator
and order of denominator are equal
(both = 4), so ROC also includes infinity.
→ causal.

41
Zero-pole plot & Geometric Evaluation of DTFT
• Consider a more general system function in the form:
(𝑧 − 𝑧1 )(𝑧 − 𝑧2 ). . . (𝑧 − 𝑧𝑀 )
𝐻(𝑧) = 𝐾
(𝑧 − 𝑝1 )(𝑧 − 𝑝2 ). . . (𝑧 − 𝑝𝑁 )
– M zeros and N poles;
– The magnitude of the system function is:
𝑧𝑎 − 𝑧1 𝑧𝑎 − 𝑧2 . . . 𝑧𝑎 − 𝑧𝑀
𝐻(𝑧𝑎 ) = 𝐾
𝑧𝑎 − 𝑝1 𝑧𝑎 − 𝑝2 . . . 𝑧𝑎 − 𝑝𝑁
𝐻(𝜔)

𝜔 (rad)
42
Quiz

43
Implementation structures for LTID system
• Take a 3rd-order LTID system as example:
𝑌(𝑧) 𝑏0 + 𝑏1 𝑧 −1 + 𝑏2 𝑧 −2 + 𝑏3 𝑧 −3
𝐻(𝑍) = =
𝑋(𝑧) 1 + 𝑎1 𝑧 −1 + 𝑎2 𝑧 −2 + 𝑎3 𝑧 −3
Direct Form I
Direct Form II
(canonic form)

44
Cascade Form 𝐻(𝑍) = 𝐻1 (𝑧) 𝐻2 (𝑧) . . . 𝐻𝑀 (𝑧)

Parallel Form

෩1 (𝑧) + 𝐻
𝐻(𝑍) = 𝐻 ෩2 (𝑧)+. . . +𝐻
෩𝑀 (𝑧)

• The sub-systems could be any order, but usually are 2nd order.
– Especially for conjugate pole and zero pairs, they are often combined to be 2nd order
sub-systems.

45
Quiz

46
Lecture 22 DFT
• 1. Definition of DFT
– 1.1 DFT definition
– 1.2 DFT: synthesis and analysis equations
– 1.3 Relationships among CTFT, DTFT and DFT
• 2. Computation of DFT
– 2.1 Computing DFT based on twiddle factor
• 3. DFT Properties
– 3.1 Periodicity
– 3.2 Parseval’s theorem
• 4. Circular Convolution
– 4.1 Circular shift
– 4.2 Circular reversal
– 4.3 Circular convolution

47
CTFT - DTFT - DFT
Time 𝒙(𝒕) 𝒙[𝒌] ෥[𝒌]
𝒙
Domain
CTFT DTFT DFT
Frequency
Domain 𝑿(𝒋𝛀) 𝑿(𝒆𝒋𝝎 ) 𝑿[𝒎]

48
Circular Shift and Reversal

𝑔 𝑛 = 𝑥[𝑛 − 𝑚]mod 𝑁

𝑔 𝑛 = 𝑥[−𝑛]mod 𝑁

49
Quiz

50
Circularly Conjugate Symmetry
For any signal 𝑥 𝑛 , it can be decomposed into two components:
𝑥 𝑛 +𝑥 ∗ [−𝑛]mod 𝑁
𝑥𝐸 𝑛 = (conjugate symmetric component)
2
𝑥 𝑛 −𝑥 ∗ [−𝑛]mod 𝑁
𝑥𝑂 𝑛 = (conjugate antisymmetric component)
2
So that
𝑥 𝑛 = 𝑥𝐸 𝑛 +𝑥𝑂 𝑛
In similar manner, its DFT can also be decomposed into
𝑋 𝑘 = 𝑋𝐸 𝑘 +𝑋𝑂 𝑘

51
Circular vs. Linear Convolution
• Circular convolution: 𝑁−1

𝑦 𝑘 = 𝑥1 𝑘 ⊛𝑁 𝑥2 𝑘 = ෍ 𝑥1 𝑛 𝑁 𝑥2 𝑘 − 𝑛 𝑁
𝑛=0
• Expressed in matrix form, take N=4 as an example:

• More convenient method?

52
Quiz

53
 Thank You ! &
GOOD LUCK☺

Please login to LMO to help fill in Module Feedback Questionnaire. Thanks!

https://core.xjtlu.edu.cn/course/view.php?id=1677

54