Signals you cannot write down. Random processes described by statistics rather than formulas: autocorrelation, power spectral density, and the Wiener-Khinchin theorem that shows they are one object seen from two sides. Then white noise, filtering a random signal, the matched filter and why only pulse energy matters, processing gain, and detection as a threshold trade. Quantisation noise and the 6.02 dB rule derived rather than quoted, and why dither trades distortion for noise deliberately. It closes with spectral estimation (the periodogram is unbiased and inconsistent; Welch buys variance with resolution), the Wiener and adaptive filters, the analytic signal, and multirate resampling. Scoped against fourier, which owns the entire deterministic side: series, transform, convolution, DFT and FFT, sampling and Nyquist-Shannon, aliasing, the z-transform, FIR against IIR filters, windowing and wavelets.
Free to start · adaptive placement finds your level · reviews timed so it stays learned.
Every idea is taught with motivation and a worked example before the drills, and an FSRS spaced-repetition engine schedules each review for the moment just before you'd forget it. A short placement check finds what you already know, so you start Statistical Signal Processing exactly where it's useful.