1/f noise

1/f noise (also called flicker noise or pink noise) is a signal or process with a power spectral density that falls off as the inverse of frequency: S(f) ∝ 1/f^α, where α is typically close to 1. It appears ubiquitously across physical, biological, and social systems — in semiconductors, heartbeats, neural activity, river flows, stock prices, and musical melodies. Its ubiquity is so striking that it has been called 'one of the oldest puzzles in contemporary physics.'

Unlike white noise (flat spectrum) or Brownian noise (1/f² spectrum), 1/f noise has no characteristic timescale. Events at all frequencies contribute equally to the total power when weighted logarithmically. This scale-invariance suggests that 1/f noise may be a signature of self-organized criticality, though the connection is debated. Other explanations include superposition of Lorentzian spectra with distributed relaxation times, multiplicative random processes, and systems with long-range temporal correlations.

The generating mechanisms of 1/f noise remain contested. In electronic devices, it is often attributed to trapping and detrapping of charge carriers at defect sites. In biological systems, it may reflect hierarchical control processes operating at multiple timescales. In financial markets, it may emerge from the interaction of agents with heterogeneous strategies and information horizons.

What is clear is that 1/f noise is not a single phenomenon with a single cause. It is a statistical signature that can be produced by many different mechanisms — a case where the same pattern masks different underlying dynamics. The systems-theoretic lesson is that scale-invariance alone is not sufficient to identify the generating mechanism. It is necessary but not diagnostic.