Chapter 1: Basic Signal Operations

𝑪𝑻: 𝒙(𝒕) → 𝒙(𝒕 − 𝒕𝟎 )

𝑫𝑻: 𝒙[𝒏] → 𝒙[𝒏 − 𝒏𝟎 ]
Time Shift
𝒕𝒐 𝒕𝒉𝒆 𝒓𝒊𝒈𝒉𝒕: 𝒕𝟎 > 𝟎
𝒕𝒐 𝒕𝒉𝒆 𝒍𝒆𝒇𝒕 𝒊𝒇 𝒕𝟎 < 𝟎
Time reversal 𝐶𝑇: 𝑥(𝑡) → 𝑥(−𝑡)
𝐷𝑇: 𝑥[𝑛] → 𝑥[−𝑛]
flip over the vertical axis
Time Scaling 𝑥(𝑡) → 𝑦(𝑡) = 𝑥(𝑎 ∙ 𝑡), 𝑎 > 0
𝑎 > 1 ∶ 𝐷𝑒𝑐𝑖𝑚𝑎𝑡𝑒𝑑 (𝑠𝑝𝑒𝑒𝑑 𝑢𝑝)
0 < 𝑎 < 1: 𝐸𝑥𝑝𝑎𝑛𝑑𝑒𝑑 (𝑠𝑙𝑜𝑤𝑒𝑑 𝑑𝑜𝑤𝑛)
Decimation 𝐷𝑇: 𝑦𝐷 [𝑛] = 𝑥[𝑀𝑛], 𝑛𝑜𝑡𝑎𝑡𝑖𝑜𝑛: ↓ 𝑀
Expansion 𝑛
𝑥 [ 𝐿 ] , 𝑛: 𝑚𝑢𝑙𝑡𝑖𝑝𝑙𝑒 𝑜𝑓 𝐿
𝑦𝐸 = { , 𝑛𝑜𝑡𝑎𝑡𝑖𝑜𝑛: ↑ 𝐿
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Linear operation order Shift, Scale, Time Reverse

Periodic Signals
CT DT
Periodic ∃ 𝑟𝑒𝑎𝑙 𝑇 > 0 𝑠. 𝑡. ∃ 𝑖𝑛𝑡𝑒𝑔𝑒𝑟 𝑁 > 0 𝑠. 𝑡.
𝑥(𝑡) = 𝑥(𝑡 + 𝑇), ∀𝑡 ∈ ℝ 𝑥[𝑛] = 𝑥[𝑛 + 𝑁], ∀𝑛 ∈ ℤ
Fundamental Period 𝑇0 = 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑇 > 0 𝑁0 = 𝑠𝑚𝑎𝑙𝑙𝑒𝑠𝑡 𝑁 > 0
Fundamental Radian 2𝜋 2𝜋
𝜔0 = 𝜔0 =
Frequency 𝑇0 𝑁0

Fundamental Frequency 1 1
𝑓0 = 𝑓0 =
𝑇0 𝑁0
 Meaning of 𝝎
Slows down the rate of oscillation, increases
|𝝎| ↓
fundamental period
Speeds up the rate of oscillation, decreases
|𝝎| ↑
fundamental period
𝑥(𝑡) is ct, zero rate of oscillation,
|𝝎| = 𝟎
fundamental period undefined

Even and Odd Signals

CT DT
Even 𝑥(−𝑡) = 𝑥(𝑡) 𝑥[−𝑛] = 𝑥[𝑛]
Odd 𝑥(−𝑡) = −𝑥(𝑡) 𝑥[−𝑛] = −𝑥[𝑛]
𝑥(𝑡) = 𝑥𝑒 (𝑡) + 𝑥𝑜 (𝑡) 𝑥[𝑛] = 𝑥𝑒 [𝑛] + 𝑥𝑜 [𝑛]
𝑥(𝑡) + 𝑥(−𝑡) 𝑥[𝑛] + 𝑥[−𝑛]`
Equations 𝑥𝑒 = 𝑥𝑒 =
2 2
𝑥(𝑡) − 𝑥(−𝑡) 𝑥[𝑛] − 𝑥[−𝑛]`
𝑥𝑜 = 𝑥𝑜 =
2 2

Right-sided Signals Is zero for 𝒕 < 𝑻

Left-sided Signals Is zero for 𝑡 > 𝑇

Unity Impulse and Unity Step Functions

CT DT
Unity Impulse Function 0, 𝑡≠0 0, 𝑛≠0
𝛿(𝑡) = { 𝛿(𝑡) = {
∞, 𝑡=0 1, 𝑛=0
∞
∫ 𝛿(𝑡) 𝑑𝑡 = 1
−∞
Unity Step Function 1, 𝑡≥0 1, 𝑛≥0
𝑢(𝑡) = { 𝑢(𝑡) = {
0, 𝑡<0 0, 𝑛<0
∞ ∞
∫ 𝑥(𝑡) ∙ 𝛿(𝑡 − 𝑡0 ) 𝑑𝑡 = 𝑥(𝑡0 ) ∑ 𝑥[𝑛]𝛿[𝑛 − 𝑛0 ] = 𝑥[𝑛0 ]
−∞ 𝑛=−∞

Properties 𝑡 ∞
𝑢(𝑡) = ∫ 𝛿(𝜏)𝑑𝜏, 𝑢[𝑛] = ∑ 𝛿[𝑛 − 𝑘]
−∞
𝑑𝑖𝑠𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦 𝑎𝑡 𝑡 = 0 𝑘=0
∞ ∞

Sifting property of 𝜹(𝒕) 𝑥(𝑡) = ∫ 𝑥(𝜏)𝛿(𝑡 − 𝜏)𝑑𝜏 𝑥[𝑛] = ∑ 𝑥[𝑘]𝛿[𝑛 − 𝑘]

−∞
𝑘=−∞
Chapter 2: System Properties
Causal (C) Output depends only on the input at 𝒕 < 𝒕𝟎 𝒐𝒓 𝒏 < 𝒏𝟎
Linear Systems (LS) If additive and scalable: 𝑎 ∙ 𝑥1 (𝑡) + 𝑏 ∙ 𝑥2 (𝑡) → 𝑎 ∙ 𝑦1 (𝑡) + 𝑏 ∙ 𝑦2 (𝑡)
Time-Invariance (TI) 𝑥(𝑡) → 𝑦(𝑡) then 𝑥(𝑡 − 𝑡0 ) → 𝑦(𝑡 − 𝑡0 ), ∀ 𝑡0
Memoryless (M) Output at any time 𝑡 depends only on the input at time 𝑡
Invertible (I) One-to-one mapping from any set of distinct inputs to any set of distinct outputs
Stable (S) 𝑥(𝑡) bounded if ∃ 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 𝐶 < ∞ s.t. |𝑥(𝑡) ≤ 𝐶|, ∀ 𝑡

LTI Systems
CT DT
∞ ∞
Convolution/Superposition
𝑦(𝑡) = ∫ 𝑥(𝜏) ∙ ℎ(𝑡 − 𝜏) 𝑑𝜏 𝑦[𝑛] = ∑ 𝑥[𝑘] ∙ ℎ[𝑛 − 𝑘]
Sum
−∞ 𝑘=−∞

Notation ‘*’ 𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛]

∞ ∞

𝑥[𝑛] = ∑ 𝑥[𝑘] ∙ 𝛿[𝑛 − 𝑘] → 𝑦[𝑛] = ∑ 𝑥[𝑘] ∙ ℎ[𝑛 − 𝑘]

𝑘=−∞ 𝑘=−∞

DT Convolution Properties

Commutative 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏] = 𝒉[𝒏] ∗ 𝒙[𝒏]

Associative 𝑦[𝑛] = (𝑥[𝑛] ∗ ℎ1 [𝑛]) ∗ ℎ2 [𝑛] = 𝑥[𝑛] ∗ (ℎ1 [𝑛] ∗ ℎ2 [𝑛])
Distributive 𝑦[𝑛] = 𝑥[𝑛] ∗ (ℎ1 [𝑛] + ℎ2 [𝑛]) = (𝑥[𝑛] ∗ ℎ1 [𝑛]) + (𝑥[𝑛] ∗ ℎ2 [𝑛])
𝑖𝑓 𝑥[𝑛] ∗ ℎ[𝑛] = 𝑦[𝑛],
Delay Accumulation 𝑡ℎ𝑒𝑛 𝑥[𝑛 − 𝑛0 ] ∗ ℎ[𝑛 − 𝑛0 ] ∗ ℎ[𝑛 − 𝑚0 ] = 𝑦[𝑛 − 𝑛0 − 𝑚0 ]
𝑓𝑜𝑟 𝑎𝑛𝑦 𝑖𝑛𝑡𝑒𝑔𝑒𝑟 𝑛0 𝑎𝑛𝑑 𝑚0
Convolving a Signal with 𝑥[𝑛] ∗ 𝛿[𝑛] = 𝑥[𝑛]
a Unit Impulse 𝑥[𝑛] ∗ 𝛿[𝑛 − 𝑛0 ] = 𝑥[𝑛 − 𝑛0 ]
 LTI System Properties in terms of Impulse Response
CT DT
Memoryless ℎ(𝑡) = 𝛼 ∙ 𝛿(𝑡), 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝛼 ∈ ℂ ℎ[𝑛] = 𝛼 ∙ 𝛿[𝑡], 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝛼 ∈ ℂ
Causal ℎ(𝑡) = 0, ∀ 𝑡 < 0 ℎ[𝑛] = 0, ∀ 𝑛 < 0
∞ ∞
Stable
∫ |ℎ(𝑡)|𝑑𝑡 < ∞ ∑ |ℎ[𝑛]| < ∞
−∞ 𝑛=−∞

Invertible (Inverting of
𝑖𝑓𝑓 ∃𝑔(𝑡) 𝑠. 𝑡. 𝑖𝑓𝑓 ∃𝑔[𝑛] 𝑠. 𝑡.
an LTI system is also
ℎ(𝑡) ∗ 𝑔(𝑡) = 𝛿(𝑡) ℎ[𝑛] ∗ 𝑔[𝑛] = 𝛿[𝑛]
called Deconvolving)

Chapter 3: Fourier Series (defined on periodic functions)

CT Signals

𝒌𝒕𝒉 eigenfunction 𝝃𝒌 (𝒕)

𝒌𝒕𝒉 eigenvalue 𝜆𝑘
∞

Signal Representation 𝑥(𝑡) = ∑ 𝑎𝑘 ∙ 𝝃𝒌 (𝒕)

𝑘=−∞
∞ ∞
𝐿𝑇𝐼
Superposition Property of LTI Systems 𝑥(𝑡) = ∑ 𝑎𝑘 ∙ 𝝃𝒌 (𝒕) → 𝑦(𝑡) = ∑ 𝑎𝑘 ∙ 𝜆𝑘 ∙ 𝝃𝒌 (𝒕)
𝑘=−∞ 𝑘=−∞
𝛿(𝑡)
𝑥(𝑡) → 𝑥(𝑡)
Specific LTI Systems
𝛿(𝑡−𝑇)
𝑥(𝑡) → 𝑥(𝑡 − 𝑇)
ℎ(𝑡)
Complex Exponentials 𝑥(𝑡) = 𝑒 𝑠𝑡 → 𝑦(𝑡) = 𝐻(𝑠)𝑒 𝑠𝑡
(Eigenfunction) 𝑠 = 𝛼 + 𝑗𝜔
Transfer Function
∞
(Corresponding eigenvalue) 𝐻(𝑠) = ∫ ℎ(𝑡)𝑒 −𝑠𝑡 𝑑𝑡
−∞
(Independent of time)
𝑠 = 𝑗𝜔, 𝜔 ∈ ℝ
Periodic Complex Exponentials
|𝑒 𝑗𝜔𝑡 | = 1
∞
Frequency Response 𝐻(𝑗𝜔) = ∫ ℎ(𝑡)𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞
2𝜋
Period 𝑇=
𝜔
 Radian Frequency 𝜔 = 2𝜋𝑓
1
Frequency (Hz) 𝑓=
𝑇

DT Signals

𝒉[𝒏]
Complex Exponentials 𝒙[𝒏] = 𝒛𝒏 → 𝒚[𝒏] = 𝑯(𝒛)𝒛𝒏
(Eigenfunction) 𝒛 = 𝒓 ∙ 𝒆𝒋𝝎
Transfer Function ∞

(Eigenvalue) 𝐻(𝑧) = ∑ ℎ[𝑘]𝑧 −𝑘

𝑘=−∞
(Independent of time)
2𝜋 2𝜋
Periodic Complex Exponentials 𝑧 𝑛 = 𝑒 𝑗𝜔𝑛 = 𝑒 𝑗∙ 𝑁 ∙𝑛 , 𝑁 ∈ ℤ, 𝜔 =
𝑁
∞
𝑗𝜔
Frequency Response 𝐻(𝑒 ) = ∑ ℎ[𝑘]𝑒 −𝑗𝜔𝑘
𝑘=−∞

Summary
∞ ∞
𝒔𝒌 𝒕
CT 𝒙(𝒕) = ∑ 𝒂𝒌 𝒆 → 𝒚(𝒕) = ∑ 𝒂𝒌 𝑯(𝒔𝒌 ) ∙ 𝒆𝒔𝒌𝒕
𝒌=−∞ 𝒌=−∞
∞ ∞

DT 𝒙[𝑡] = ∑ 𝒂𝒌 𝒛𝒏𝒌 → 𝒚[𝑡] = ∑ 𝒂𝒌 𝑯(𝒛𝒌 ) ∙ 𝒛𝒏𝒌

𝒌=−∞ 𝒌=−∞

Periodic Signals
𝑒 𝑠𝑡 , 𝑤ℎ𝑒𝑛 𝑠 = 𝜔𝑗: 𝑝𝑢𝑟𝑒 𝑖𝑚𝑎𝑔𝑖𝑛𝑎𝑟𝑦
CT {
𝑠𝑖𝑔𝑛𝑎𝑙𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑜𝑟𝑚: 𝑒 𝑗𝜔𝑡 𝑤𝑖𝑡ℎ |𝑒 𝑗𝜔𝑡 | = 1
𝑧 𝑛 , 𝑤ℎ𝑒𝑛 𝑧 = 𝑒 𝑗𝜔 : 𝑝𝑢𝑟𝑒 𝑝ℎ𝑎𝑠𝑒 (𝑟 = 1)
DT {
𝑠𝑖𝑔𝑛𝑎𝑙𝑠 𝑜𝑓 𝑡ℎ𝑒 𝑓𝑜𝑟𝑚: 𝑒 𝑗𝜔𝑛 𝑤𝑖𝑡ℎ |𝑒 𝑗𝜔𝑛 | = 1
Fourier Series Pair in CT Domain 𝒙(𝒕) ↔ 𝒂𝒌
∞ ∞
𝟐𝝅
Synthesis Equation 𝒙(𝒕) = ∑ 𝒂𝒌 ∙ 𝒆 𝒋𝒌𝝎𝟎 𝒕
= ∑ 𝒂𝒌 ∙ 𝒆𝒋𝒌 𝑻 𝒕

𝒌=−∞ 𝒌=−∞

1 1 𝟐𝝅
Analysis Equation 𝑎𝑘 = ∫ 𝑥(𝑡)𝑒 −𝑗𝑘𝜔0𝑡 𝑑𝑡 = ∫ 𝑥(𝑡)𝑒 −𝑗𝑘 𝑻 𝑡 𝑑𝑡
𝑇 𝑇 𝑇 𝑇
∞ ∞
𝒋𝒌𝝎𝟎 𝒕
𝑥(𝑡) = ∑ 𝒂𝒌 ∙ 𝒆 → 𝒚(𝒕) = ∑ 𝒂𝒌 𝑯(𝒋𝒌𝝎𝟎 ) ∙ 𝒆𝒋𝒌𝝎𝟎 𝒕
𝒌=−∞ 𝒌=−∞
Use of Fourier Series ∞
𝐻(𝒋𝒌𝝎𝟎 ) = ∫ ℎ(𝑡)𝑠 −𝑗𝑘𝝎𝟎 𝑡 𝑑𝑡
−∞

Fourier Series Pair in DT Domain 𝒙[𝒏] ↔ 𝒂𝒌

𝟐𝝅
Synthesis Equation 𝒙[𝒏] = ∑ 𝒂𝒌 ∙ 𝒆𝒋 𝑵 𝒌𝒏
= ∑ 𝒂𝒌 ∙ 𝒆𝒋𝒌𝝎𝟎 𝒏
𝒌=<𝑵> 𝒌=<𝑵>

1 𝟐𝝅 1
𝑎𝑘 = ∑ 𝒙[𝒏] ∙ 𝒆−𝒋 𝑵 𝒌𝒏 = ∑ 𝒙[𝒏] ∙ 𝒆−𝒋𝒌𝝎𝟎 𝒏
𝑁 𝑁
𝒏=<𝑵> 𝒏=<𝑵>
Analysis Equation
: 𝑆𝑢𝑚𝑚𝑖𝑛𝑔 𝑜𝑣𝑒𝑟 𝑎 𝑝𝑒𝑟𝑖𝑜𝑑 𝑁. 𝑆𝑖𝑛𝑐𝑒 𝑡ℎ𝑒 𝑠𝑖𝑔𝑛𝑎𝑙
∑ 𝑖𝑠 𝑟𝑒𝑝𝑒𝑎𝑡𝑒𝑑 𝑒𝑣𝑒𝑟𝑦 𝑁 𝑠𝑎𝑚𝑝𝑙𝑒𝑠, 𝑖𝑡 𝑚𝑎𝑘𝑒𝑠 𝑛𝑜
𝑛=<𝑁> 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑤ℎ𝑖𝑐ℎ 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒 𝑐ℎ𝑜𝑜𝑠𝑒

Spectral Coefficients 𝑎𝑘
 Properties of Fourier Series
CT DT
𝑥1 (𝑡) ↔ 𝑎𝑘 𝑥1 [𝑛] ↔ 𝑎𝑘
Linearity 𝑥2 (𝑡) ↔ 𝑏𝑘 𝑥2 [𝑛] ↔ 𝑏𝑘
𝑐1 𝑥1 (𝑡) + 𝑐2 𝑥2 (𝑡) ↔ 𝑐1 𝑎𝑘 + 𝑐2 𝑏𝑘 𝑐1 𝑥1 [𝑛] + 𝑐2 𝑥2 [𝑛] ↔ 𝑐1 𝑎𝑘 + 𝑐2 𝑏𝑘
𝑥(𝑡) ↔ 𝑎𝑘 𝑥[𝑛] ↔ 𝑎𝑘
2𝜋 2𝜋
Time Shift 𝑥(𝑡 − 𝑡0 ) ↔ 𝑎𝑘 𝑒 −𝑗𝑘𝜔0𝑡0 = 𝑎𝑘 𝑒 −𝑗𝑘 𝑇 𝑡0 𝑥[𝑛 − 𝑛0 ] ↔ 𝑎𝑘 𝑒 −𝑗𝑘𝜔0𝑛0 = 𝑎𝑘 𝑒 −𝑗𝑘 𝑁 𝑛0
2𝜋 2𝜋
𝜔0 = , 𝑇 𝑓𝑢𝑛𝑑𝑎𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑒𝑟𝑖𝑜𝑑 𝜔0 = , 𝑁 𝑓𝑢𝑛𝑑𝑎𝑚𝑒𝑛𝑡𝑎𝑙 𝑝𝑒𝑟𝑖𝑜𝑑
𝑇 𝑁
𝑥(𝑡) ↔ 𝑎𝑘 𝑥[𝑛] ↔ 𝑎𝑘
Time Reversal
𝑥(−𝑡) ↔ 𝑎−𝑘 𝑥[−𝑛] ↔ 𝑎−𝑘
𝑥(𝑡) ↔ 𝑎𝑘 𝑥[𝑛] ↔ 𝑎𝑘
Conjugation ∗ ∗
𝑥 ∗ (𝑡) ↔ 𝑎−𝑘 𝑥 ∗ [𝑛] ↔ 𝑎−𝑘
𝑥(𝑡) ↔ 𝑎𝑘 𝑥[𝑛] ↔ 𝑎𝑘
𝑦(𝑡) ↔ 𝑏𝑘 𝑦[𝑛] ↔ 𝑏𝑘
Multiplication ∞

𝑥(𝑡)𝑦(𝑡) ↔ ∑ 𝑎𝑘 𝑏𝑘−𝑙 𝑥[𝑛]𝑦[𝑛] ↔ ∑ 𝑎𝑘 𝑏𝑘−𝑙

𝑙=−∞ 𝑙=<𝑁>

𝑥(𝑡) ↔ 𝑎𝑘
𝑑 2𝜋
𝑥(𝑡) ↔ 𝑗𝑘𝜔0 𝑎𝑘 = 𝑗𝑘 𝑎
𝑑𝑡 𝑇 𝑘
Differentiation and
𝑥(𝑡) ↔ 𝑎𝑘
Integration
𝑡
1 1
∫ 𝑥(𝜏)𝑑𝜏 ↔ ( ) 𝑎𝑘 = ( )𝑎
𝑗𝑘𝜔0 2𝜋 𝑘
−∞ 𝑗𝑘 𝑇

𝑥(𝑡) ↔ 𝑎𝑘 𝑥[𝑛] ↔ 𝑎𝑘
Parseval Relation 1 ∞ 1
∫ |𝑥(𝑡)|2 𝑑𝑡 = ∑ |𝑎𝑘 |2 ∑ |𝑥[𝑛]|2 = ∑ |𝑎𝑘 |2
𝑇 𝑇 𝑘=−∞
𝑁 𝑛=<𝑁> 𝑘=<𝑁>
Chapter 4: Fourier Transform (applies to both periodic and non-periodic signals)
Non-periodic signal 𝑥(𝑡) can be treated as a periodic signal with 𝑇 → ∞.

1
Sine Formula sin 𝜃 = ∙ (𝑒 𝑗𝜃 − 𝑒 −𝑗𝜃 )
2𝑗
1
Cosine Formula cos 𝜃 = ∙ (𝑒 𝑗𝜃 + 𝑒 −𝑗𝜃 )
2

Euler’s Formula 𝑒 𝑗𝜃 = cos 𝜃 + 𝑗 ∙ sin 𝜃

sin(𝜋𝑥)
Sinc Pulse 𝑠𝑖𝑛𝑐(𝑥) =
𝜋𝑥

Continuous Time Fourier Transform

∞
Synthesis Equation (Inverse Fourier 𝟏
𝒙(𝒕) = ∙ ∫ 𝑋(𝑗𝜔) ∙ 𝑒 𝑗𝜔𝑡 𝒅𝝎
Transform) 𝟐𝝅 −∞
∞
Analysis Equation (Fourier Transform) 𝑋(𝑗𝜔) = ∫ 𝒙(𝒕) ∙ 𝒆−𝒋𝝎𝒕 𝑑𝑡
−∞

Notation 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)

∞ ∞
𝑗𝑘𝜔0 𝑡
Relationship Between FT and FS 𝑥(𝑡) = ∑ 𝑎𝑘 ∙ 𝑒 ↔ 𝑿(𝒋𝝎) = 𝟐𝝅 ∑ 𝑎𝑘 ∙ 𝜹(𝝎 − 𝒌𝝎𝟎 )
𝑘=−∞ 𝑘=−∞
2𝜋
Fundamental Frequency 𝜔0 =
𝑇

Discreate-Time Fourier Transform

∞

DTFT (Analysis Equation) 𝑋(𝑒 𝑗𝜔 ) = ∑ 𝒙[𝒏] ∙ 𝒆−𝒋𝝎𝒏

𝒏=−∞

𝟏
Inverse DTFT (Synthesis Equation) 𝒙[𝒏] = ∫ 𝑿(𝒆𝒋𝝎 ) ∙ 𝒆𝒋𝝎𝒏 𝒅𝝎
𝟐𝝅 𝟐𝝅

Notation 𝒙[𝒏] ↔ 𝑿(𝒆𝒋𝝎 )

 Properties
Name CTFT DTFT
Periodicity 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔+2𝜋) )

If 𝑥1 (𝑡) ↔ 𝑋1 (𝑗𝜔) If 𝑥1 [𝑛] ↔ 𝑋1 (𝑒 𝑗𝜔 )

Linearity 𝑥2 (𝑡) ↔ 𝑋2 (𝑗𝜔) 𝑥2 [𝑛] ↔ 𝑋2 (𝑒 𝑗𝜔 )
𝑎 ∙ 𝑥1 (𝑡) + 𝑏 ∙ 𝑥2 (𝑡) ↔ 𝑎 ∙ 𝑋1 (𝑗𝜔) + 𝑏 ∙ 𝑋2 (𝑗𝜔) 𝑎 ∙ 𝑥1 [𝑛] + 𝑏 ∙ 𝑥2 [𝑛] ↔ 𝑎 ∙ 𝑋1 (𝑒 𝑗𝜔 ) + 𝑏 ∙ 𝑋2 (𝑒 𝑗𝜔 )
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)
If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
−𝑗𝜔𝑡0
Then 𝑦(𝑡) = 𝑥(𝑡 − 𝑡0 ) ↔ 𝑒 ∙ 𝑋(𝑗𝜔)
𝑥[𝑛 − 𝑛0 ] ↔ 𝑒 −𝑗𝜔𝑛0 𝑋(𝑗𝜔)
−𝑗𝜔𝑡0
𝑒 ∙ 𝑋(𝑗𝜔) = |𝑋(𝑗𝜔)| ∙ 𝑒 𝑗∡𝑋(𝑗𝜔) ∙ 𝑒 −𝑗𝜔𝑡0
Time Shift Phase shift:
= |𝑋(𝑗𝜔)| ∙ 𝑒 𝑗(∡𝑋(𝑗𝜔)−𝜔𝑡0 )
If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
A time shift in time-domain contributes to a
𝑒 𝑗𝜔0𝑛 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗(𝜔−𝜔0) )
linear phase shift in frequency domain
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔) If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
Conjugation Then 𝑥 ∗ (𝑡) ↔ 𝑋 ∗ (−𝑗𝜔) Then 𝑥 ∗ [𝑛] ↔ 𝑋 ∗ (𝑒 −𝑗𝜔 )
If 𝑥(𝑡) is real, 𝑥(𝑡) = 𝑥 ∗ (𝑡), 𝑋(𝑗𝜔) = 𝑋 ∗ (−𝑗𝜔)
𝑘 −positive integer
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)
𝑛
Time Scaling / 1 𝜔 𝑥 [ ] , 𝑖𝑓 𝑛 = 𝑟𝑘, 𝑟 𝑖𝑛𝑡𝑒𝑔𝑒𝑟
Then 𝑥(𝑎𝑡) ↔ |𝑎| 𝑋 (𝑗 𝑎 ) 𝑥(𝑘) [𝑛] ={ 𝑘
Time 0, 𝑛 ≠ 𝑟𝑘
Expansion Compressing in time-domain ⇔ Stretching in ∞
𝑗𝜔
frequency domain 𝑋(𝑘) (𝑒 ) = ∑ 𝒙(𝒌) [𝒏] ∙ 𝒆−𝒋𝝎
𝑛=−∞

If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
Time Reversal
𝑥[−𝑛] ↔ 𝑋(𝑒 −𝑗𝜔 )
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)
Differentiation 𝑑𝑥(𝑡) If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
↔ 𝑗𝜔𝑋(𝑗𝜔)
and 𝑑𝑡 𝑑𝑋(𝑒 𝑗𝜔 )
𝑡 𝑛𝑥[𝑛] ↔ 𝑗
Integration 1 𝑑𝜔
∫ 𝑥(𝜏) 𝑑𝜏 ↔ 𝑋(𝑗𝜔) + 𝜋𝑋(0)𝛿(𝜔)
−∞ 𝑗𝜔
If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔) ∞
Parseval’s 𝟏 𝟐
∞
1 ∞ ∑ |𝒙[𝒏]|𝟐 = ∫ |𝑿(𝒆𝒋𝝎 )| 𝒅𝝎
Relation ∫ |𝑥(𝑡)|2 𝑑𝑡 = ∫ |𝑋(𝑗𝜔)|2 𝑑𝜔 𝟐𝝅 𝟐𝝅
𝑛=−∞
−∞ 2𝜋 −∞
𝟏 𝟐
Spectral Energy Density: 𝟐𝝅 |𝑿(𝒆𝒋𝝎 )|
 If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)
ℎ(𝑡) ↔ 𝐻(𝑗𝜔) If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
Convolution 𝑦(𝑡) = 𝑥(𝑡) ∗ ℎ(𝑡) ↔ 𝑌(𝑗𝜔) = 𝑋(𝑗𝜔)𝐻(𝑗𝜔) ℎ[𝑛] ↔ 𝐻(𝑒 𝑗𝜔 )
∞
𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛] ↔ 𝑌(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗𝜔 )𝐻(𝑒 𝑗𝜔 )
𝐻(𝑗𝜔) = ∫ ℎ(𝑡) ∙ 𝑒 −𝑗𝜔𝑡 𝑑𝑡
−∞

If 𝑥[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )
𝑦[𝑛] ↔ 𝑌(𝑒 𝑗𝜔 )
If 𝑥(𝑡) ↔ 𝑋(𝑗𝜔)
𝑧[𝑛] = 𝑥[𝑛]𝑦[𝑛] ↔
𝑦(𝑡) ↔ 𝑌(𝑗𝜔)
Multiplication 1
1 𝑍(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗𝜔 ) ∗ 𝑌(𝑒 𝑗𝜔 )
𝑧(𝑡) = 𝑥(𝑡)𝑦(𝑡) ↔ 𝑍(𝑗𝜔) = 𝑋(𝑗𝜔) ∗ 𝑌(𝑗𝜔) 2𝜋
2𝜋
1
= ∫ 𝑋(𝑒 𝑗𝜃 )𝑌(𝑒 𝑗(𝜔−𝜃) ) 𝑑𝜃
2𝜋 2𝜋

Chapter 5: Sampling
∞

Sampling Function 𝒑(𝒕) = ∑ 𝜹(𝒕 − 𝒏𝑻)

𝒏=−∞
∞

A/D Convertor 𝑥𝑝 (𝑡) = ∑ 𝑥(𝑛𝑇) ∙ 𝛿(𝑡 − 𝑛𝑇)

𝑛=−∞

CT Signal can be viewed as DT if we

𝑥𝑝 (𝑡) ≡ 𝑥(𝑛𝑇)
define:
∞ ∞
Frequency Domain Analysis A/D 𝟐𝝅 𝟐𝝅
𝑝(𝑡) = ∑ 𝜹(𝒕 − 𝒏𝑻) ↔ 𝑷(𝒋𝝎) = ∑ 𝜹(𝝎 − 𝒌 ∙ )
Convertor 𝑻 𝑻
𝒏=−∞ 𝒌=−∞
2𝜋
Fundamental Frequency of 𝒑(𝒕) 𝜔0 =
𝑇
Sampling Frequency 2𝜋
𝜔𝑠 = 𝜔0 = (𝑟𝑎𝑑𝑖𝑎𝑛)
𝑇
1
Fourier Series Coefficients for 𝒑(𝒕) 𝑎𝑘 = , ∀𝑘
𝑇
Convolution 1
𝑥𝑝 (𝑡) = 𝑥(𝑡)𝑝(𝑡) ↔ 𝑋𝑝 (𝑗𝜔) = 𝑋(𝑗𝜔) ∗ 𝑃(𝑗𝜔)
2𝜋
(W-the highest frequency that the signal
2𝜋
x(t) has in the frequency domain) 𝑓𝑜𝑟 2𝑊 =
𝑇
Aliasing The overlap effect at edges of the replicas
 Sampling Theorem 𝑥(𝑡) 𝑖𝑠 𝑏𝑎𝑛𝑑 − 𝑙𝑖𝑚𝑖𝑡𝑒𝑑 𝑖𝑓
(W-the bandwidth of the signal) 𝑥(𝑡) ↔ 𝑋(𝑗𝜔) = 0, ∀|𝜔| > 𝑊
𝑥(𝑡) 𝑖𝑠 𝑏𝑎𝑛𝑑 − 𝑙𝑖𝑚𝑖𝑡𝑒𝑑 𝑤𝑖𝑡ℎ 𝑋(𝑗𝜔) = 0, ∀|𝜔| > 𝑊,
𝑡ℎ𝑒𝑛 𝑡ℎ𝑒 𝑚𝑖𝑛𝑖𝑚𝑢𝑚 𝑠𝑎𝑚𝑝𝑙𝑖𝑛𝑔 𝑟𝑎𝑡𝑒 𝜔𝑠 𝑠𝑢𝑐ℎ 𝑡ℎ𝑎𝑡 𝑛𝑜 𝑎𝑙𝑖𝑎𝑠𝑖𝑛𝑔
Sampling Theorem
𝑜𝑐𝑐𝑢𝑟𝑠 𝑖𝑛 𝑋(𝑗𝜔)𝑖𝑠 𝑡𝑤𝑖𝑐𝑒 𝑡ℎ𝑒 ℎ𝑖𝑔ℎ𝑒𝑠𝑡 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑜𝑓 𝑥(𝑡) 𝑖𝑛
𝑡ℎ𝑒 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦 𝑑𝑜𝑚𝑎𝑖𝑛: 𝜔𝑠 ≥ 2𝑊
The frequency that must be exceeded by the sampling frequency
Nyquist Rate (2W) 𝜔𝑠
𝜔𝑠 = 2𝑊
Nyquist Frequency (W) Half of the Nyquist rate
𝑾 = 𝟐𝝅𝒇 𝜔𝑠 = 𝑊
∞

𝑥𝑟 (𝑡) = ∑ 𝑥(𝑛𝑇) ∙ ℎ(𝑡 − 𝑛𝑇)

𝑛=−∞

For an ideal LPF 𝐻(𝑗𝜔):

Reconstruction of a Signal from Its
𝑇 ∙ sin(𝜔𝑐 𝑡) 𝜔𝑐 𝑇 𝜔𝑐 𝑡
Samples (Interpolation) ℎ(𝑡) = = 𝑠𝑖𝑛𝑐 ( ) ↔ 𝐻(𝑗𝜔)
𝜋𝑡 𝜋 𝜋
∞
𝜔𝑐 𝑇 𝜔𝑐 (𝑡 − 𝑛𝑇)
𝑥𝑟 (𝑡) = ∑ 𝑥(𝑛𝑇) ∙ 𝑠𝑖𝑛𝑐 ( )
𝜋 𝜋
𝑛=−∞

Chapter 6: Communication Systems

Embedding the information-bearing signal (modulating signal 𝒙(𝒕)) into
Modulation
second signal (carrier signal 𝒄(𝒕))
Modulated Signal: 𝑦(𝑡) = 𝑥(𝑡) ∙ 𝑐(𝑡)
Amplitude Modulation (AM):
Complex Exponential Signal 𝑐(𝑡) = 𝑒 𝑗(𝜔𝑐∙𝑡+𝜃𝑐)
Carrier Frequency 𝜔𝑐 = 2𝜋𝑓𝑐
Carrier Phase 𝜃𝑐
1
𝑦(𝑡) = 𝑥(𝑡) ∙ 𝑐(𝑡) ↔ 𝑌(𝑗𝜔) = 𝑋(𝑗𝜔) ∗ 𝐶(𝑗𝜔)
2𝜋
Multiplication Property of
If 𝜃𝑐 = 0 ⇒ 𝑐(𝑡) = 𝑒 𝑗𝜔𝑐𝑡 ↔ 𝐶(𝑗𝜔) = 2𝜋𝛿(𝜔 − 𝜔𝑐 )
The Fourier Transform
1
𝑌(𝑗𝜔) = 𝑋(𝑗𝜔) ∗ [2𝜋 ∙ 𝛿(𝜔 − 𝜔𝑐 )] = 𝑋(𝑗(𝜔 − 𝜔𝑐 ))
2𝜋
 Demodulation Recovering the modulating signal 𝑥(𝑡) from the modulated signal 𝑦(𝑡)
Modulated Signal 𝑦(𝑡) = 𝑥(𝑡)𝑐(𝑡) = 𝑥(𝑡) ∙ 𝑒 𝑗𝜔𝑐𝑡
Demodulated Signal 𝑥𝑟 (𝑡) = 𝑥(𝑡) = 𝑦(𝑡) ∙ 𝑒 −𝑗𝜔𝑐𝑡
1
𝑋𝑟 (𝑗𝜔) = 𝑌(𝑗𝜔) ∗ [2𝜋 ∙ 𝛿(𝜔 + 𝜔𝑐 )]
Frequency Domain Analysis 2𝜋
= 𝑌(𝑗𝜔) ∗ 𝛿(𝜔 + 𝜔𝑐 )

Chapter 7: Z-Transform
∞

𝑋(𝑧) = ∑ 𝑥[𝑛] ∙ 𝑧 −𝑛
Z-Transform 𝑛=−∞

where 𝑧 = 𝑟𝑒 𝑗𝜔 , 𝑟 ∈ ℝ, 𝜔 ∈ ℝ
Notation 𝑥[𝑛] ↔ 𝑋(𝑧)
∞

DTFT of 𝒙[𝒏] can be described as a 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑧)|𝑧=𝑒 𝑗𝜔 = ∑ 𝑥[𝑛] ∙ 𝑒 −𝑗𝜔𝑛

𝑛=−∞
special case of the Z-Transform, when
DTFT of 𝑥[𝑛] can be viewed as the Z-Transform of 𝑥[𝑛]
𝒓=𝟏
evaluated at the points on the unit circle on the complex Z-plane
∞
−𝑛
𝑋(𝑟𝑒 𝑗𝜔 ) = ∑ 𝑥[𝑛] ∙ (𝑟𝑒 𝑗𝜔 )
𝑛=−∞
Relationship Between DTFT and Z- ∞
Transform = ∑ (𝑟 −𝑛 ∙ 𝑥[𝑛]) ∙ 𝑒 −𝑗𝜔𝑛
𝑛=−∞

∴ 𝑋(𝑟𝑒 𝑗𝜔 ) = 𝐷𝑇𝐹𝑇{𝑟 −𝑛 𝑥[𝑛]}

∞

𝑋(𝑧) < ∞ ⇒ ∑ |𝒙[𝒏]| ∙ 𝒓−𝒏 < ∞

Region of Convergence (RoC) 𝒏=−∞

Convergence may happen for some values 𝒓 = |𝒛|

Properties of Z-Transform
𝑥1 [𝑛] ↔ 𝑋1 (𝑧), 𝑅𝑜𝐶 = 𝑅1
𝑥2 [𝑛] ↔ 𝑋2 (𝑧), 𝑅𝑜𝐶 = 𝑅2
Linearity
Then 𝑎 ∙ 𝑥1 [𝑛] + 𝑏 ∙ 𝑥2 [𝑛] ↔ 𝑎 ∙ 𝑋1 (𝑧) + 𝑏 ∙ 𝑋2 (𝑧),
𝑅𝑜𝐶 = 𝑅1 ∩ 𝑅2
If 𝑥[𝑛] ↔ 𝑋(𝑧), 𝑅𝑜𝐶 = 𝑅
Time Shift
Then 𝑥[𝑛 − 𝑛0 ] ↔ 𝑋(𝑧)𝑧 −𝑛0 , 𝑅𝑜𝐶 = 𝑅
If 𝑥[𝑛] ↔ 𝑋(𝑧), 𝑅𝑜𝐶 = 𝑅
𝑧
Then 𝑧0𝑛 ∙ 𝑥[𝑛] ↔ 𝑊(𝑧) = 𝑋 (𝑧 ) , 𝑅𝑜𝐶 = |𝑧0 | ∙ 𝑅
0

𝑧
Scaling in the Z-Domain if 𝑧 ∈ 𝑅𝑜𝐶 {𝑋 (𝑧)} ⇒ |𝑧0 |𝑧 ∈ 𝑅𝑜𝐶 {𝑋 ( )}
𝑧0

if 𝑧 = 𝑎 is a pole (or zero) of 𝑋(𝑧) ⇒ 𝑧 = 𝑧0 is also a pole (or zero) of

𝑧
𝑋 (𝑧 )
0

Special Case
𝑒 𝑗𝜔0𝑛 ∙ 𝑥[𝑛] ↔ 𝑋(𝑒 −𝑗𝜔0 ∙ 𝑧)
(𝒛𝟎 = 𝒆𝒋𝝎𝟎 )
If 𝑥[𝑛] ↔ 𝑋(𝑧), 𝑅𝑜𝐶 = 𝑅
1 1
Time Reversal Then 𝑥[−𝑛] ↔ 𝑊(𝑧) = 𝑋 (𝑧) , 𝑅𝑜𝐶 = 𝑅
1 1
meaning: if 𝑧0 ∈ 𝑅𝑜𝐶{𝑋(𝑧)} ⇒ 𝑧 ∈ 𝑅𝑜𝐶 {𝑋 (𝑧)}
0

If 𝑥[𝑛] ↔ 𝑋(𝑧), 𝑅𝑜𝐶 = 𝑅

Then: 𝑥 ∗ [𝑛] ↔ 𝑋 ∗ (𝑧 ∗ ), 𝑅𝑜𝐶 = 𝑅
Conjugation For real 𝑥[𝑛] = 𝑥 ∗ [𝑛] ⇒ 𝑋(𝑧) = 𝑋 ∗ (𝑧 ∗ )
meaning if 𝑧 = 𝑧0 is a pole (or zero) of 𝑋(𝑧) ⇒ 𝑧 = 𝑧0∗ is also a pole (or
zero) of 𝑋(𝑧)
If 𝑥[𝑛] ↔ 𝑋(𝑧), 𝑅𝑜𝐶 = 𝑅
𝑛
𝑥 [𝑘 ] , 𝑖𝑓 𝑛 = 𝑟𝑘, 𝑟 𝑖𝑛𝑡𝑒𝑔𝑒𝑟
and 𝑥(𝑘) [𝑛] = {
0, 𝑛 ≠ 𝑟𝑘
Time Expansion
Then 𝑥(𝑘) [𝑛] ↔ 𝑋(𝑘) (𝑧) = 𝑋(𝑧 𝑘 ), 𝑅𝑜𝐶 = 𝑅1/𝑘
1
𝐼𝑓 𝑧 ∈ 𝑅𝑜𝐶{𝑋(𝑧)} ⇒ 𝑧 𝑘 ∈ 𝑅𝑜𝐶{𝑋(𝑧 𝑘 )}
If 𝑧 = 𝑎 is a pole (or zero) of 𝑋(𝑧) ⇒ 𝑧 = 𝑎𝑎/𝑘 is a pole (or zero) of 𝑋(𝑧 𝑘 )
If 𝑥1 [𝑛] ↔ 𝑋1 (𝑧), 𝑅𝑜𝐶 = 𝑅1
Convolution
𝑥2 [𝑛] ↔ 𝑋2 (𝑧), 𝑅𝑜𝐶 = 𝑅2
 Then 𝑥1 [𝑛] ∗ 𝑥2 [𝑛] ↔ 𝑋1 (𝑧)𝑋2 (𝑧), 𝑅𝑜𝐶 = 𝑹𝟏 ∩ 𝑹𝟐

System Properties Using Z-Transform

A discreate-time LTI system is causal if and only if the RoC of its
Causality
system function 𝐻(𝑧) is the exterior of a circle, including infinity
A discreate-time LTI system is stable iff the RoC of its system function
Stability
𝑯(𝒛) includes the unit circle, i.e the values of 𝑧 with 𝑧 = 1
A causal discreate-time LTI system with rational system function is
Causal→Stable
stable iff all the poles of 𝑯(𝒛) are inside the unit circle
Useful Formulas:
𝜋
sin(𝜃) = cos ( − 𝜃)
2
𝜋
cos(𝜃) = sin ( − 𝜃)
2
cos(𝜃 ± 𝜋) = −cos (𝜃)
sin(2𝜃) = 2 ∙ 𝐬𝐢𝐧 𝜃 𝐜𝐨𝐬 𝜃
cos(2𝜃) = cos 2 𝜃 − 𝑠𝑖𝑛2 𝜃
= 2 ∙ cos2 𝜃 − 1
= 1 − 2 ∙ 𝑠𝑖𝑛2 𝜃
2 ∙ tan 𝜃
tan(2𝜃) =
1 − 𝑡𝑎𝑛2 𝜃
cos(3𝜃) = 4 ∙ cos3 𝜃 − 3 ∙ cos 𝜃
sin(3𝜃) = 3 ∙ 𝐬𝐢𝐧 𝜃 − 4 ∙ sin3 𝜃
3 ∙ tan 𝜃 − 𝑡𝑎𝑛3 𝜃
tan(3𝜃) =
1 − 3 ∙ 𝑡𝑎𝑛2 𝜃
1
cos(𝑎) cos(𝑏) = ∙ (𝑐𝑜𝑠(𝑎 + 𝑏) + 𝑐𝑜𝑠(𝑎 − 𝑏))
2
1
sin(𝑎) sin(𝑏) = ∙ (𝑐𝑜𝑠(𝑎 − 𝑏) − 𝑐𝑜𝑠(𝑎 + 𝑏))
2
1
sin(𝑎) cos(𝑏) = ∙ (𝑠𝑖𝑛(𝑎 + 𝑏) + 𝑠𝑖𝑛(𝑎 − 𝑏))
2
1
cos(𝑎) sin(𝑏) = ∙ (𝑠𝑖𝑛(𝑎 + 𝑏) − 𝑠𝑖𝑛(𝑎 − 𝑏))
2
sin(𝑎 + 𝑏) = 𝐬𝐢𝐧 𝑎 ∙ 𝐜𝐨𝐬 𝑏 + 𝐜𝐨𝐬 𝑎 ∙ 𝐬𝐢𝐧 𝑏
cos(𝑎 + 𝑏) = 𝐜𝐨𝐬 𝑎 ∙ 𝐜𝐨𝐬 𝑏 − 𝐬𝐢𝐧 𝑎 ∙ 𝐬𝐢𝐧 𝑏
sin(𝑎 − 𝑏) = sin 𝑎 ∙ cos 𝑏 − cos 𝑎 ∙ sin 𝑏
𝑐𝑜𝑠(𝑎 − 𝑏) = cos 𝑎 ∙ cos 𝑏 + sin 𝑎 ∙ sin 𝑏