Overview: Stochastic Signals and Scale-Free 1/f Dynamics
In biomedical signal processing, neurophysics, and electronic instrumentation, noise is rarely a simple, uncorrelated random variable. While uncorrelated white noise exhibits a flat power spectral density ($S(f) = \text{constant}$), natural biological phenomena—most prominently electroencephalographic (EEG) scalp potentials, heart rate variability (HRV), and membrane channel gating kinetics—demonstrate scale-free, self-similar temporal fluctuations known as $1/f^\alpha$ noise (where $\alpha \approx 1$). Often referred to as pink noise or flicker noise, inverse-frequency dynamics bridge the deterministic world of simple harmonic oscillators and the purely unpredictable realm of white noise.
In resting-state EEG, baseline cortical activity is characterized by a prominent $1/f$ aperiodic spectral background upon which specific physiological rhythms (such as delta 0.5–4 Hz, theta 4–8 Hz, alpha 8–13 Hz, beta 13–30 Hz, and gamma >30 Hz) are superimposed. The steeper the aperiodic exponent $\alpha$, the greater the structural integration and inhibitory tone across large-scale neural assemblies. This interactive simulator provides a high-fidelity laboratory sandbox to generate, shape, filter, and mathematically dissect various noise topographies in real time—ranging from uniform white noise and Gaussian distributions to periodic oscillatory interference and custom fractional $1/f^\alpha$ noise architectures.
Technical Details & Implementation Architecture
To provide smooth real-time execution without external network dependencies or layout shifts, the engine relies on client-side matrix computations, Box-Muller transforms, multi-rate Paul Kellet IIR pinking filter cascades, and Cooley-Tukey Radix-2 Fast Fourier Transforms.
Gaussian Random Generation: Independent normal variables $Z \sim \mathcal{N}(0, 1)$ are synthesized using the exact Box-Muller transform:
Z = \sqrt{-2 \ln U_1} \cdot \cos(2\pi U_2), \quad U_1, U_2 \sim \text{Uniform}(0, 1)
Pink (1/f) Noise Synthesis: True $1/f$ spectral slope requires an infinite sum of exponential relaxation processes. The simulator uses a multi-stage pole-zero filter system configured to approximate a $-3\text{ dB/octave}$ ($-10\text{ dB/decade}$) roll-off across several decades of frequency response without phase distortion.
DPI Normalization & Safety Bounds: Canvas rendering contexts query window.devicePixelRatio to configure bitmap buffer dimensions dynamically, guaranteeing anti-aliased vector rendering on Retina and 4K displays. Computational guardrails wrap every numerical iteration to intercept non-finite or NaN floating-point overflow events, seamlessly restoring baseline default states.