Professional Documents
Culture Documents
Fadi Karameh
June 2020
EECE340 © Fadi Karameh
• We have so far assumed that the DT system is
causal, that is y[n] cannot depend on
x[n+1],x[n+2] etc..
𝑦 𝑛 = 0.5 𝑥 𝑛 + 2𝑥 𝑛 − 1 + 𝑥[𝑛 + 1]
0.5
1
-1 0 1 n
3D plot of blurring
𝑥[𝑛] 𝑥(𝑡)
𝑎2 𝑎 𝑎 𝑎 𝑎2
A blurring effect of the lens could be
described as 𝑎2 𝑎 1 𝑎 𝑎2
𝑎2 𝑎 𝑎 𝑎 𝑎2
(a<1) 𝑎2 𝑎2 𝑎2 𝑎2 𝑎2
-2 -1 0 1 2 n
𝑦 𝑛 = 𝑥 𝑛 + 𝑎𝑥 𝑛 − 1 + 𝑎𝑥 𝑛 + 1 + 𝑎2 𝑥 𝑛 − 2 + 𝑎2 𝑥[𝑛 + 2]
𝑦𝑛 = 𝑘 𝑥[𝑛 − 𝑘]
𝑘=−2
EECE340 © Fadi Karameh
General DT LTI system representation
∞
𝑦𝑛 = 𝑘 𝑥[𝑛 − 𝑘] ≜ 𝑛 ∗ 𝑥[𝑛]
𝑘=−∞
𝑦𝑛 = 𝑘 𝑥[𝑛 − 𝑘] ≜ 𝑛 ∗ 𝑥[𝑛]
𝑘=−∞
• DT convolution properties:
1. Commutative:
𝑦 𝑛 = 𝑛 ∗ 𝑥 𝑛 = 𝑥 𝑛 ∗ [𝑛]
𝑥[𝑛] y[𝑛]
𝑛 h[𝑛] y[𝑛]
𝑥𝑛
𝑥[𝑛] y[𝑛]
g𝑛 h𝑛
𝑥[𝑛] y[𝑛]
𝑒𝑞 𝑛
EECE340 © Fadi Karameh
• DT convolution properties:
3. Distributive:
𝑦 𝑛 = 𝑛 ∗ 𝑚 𝑛 +𝑥 𝑛 = 𝑛 ∗𝑚+ 𝑛 ∗𝑥 𝑛
1 𝑛
𝑥𝑛 = 𝑢[𝑛 +2]
2
-2
• Left-sided sequence:
𝑥 𝑛 = 2 𝑛 𝑢[−𝑛 − 1]
-1
𝑥 𝑛 = 0, 𝑛 >𝑀
-M M
𝑎𝑚 𝑦 𝑛 + 𝑚 = 𝑏𝑙 𝑥 𝑛 + 𝑙
𝑚=−𝑀 𝑙=−𝑁
𝑋 𝑧 = 𝑥[𝑛]𝑧 −𝑛
𝑛=−∞
∞ 𝑛
1 ∞
1 −1
𝑛
= 𝑧 −𝑛 = 𝑧
2 2
𝑛=0 𝑛=0
1
= 1 −1 1
1 f𝑜𝑟 𝑧 <1 𝑜𝑟 𝑧 >
1 − 𝑧 −1 2 2
2
-4
-2
…
-2 -1
. . .
0 1 2 n
𝑋 𝑧 = 𝑥[𝑛]𝑧 −𝑛
𝑛=−∞
∞ 𝑛
1
𝑋 𝑧 =− 𝑢[−𝑛 − 1]𝑧 −𝑛
2
𝑛=−∞
−1 𝑛 −1 𝑛
1 1 −1
=− 𝑧 −𝑛 =− 𝑧
2 2
𝑛=−∞ 𝑛=−∞
∞ ∞
𝑛 𝑛
=− 2𝑧 =1− 2𝑧
𝑛=1 𝑛=0
EECE340 © Fadi Karameh
1 𝑛
• Find the Z transform of 𝑥 𝑛 = − 2
𝑢[−𝑛 − 1]
∞
𝑛
𝑋(𝑧) = 1 − 2𝑧
𝑛=0
1
1 𝑓𝑜𝑟 2𝑧 < 1 or 𝑧 <
=1− 2
1 − 2𝑧
−2𝑧
=
1 − 2𝑧
1
= 1
1 𝑅𝑂𝐶: 𝑧 <
1 − 𝑧 −1 2
2
EECE340 © Fadi Karameh
1 𝑛 𝑅𝑂𝐶𝑥
𝑥 𝑛 =− 𝑢[−𝑛 − 1]
2
1
𝑋(𝑧) = 1 1
1 −1 𝑅𝑂𝐶: 𝑧 <
1− 𝑧 2 2
2
𝑅𝑂𝐶𝑦
1 𝑛
𝑦𝑛 = 𝑢[𝑛]
2
1 1
𝑌(𝑧) = 1
1 −1 ROC: 𝑧 > 2
1− 𝑧 2
2
(left-sided sequence)
𝑎
1
𝑋(𝑧) =
1 − 𝑎𝑧 −1
𝑥 𝑛 = 𝑎 𝑛 𝑢[𝑛]
(right-sided sequence)
ROC: 𝑧 > 𝑎 𝑎
EECE340 © Fadi Karameh
• Example: Double sided sequences
𝑥 𝑛 = 3 𝑛 𝑢[−𝑛 − 1] + 2 𝑛 𝑢[𝑛]
𝑥 𝑛 = 𝑥1 𝑛 + 𝑥2 [𝑛] 𝑅𝑂𝐶𝑥
3
−1 2 3
𝑋1 (𝑧) = 1
1 − 3𝑧 −1 𝑋2 (𝑧) =
1 − 2𝑧 −1
ROC: 𝑧 < 3 ROC: 𝑧 > 2
Doughnut
𝑋 𝑧 = 𝑋1 𝑧 + 𝑋2 (𝑧)
𝑅𝑂𝐶𝑥 2 < 𝑧 < 3 EECE340 © Fadi Karameh
• Example: Double sided sequences
𝑥 𝑛 = 2 𝑛 𝑢[−𝑛 − 1] + 3 𝑛 𝑢[𝑛]
𝑥 𝑛 = 𝑥1 𝑛 + 𝑥2 [𝑛]
−1 3
𝑋1 (𝑧) = 1 2 3
1 − 2𝑧 −1 𝑋2 (𝑧) =
1 − 3𝑧 −1
ROC: 𝑧 < 2 ROC: 𝑧 > 3
𝑧 −1
𝐻 𝑧 =1+
1 + 2𝑧 −1
𝛿𝑛 ??
1
𝐻1 𝑧 =
1 + 2𝑧 −1 noncausal
𝑛 = 𝛿 𝑛 + 1 𝑛 − 1 = 𝛿 𝑛 + −2 𝑛−1 𝑢[𝑛 − 1]
0
EECE340 © Fadi Karameh
• If we know that the system is BIBO stable
∞
[𝑘] < ∞
𝑘=−∞
𝑧 1 1
𝐻 𝑧 = 𝑅𝑂𝐶 ∶ < 𝑧 <
1 1 6 3
𝑧+ 𝑧−
3 6
1. Find the corresponding LDE and unit sample
response.
2. Is the system causal?
3. Is the system stable?
1 1
(𝑧 2 + 𝑧 − ) 𝑌(𝑧) = 𝑧𝑋(𝑧)
6 18
Corresponding LDE is :
1 1
𝑦 𝑛 + 2 + 𝑦 𝑛 + 1 − 𝑦 𝑛 = 𝑥[𝑛 + 1]
6 18
To find h[n]:
𝑧 −1 −2 2
𝐻 𝑧 = = +
1 1 1 −1 1
1 + 𝑧 −1 1 − 𝑧 −1 1+ 𝑧 1 − 𝑧 −1
3 6 3 6
−2 1
𝐻1 𝑧 = 𝑅𝑂𝐶1 ∶ 𝑧 <
1 3
1 + 𝑧 −1
3
2 1
𝐻2 𝑧 = 𝑅𝑂𝐶2 : 𝑧 >
1 −1 6
1− 𝑧
6
2 1
𝐻2 𝑧 = 𝑅𝑂𝐶2 : 𝑧 > right- sided sequence
1 6
1 − 𝑧 −1
6
𝑛
1 ( 𝑛 = 0 ∀ 𝑛 < 0)
2 𝑛 = 2 𝑢𝑛 causal
6
𝑛 𝑛
1 1
𝑛 = 1 𝑛 + 2 𝑛 = 2 − 𝑢 −𝑛 − 1 + 2 𝑢𝑛
3 6
h[n] is non causal EECE340 © Fadi Karameh
2 1 𝑅𝑂𝐶2
𝐻2 𝑧 = 𝑅𝑂𝐶2 : 𝑧 >
1 −1 6
1− 𝑧
6
𝟏
𝑛 𝟔
1
2 𝑛 = 2 𝑢𝑛
6
decaying
2 [𝑛] 2
2 1
𝑛
2 2 → 0 𝑎𝑠 𝑛 → ∞
6 6
36
BIBO stable
𝑛
1 2
3 2 [𝑛] < ∞
-1 0
𝑛
1 𝟏
1 𝑛 = 2 − 𝑢 −𝑛 − 1 𝟑
3
1 [𝑛] 2*72
2*9 growing
𝑛
1
2 − → ∞ 𝑎𝑠 𝑛 → −∞
3
-3 -2 -1 0
𝑛 NOT BIBO stable
-2*3 1 [𝑛] → ∞
-2*27
EECE340 © Fadi Karameh
• Therefore
𝑧 −1 1 1
𝐻 𝑧 = 𝑅𝑂𝐶 ∶ < 𝑧 <
1 −1 1 −1 6 3
1+ 𝑧 𝑧− 𝑧
3 6
𝑧 −1
𝐻 𝑧 = 1 1
1 1
1 + 𝑧 −1 𝑧 − 𝑧 −1 𝑅𝑂𝐶 ∶ < 𝑧 <
3 6 6 3
3
+3
3 𝑛 = 2 −3 𝑛 𝑢 −𝑛 − 1 3
3 [𝑛]
2*(1/9) decaying
2 −3 𝑛 → 0 𝑎𝑠 𝑛 → −∞
-3 -2 -1 0
𝑛 BIBO stable
-2*1/3 3 [𝑛] < ∞
-2*(1/27)
EECE340 © Fadi Karameh
−2 3 𝑛 = 2 −3 𝑛 𝑢 −𝑛 − 1
𝐻3 𝑧 =
1 + 3𝑧 −1
𝑅𝑂𝐶3
×
-3 STABLE
1
−2
𝐻1 𝑧 = 1
𝑛
1 1 𝑛 = 2 − 𝑢 −𝑛 − 1
1 + 𝑧 −1 3
3
𝑅𝑂𝐶1
𝟏×
− 1 UNSTABLE
𝟑
EECE340 © Fadi Karameh
• Therefore the new system G
𝑧 −1 1
G 𝑧 = 1 𝑅𝑂𝐶 ∶ < 𝑧 <3
1+3𝑧 −1 1− 𝑧 −1 6
6
𝑅𝑂𝐶2
×𝟏 STABLE
𝟔
1
2 4 𝑛 = 2 6 𝑛 𝑢 𝑛
𝐻4 𝑧 = 𝑅𝑂𝐶4 : 𝑧 >6
1 − 6𝑧 −1
𝑅𝑂𝐶4
UNSTABLE
×
1 6
EECE340 © Fadi Karameh
Right-sided unit sample response
1
H 𝑧 = 𝑧 > 𝑝
1−𝑝𝑧 −1
𝑅𝑂𝐶
𝑅𝑂𝐶
p 1
1 p
1
EECE340 © Fadi Karameh
Left-sided unit sample response
1
H 𝑧 = 𝑧 < 𝑝
1−𝑝𝑧 −1
𝑅𝑂𝐶
𝑅𝑂𝐶
p
1 p
1
EECE340 © Fadi Karameh
Double sided unit sample response
1
H 𝑧 = 𝑝1 < 𝑧 < 𝑝2
1−𝑝1 𝑧 −1 1−𝑝2 𝑧 −1
𝑛 𝑛
𝑛 ∝ 𝑝1 𝑢 𝑛 + 𝑝2 𝑢[−𝑛 − 1]
𝒑𝟐 > 𝟏
Stable
𝒑𝟏X X 𝒑𝟏 < 𝟏
1 𝒑𝟐
→ ∞
𝑛 ∝ 𝑝1 𝑛 𝑢 𝑛 + 𝑝2 𝑛 𝑢[−𝑛 − 1]
𝒑𝟏 < 𝟏
Unstable
𝒑𝟐 < 𝟏
1 𝒑𝟏 X X
𝒑𝟐
→ ∞
𝑛 ∝ 𝑝1 𝑛 𝑢 𝑛 + 𝑝2 𝑛 𝑢[−𝑛 − 1]
𝒑𝟏 > 𝟏
Unstable
1 X X 𝒑𝟐 > 𝟏
𝒑𝟏 𝒑𝟐
5 −1 (Casual unstable)
1 + 𝑧 + 𝑧 −2 𝑌 𝑧 = 𝑧 −1 𝑋 𝑧 Or
2
(Stable, noncausal )
𝑧 −1 𝑧 −1
H z = 5 = 1 Stable if left sided
1+ 𝑧 −1 +𝑧 −2 1+ 𝑧 −1 1+2𝑧 −1
2 2
Stable if right sided EECE340 © Fadi Karameh
Application: finding inverse systems
m[𝑛]
y[𝑛]
h[𝑛] h 𝑛 : filtering effect
( DT LTI operation)
y[𝑛] x 𝑛 = 𝑚[𝑛]
g[𝑛] Design g 𝑛 such that 𝑚 𝑛
? is retrieved or recovered at 𝑥 𝑛
M(𝑧) X(𝑧)
H(𝑧) 𝑌(𝑧) G(z)
X 𝑧 = 𝐺 𝑧 𝑌 𝑧 = 𝐺(𝑧)𝐻 𝑧 𝑀(𝑧)
1
𝐺 𝑧 𝐻 𝑧 =1 𝐺 𝑧 = = 𝐻𝑖𝑛𝑣 (𝑧)
𝐻(𝑧)
3D plot of blurring
𝑥[𝑛] 𝑥(𝑡)
𝑎2 𝑎 𝑎 𝑎 𝑎2
A blurring effect of the lens could be
described as 𝑎2 𝑎 1 𝑎 𝑎2
𝑎2 𝑎 𝑎 𝑎 𝑎2
(a<1) 𝑎2 𝑎2 𝑎2 𝑎2 𝑎2
𝐻𝑖𝑛𝑣 𝑧 =
2 5 2
𝐻𝑖𝑛𝑣 𝑧 = − 𝛿 𝑛 + 1 + 𝛿 𝑛 − 𝛿[𝑛 − 1]
3 6 3
m[𝑛] y[𝑛]
h[𝑛] y[𝑛] x[𝑛]
g[𝑛]