Deterministic VS Random
Deterministic model
Signal which can be predicted without uncertainty by ”simple”descriptor (function, frequencies…)
Examples: song record, photo…
Tools: Fourier, Wavelets, Time-Frequency analysis, Ceptral analysis (Hilbert/Harmonic analysis)
Random model
Signal issued from “stochastic process”: each “realization” share some common properties, but each signal is different.
Example: noise
Tools: Probabilty, statistics, Estimation
In practice
Both model can be employed and mixed
Tools: Optimization, Bayesian approach