HN
Today

An Interesting Fourier Transform – 1/F Noise

This article delves into the fascinating world of Fourier transforms applied to power law functions, revealing a surprising property: the transform of a power law is often another power law. It highlights how this mathematical relationship might hold the key to solving the 80-year-old mystery of 1/f noise. The technical exploration into signal processing and unresolved physics makes it particularly compelling for the Hacker News audience.

13
Score
2
Comments
#6
Highest Rank
5h
on Front Page
First Seen
Aug 10, 7:00 AM
Last Seen
Aug 10, 11:00 AM
Rank Over Time
106667

The Lowdown

This piece explores the intricate properties of the Fourier transform when applied to power law functions, a concept common in scientific and engineering fields. It meticulously breaks down the mathematical relationships and their implications, ultimately connecting these abstract concepts to one of the enduring mysteries in physics: 1/f noise.

  • The fundamental insight is that the Fourier transform of a power law, t^α, yields another power law, ω^-(α+1), in the frequency domain.
  • However, this simple relation comes with nuances, including the presence of a unit step function, a non-power-law phase term, and a scaling factor involving the Gamma function.
  • Visualized on a log-log plot, the frequency domain magnitude of these transforms appears as a straight line with a slope of -(α+1).
  • Specific cases like α=0 (perfect integrator) and α=1 (two-stage integrator) are discussed, showing how well-behaved these transforms can be despite boundary conditions.
  • A particularly interesting edge case occurs as α approaches -1, where the Gamma function becomes undefined, and the time domain function approaches a delta function.
  • The author identifies a unique point at α = -0.5, where both the time domain (t^-0.5) and the frequency domain (magnitude ω^-0.5) exhibit the same power law decay rate.
  • This observation is then linked to 1/f noise, a pervasive and unexplained phenomenon found in diverse systems, whose amplitude spectrum decays as 1/f^(1/2) (or ω^-0.5).
  • The compelling hypothesis emerges that 1/f noise might be its own Fourier transform, similar to how a Gaussian curve transforms into itself, potentially offering a critical clue to its origins.
  • Despite this compelling connection, the article acknowledges remaining challenges, such as the unknown phase of 1/f noise and the lack of a clear physical interpretation for the corresponding u(t)t^-0.5 time-domain signal.

Ultimately, the article presents a potent mathematical connection that could be instrumental in deciphering the long-standing enigma of 1/f noise, suggesting that a self-transforming property might be a fundamental characteristic of this ubiquitous phenomenon.