Introduction to Z-Transform

Introduction

The Z-Transform occupies a central position in discrete-time signal processing, serving the same foundational role that the Laplace Transform plays in continuous-time systems. If you have ever worked with digital filters, control algorithms running on microprocessors, or software-based audio processing, you have encountered systems best analyzed through the Z-Transform framework. It provides the mathematical machinery to convert difference equations into algebraic expressions, analyze stability through geometric relationships in the complex plane, and design digital systems with specified frequency characteristics.

The beauty of the Z-Transform lies in its ability to turn the recursive, sequential nature of discrete-time computation into static algebraic relationships. A difference equation that requires tracking previous outputs and inputs becomes a simple rational function in the complex variable $z$, which can be factored, decomposed, and manipulated using familiar algebraic techniques.

Formal Definition

The Z-Transform of a discrete-time sequence $x[n]$ is defined as:

$$X(z) = \sum_{n=-\infty}^{\infty} x[n] z^{-n}$$

where $z$ is a complex variable. Mathematically, this is a Laurent series — a power series that includes both positive and negative powers of $z^{-1}$. The transform maps a sequence of numbers (the time-domain signal) into a function of a complex variable (the z-domain representation).

For causal sequences where $x[n] = 0$ for $n < 0$, the summation starts at $n = 0$:

$$X(z) = \sum_{n=0}^{\infty} x[n] z^{-n} = x[0] + x[1]z^{-1} + x[2]z^{-2} + \cdots$$

This reveals that $z^{-1}$ represents a unit delay: multiplying by $z^{-1}$ shifts the sequence forward by one position in the power series.

The Complex z-Plane

The complex variable $z$ can be expressed in polar form as:

$$z = re^{j\Omega}$$

where $r = |z|$ is the radial distance from the origin and $\Omega$ is the angle. This polar representation creates a natural geometric framework:

• The unit circle ($|z| = 1$, or $r = 1$): This is the most important contour in the z-plane. Evaluating $X(z)$ on the unit circle gives the Discrete-Time Fourier Transform: $X(e^{j\Omega}) = X(z)\big|_{z=e^{j\Omega}}$
• Inside the unit circle ($|z| < 1$): Corresponds to decaying exponentials
• Outside the unit circle ($|z| > 1$): Corresponds to growing exponentials
• On the unit circle ($|z| = 1$): Corresponds to sustained oscillations

This geometric interpretation makes stability analysis intuitive: if all poles of a system's transfer function lie inside the unit circle, the system is stable because all natural response components decay over time.

Common Z-Transform Pairs

These fundamental pairs form the building blocks for most Z-Transform analysis:

Notice that $a^n u[n]$ and $-a^n u[-n-1]$ produce the same algebraic expression $\frac{z}{z-a}$ but with different ROCs. This illustrates why specifying the ROC is essential for uniquely identifying the time-domain sequence.

Relationship to the Laplace Transform

The Z-Transform relates to the Laplace Transform through the exponential mapping:

$$z = e^{sT_s}$$

where $T_s$ is the sampling period. This mapping transforms the continuous s-plane into the discrete z-plane with the following correspondences:

• Left half of s-plane (Re$\{s\} < 0$) → Inside the unit circle ($|z| < 1$): stable poles
• Right half of s-plane (Re$\{s\} > 0$) → Outside the unit circle ($|z| > 1$): unstable poles
• Imaginary axis ($s = j\omega$) → Unit circle ($|z| = 1$): frequency response

This mapping is many-to-one: the entire $j\omega$ axis maps onto the unit circle with periodicity $2\pi/T_s$, which is why discrete-time systems exhibit frequency aliasing.

Transfer Function of Discrete-Time Systems

For a discrete-time LTI system, the input-output relationship in the z-domain is:

$$Y(z) = H(z) \cdot X(z)$$

The transfer function $H(z) = Y(z)/X(z)$ completely characterizes the system. Given a general linear constant-coefficient difference equation:

$$y[n] + a_1 y[n-1] + \cdots + a_N y[n-N] = b_0 x[n] + b_1 x[n-1] + \cdots + b_M x[n-M]$$

Taking the Z-Transform (using the time-shift property) yields:

$$H(z) = \frac{b_0 + b_1 z^{-1} + \cdots + b_M z^{-M}}{1 + a_1 z^{-1} + \cdots + a_N z^{-N}}$$

This rational function in $z^{-1}$ can be factored into poles and zeros, providing immediate insight into system behavior.

Pole-Zero Analysis

The transfer function can be written in factored form:

$$H(z) = \frac{b_0 \prod_{k=1}^{M}(z - z_k)}{\prod_{k=1}^{N}(z - p_k)}$$

where $z_k$ are the zeros and $p_k$ are the poles. The pole-zero plot in the z-plane reveals:

• Pole locations: Determine stability and natural response modes
• Zero locations: Determine where the frequency response has nulls
• Pole-zero proximity to unit circle: Determines resonant frequencies and bandwidth
• Pole angles: Determine oscillation frequencies in the impulse response

Practical Significance

The Z-Transform is the primary tool for:

1. Digital filter design: Specifying desired frequency response and deriving filter coefficients
2. Control system analysis: Analyzing stability and performance of discrete-time controllers
3. Difference equation solving: Converting recursive equations to algebraic form
4. Frequency response computation: Evaluating $H(z)$ on the unit circle
5. Signal reconstruction: Understanding sampling and interpolation through z-domain analysis

Key Takeaways

• The Z-Transform $X(z) = \sum x[n]z^{-n}$ is the discrete-time counterpart of the Laplace Transform
• The unit circle in the z-plane corresponds to the frequency axis — evaluating $X(z)$ there gives the DTFT
• Poles inside the unit circle indicate stability; poles outside indicate instability
• The mapping $z = e^{sT_s}$ connects continuous-time and discrete-time analysis
• $z^{-1}$ is the unit delay operator: $\mathcal{Z}\{x[n-1]\} = z^{-1}X(z)$
• Difference equations become polynomial equations in $z$, solvable by algebraic methods
• The Z-Transform forms the mathematical foundation for all digital filter design and analysis