DSP 101
M4 · L4
Module 4 — The Z-Transform
Inverse Z-Transform
Methods

Three systematic techniques for recovering a time-domain sequence from its Z-domain representation: partial fractions, long division, and the formal contour integral.

1 / 9
DSP 101
M4 · L4
Motivation
From Z-domain
back to Time

Design happens in the Z-domain. Implementation happens in time. The inverse Z-transform bridges the two — recovering h[n] from a pole-zero specification or solving difference equations.

Critical Note
The ROC is mandatory — the same X(z) corresponds to different sequences for different ROCs.
2 / 9
DSP 101
M4 · L4
Method 1
Partial Fraction Expansion

Decompose X(z)/z into first-order terms. Each residue A_k = (z − p_k)X(z)/z at z = p_k. ROC determines causality of each term.

Decomposition
X(z) = \sum_{k=1}^{N} \frac{A_k\,z}{z - p_k}
3 / 9
DSP 101
M4 · L4
ROC Rules
Causality from ROC

The same pole p gives different sequences depending on the region of convergence chosen.

|z|>|p|
Causal: aⁿu[n]
|z|<|p|
Anti-causal: −aⁿu[−n−1]
4 / 9
DSP 101
M4 · L4
Method 2
Long Division &
Power Series

Divide numerator by denominator in ascending powers of z⁻¹. Each quotient coefficient is a direct time-domain sample: the coefficient of z⁻ⁿ is x[n].

Best for
First few samples, hard-to-factor polynomials, and verifying partial-fraction results.
5 / 9
DSP 101
M4 · L4
Method 3
The Contour Integral

Formal definition: x[n] = (1/2πj) ∮ X(z) zn−1 dz. By Cauchy's residue theorem this equals the sum of residues of X(z)zn−1 at poles inside the contour C — identical to partial fractions.

Practical insight
The ROC selects which poles lie inside C, explaining why different ROCs give different sequences.
6 / 9
DSP 101
M4 · L4
Worked Example
Conjugate Poles → Sinusoid

H(z) = z²/(z² − 0.9z + 0.81) has poles at 0.9e±jπ/3. The impulse response is a damped sinusoid: h[n] = (0.9)ncos(nπ/3) u[n].

General pattern
Conjugate poles at radius r, angle θ → rⁿcos(nθ + φ)u[n]. Radius = decay rate; angle = oscillation frequency.
7 / 9
DSP 101
Knowledge Check

Check what stuck

Four questions from this lesson. Answer to see why — the explanation appears whether you were right or wrong. Nothing is scored or saved.

Question 1 of 0
Score 0/0

8 / 9
DSP 101
M4 · L4
Key Takeaways
What You Learned
  • ROC is essential — same X(z) → different x[n] for different ROCs
  • Partial fractions: decompose into first-order terms, read from table
  • Long division: coefficients of power series = time-domain samples
  • Contour integral: formal definition, equals residue sum = partial fractions
  • Conjugate poles at (r, ±θ) → damped sinusoid rⁿcos(nθ + φ)u[n]
  • |z| > |p| → causal; |z| < |p| → anti-causal
Module 4 Complete
You now have the full Z-transform toolkit. Module 5 applies it to digital filter design.
Read full article →
9 / 9
1 / 8 muted