Every time you stream music, compress an image, or use autocorrect, signal processing is running behind the scenes. The math can look intimidating — Fourier transforms, convolutions, Nyquist rates — but seeing these concepts in action changes everything.
Here are 6 free interactive visualizations that turn abstract DSP concepts into something you can actually feel.
1. Fourier Transform Family
URL: https://elysiatools.com/en/visualizations/fourier-transform-family
The Fourier Transform is arguably the most important algorithm in all of engineering. It converts any signal from the time domain (amplitude over time) into the frequency domain (what frequencies are present and how strong).
This interactive tool shows you exactly how that works:
- Time ↔ Frequency views — toggle between them and watch signals transform
- Fourier Series approximation — see how sine waves add up to reconstruct any periodic signal
- FFT Butterfly Diagram — visualize the Cooley-Tukey radix-2 algorithm that makes FFT computationally efficient
- DFT vs FFT performance — compare O(N²) vs O(N log N) complexity side by side
- Real-time audio spectrum — feed your microphone and watch the frequency content update live
Real-world impact: MP3 and AAC audio compression work by discarding frequencies you can't hear. JPEG image compression does the same in 2D. Every WiFi, 4G, and 5G connection uses OFDM — a frequency-domain technique built directly on the Fourier Transform. Without it, the internet as we know it doesn't exist.
2. Convolution
URL: https://elysiatools.com/en/visualizations/convolution
Convolution is the operation at the heart of almost every system: audio filters, image blurs, neural networks, and control systems all use it. But the definition — (f × g)(t) = ∫f(τ)g(t−τ)dτ — doesn't exactly scream "intuitive."
This visualization breaks it down step by step:
- Flip the kernel — the second signal is literally reversed
- Slide to position — move it along the first signal
- Multiply — at each overlap, multiply the values
- Sum — add up all the products
You can draw your own signals and watch the animation play through each step. This is the "aha moment" that college DSP courses spend weeks trying to deliver.
In practice: Convolution is how Instagram's blur filter works, how reverb is added to audio, and how the feature maps in a convolutional neural network (CNN) scan an image for edges and shapes.
3. PID Controller
URL: https://elysiatools.com/en/visualizations/pid-controller
PID (Proportional-Integral-Derivative) controllers are everywhere — from the thermostat in your home to the autopilot in a 747 to the cruise control in your car. They're how systems maintain a desired state by continuously adjusting based on error.
This interactive tool lets you tune a real control loop in real time:
- Proportional (P) — reacts to current error. Higher Kp = faster response, but can cause oscillation
- Integral (I) — accumulates past error to eliminate steady-state drift. Higher Ki = faster correction, but can cause overshoot
- Derivative (D) — predicts future error based on change rate. Higher Kd = less overshoot, but amplifies noise
You adjust the three parameters and watch the system response (rise time, settling time, overshoot, oscillation) change instantly.
In practice: A drone's altitude hold, a robot arm's precise positioning, and a solar panel tracking the sun all depend on well-tuned PID loops. Understanding them makes you a better engineer, period.
4. Sampling Theorem (Nyquist-Shannon)
URL: https://elysiatools.com/en/visualizations/sampling-theorem
Here's a shocking fact: a CD encodes audio at 44,100 samples per second — not because someone picked a round number, but because of a precise mathematical theorem.
The Nyquist-Shannon Sampling Theorem states: a continuous signal can be perfectly reconstructed from its samples only if the sampling rate is at least twice the maximum frequency in the signal.
This visualization shows you exactly what happens when you violate that rule:
- Aliasing in action — reduce the sampling rate and watch high-frequency components fold back into low frequencies as moiré-like artifacts
- Reconstruction methods — see how different interpolation algorithms attempt to rebuild the original signal from samples
- Nyquist frequency indicator — always know whether your current settings are above or below the safe threshold
In practice: If you've ever seen wagon-wheel aliasing in a film (spokes appearing to spin backward), that's aliasing. The 48 kHz sampling rate for professional audio, the frame rate of your camera sensor, and the GPS update rate in your phone — all determined by Nyquist.
5. Wavelet Transform
URL: https://elysiatools.com/en/visualizations/wavelet-transform
Fourier Transform tells you what frequencies exist in a signal — but it loses all time information. A chord and a single arpeggio can produce identical frequency spectra. Wavelet Transform fixes this by giving you both time and frequency simultaneously.
This visualization introduces two modes:
- Continuous Wavelet Transform (CWT) — uses scaled and translated versions of a "mother wavelet" to analyze signals at multiple resolutions
- Discrete Wavelet Transform (DWT) — efficient, power-of-2 decomposition used in compression
You can draw a custom signal and watch it decompose across scales. Different wavelets are available:
- Haar Wavelet — the simplest, piecewise constant, foundational for understanding
- Morlet Wavelet — sine wave modulated by Gaussian, ideal for time-frequency analysis
In practice: JPEG2000 image compression uses wavelets instead of the DCT (Discrete Cosine Transform) used in standard JPEG. ECG machines use wavelet denoising to clean up biomedical signals. Seismologists use them to detect and locate earthquakes.
The Signal Processing Toolkit You Didn't Know You Needed
All of these visualizations run entirely in your browser. No account. No download. No setup.
https://elysiatools.com/en/visualizations/fourier-transform-family
The common thread? All of them bridge the gap between the equations you'll find in a textbook and the intuition you need to apply them in real systems. Whether you're building audio plugins, training a neural network, designing a control system, or just want to understand why your phone's autocorrect works — these tools give you the mental model you need.
What's the one DSP concept you've always wanted to really understand? The answer might be a click away.
All tools are free at ElysiaTools.com — no sign-up required.
Top comments (0)