Signal Visualization

Introduction

Effective visualization transforms abstract mathematical signals into intuitive understanding. A well-crafted plot can reveal periodicities, transients, spectral content, and system behavior that equations alone cannot convey. Signal processing relies on multiple complementary visualization techniques because no single plot type captures all aspects of a signal. Time-domain plots show temporal evolution, frequency-domain plots show spectral content, and time-frequency displays (spectrograms) show how the spectrum evolves over time.

Mastering signal visualization is a practical skill that serves you in coursework (verifying calculations), laboratory work (debugging measurements), and professional practice (presenting results to colleagues and clients).

Time-Domain Visualization

Continuous-Time Waveform Plot

The most basic visualization: signal amplitude versus time. Use for:

• Observing transient behavior (startup, decay)
• Identifying periodicity and period
• Detecting discontinuities and anomalies
• Comparing input and output waveforms

Best practices:

• Label axes with units (Time in seconds, Amplitude in volts)
• Choose time span to show several complete cycles (for periodic signals)
• Use grid lines for reading values
• Overlay multiple signals for comparison (different colors/styles)

Discrete-Time Stem Plot

For discrete sequences, stem plots show individual sample values as vertical lines with dots. This correctly represents the discrete nature — connecting dots with lines (as in plot) falsely implies values exist between samples.

Multi-Panel Plots

Compare related signals using subplots:

Frequency-Domain Visualization

Magnitude Spectrum

Plot $|X(j\omega)|$ or $|X(f)|$ vs frequency. Use linear scale for narrowband signals, dB scale ($20\log_{10}|X|$) for wideband signals or when dynamic range is large.

Phase Spectrum

Plot $\angle X(j\omega)$ vs frequency. Unwrap phase for continuous display:

Bode Plot (System Frequency Response)

Two-panel plot showing magnitude (dB) and phase (degrees) vs log-frequency. Standard for characterizing filters and systems:

Time-Frequency Visualization

Spectrogram

Shows how frequency content changes over time — essential for non-stationary signals like speech, music, and biomedical data:

The spectrogram reveals:

• Formant transitions in speech
• Harmonic structure of musical instruments
• Transient events (clicks, impulses)
• Frequency modulation (chirps, vibrato)

Scalogram (Wavelet Time-Frequency)

Wavelets provide better time resolution at high frequencies and better frequency resolution at low frequencies — complementary to the fixed-resolution spectrogram.

Pole-Zero Plots

Display system poles (×) and zeros (○) on the complex plane with the unit circle:

Reveals stability (all poles inside unit circle), resonant frequencies (pole angles), and bandwidth (pole distances from unit circle).

Signal Space and Phase Portraits

Constellation Diagrams

For digital communications: plot the in-phase vs quadrature components of received symbols. Reveals modulation scheme (QAM, PSK) and signal quality.

Phase Portrait

Plot $x(t)$ vs $\dot{x}(t)$ (or $x[n]$ vs $x[n-1]$) to visualize system dynamics:

• Limit cycles appear as closed loops
• Stable equilibria appear as spiral convergence
• Chaos appears as complex non-repeating patterns

3D and Animated Visualization

Surface plots: Display the z-transform $|X(z)|$ over the z-plane to visualize how poles create peaks and zeros create valleys.

Animated convolution: Step through the flip-and-slide process frame by frame to build intuition for how convolution works.

Practical Visualization Guidelines

1. Always include axis labels and titles — a plot without labels is useless
2. Choose appropriate scales: linear for signals, log/dB for spectra with large dynamic range
3. Use color meaningfully: different signals in different colors, consistent across subplots
4. Match visualization to the question: use time plots for transient questions, frequency plots for spectral questions
5. Zoom appropriately: show enough context to interpret, but enough detail to read

Key Takeaways

• Time-domain plots reveal temporal structure; frequency-domain plots reveal spectral content
• Spectrograms bridge both domains for non-stationary signal analysis
• Stem plots (not line plots) correctly represent discrete-time sequences
• Bode plots (magnitude + phase vs log-frequency) are standard for system characterization
• Pole-zero plots provide immediate stability and frequency response insight
• Always label axes, include units, and choose scales appropriate to the data