Compute poles, zeros, region of convergence (ROC), stability, impulse response, and frequency response magnitude (dB) from rational transfer function H(z) = B(z)/A(z). Visualize pole-zero plot on the z-plane with unit circle.
In digital signal processing, the Z-transform converts a discrete-time signal into a complex frequency domain representation: \( H(z) = \frac{b_0 + b_1 z^{-1} + \dots + b_M z^{-M}}{a_0 + a_1 z^{-1} + \dots + a_N z^{-N}} \). The roots of the denominator (poles) determine system stability and natural response, while the roots of the numerator (zeros) shape frequency selectivity. For a causal system, the region of convergence (ROC) is \( |z| > \max|p_i| \), and the system is BIBO stable iff all poles lie strictly inside the unit circle (\( |p_i| < 1 \)).
Frequency response: \( H(e^{j\omega}) = \frac{B(e^{j\omega})}{A(e^{j\omega})} \)
Magnitude in dB: \( 20\log_{10}|H(e^{j\omega})| \)
Given numerator coefficients \([b_0, b_1, ..., b_M]\) and denominator \([a_0, a_1, ..., a_N]\) (normalized so \(a_0=1\)), we construct polynomial forms: \(B(z) = b_0 z^M + b_1 z^{M-1} + ... + b_M\) and \(A(z) = a_0 z^N + a_1 z^{N-1} + ... + a_N\). Roots are solved using a robust eigenvalue method (companion matrix) via the numeric.roots library. Zeros are roots of \(B(z)\), poles are roots of \(A(z)\). The ROC is determined by the outermost pole magnitude: for causal systems ROC = \(|z| > r_{max}\). Stability is satisfied when all pole magnitudes < 1. The impulse response is computed via recursive difference equation: \(h[n] = b_n - \sum_{k=1}^{N} a_k h[n-k]\) for \(n \ge 0\), with \(b_n=0\) beyond M. Frequency response is evaluated at 512 points on the unit circle directly from the rational function.
An audio engineer designs a simple reverberation filter using the transfer function \( H(z) = \frac{1}{1 - 0.7 z^{-1}} \). Our calculator reveals a pole at \(z = 0.7\) (inside unit circle → stable), ROC \(|z|>0.7\). The impulse response decays exponentially: \(h[n] = 0.7^n\), producing a natural echo decay. The magnitude response shows a lowpass characteristic with -3dB around 0.3π rad/sample. Such filters are building blocks for digital audio effects.