An Interesting Fourier Transform - 1/f Noise

Steve SmithNovember 23, 200713 comments

Power law functions are common in science and engineering. A surprising property is that the Fourier transform of a power law is also a power law. But this is only the start- there are many interesting features that soon become apparent. This may even be the key to solving an 80-year mystery in physics.

It starts with the following Fourier transform:

Fourier transform of power law

This article is available in PDF format for easy printing

The general form is tα ↔ ω-(α+1), where α is a constant. For example, t2 ↔ ω–3 and t -0.75 ↔ ω–0.25. Unfortunately, there are additional terms that distort this simple relation. First, the left side contains the term, u(t), the unit step function. This is defined as u(t) = 0 for t < 0, and u(t) = 1 for t ≥ 0. In other words, this makes the time domain a one-sided power law. Second, the power law in the frequency domain only pertains to the magnitude; there is a phase term that doesn't resemble a power law at all. Third, there is a scaling factor in the frequency domain, Γ(α +1). This is the Gamma function, which is essentially a continuous version of factorials. A graph of the Gamma function is shown below. Don't worry too much about this strange function. Think of it simply as a constant that scales the amplitude of the frequency domain, depending on the value of α.

The Gamma function

The table below shows nine cases of this Fourier transform, with alpha running from -2.0 to 2.0, and rough sketches of the curves. The frequency domain also shows a rough sketch of the magnitude graphed on a log-log plot, which turns out to be is a straight line with a slope of -(α+1). Take a few minutes to examine this figure, especially noting the symmetry between the time and frequency domains.

FT power law pairs

One of these cases should be familiar to you, where α=0. This is the Fourier transform of the unit step function, with a magnitude of 1/ω, and a phase of -π/2. As you probably recall, this describes the impulse and frequency response of the perfect integrator. Now consider the case of an integrator followed by another integrator. The impulse response of this two stage combination is the unit step response convolved with itself. In the frequency domain the magnitude becomes 1/ω × 1/ω, and the phase becomes 2 × (-π/2). This two-integrator cascade is shown on the graph for α=1, where the impulse response is a linearly increasing line, and the frequency spectrum is 1/ω2 , with φ = -π. Likewise, α=2 represents a cascade of three integrators, and so on. It is interesting that these Fourier transforms are so well behaved, in spite of both domains containing nasty features (such as: t2 as t → ∞ , and ω-3 as ω → 0 ).

Here is something even more interesting. As you approach α = -1, the time domain approaches a shape of t-1, and the frequency domain approaches a flat magnitude with a zero phase. However, a flat magnitude and zero phase corresponds to a delta function, δ(t), in the time domain. That is, in the limit as α → -1, u(t)t = δ(t). This is because the sharp point of t grows rapidly as α → -1, dominating the entire function. But what happens exactly at α = -1? How could the sharp point ever completely negate the seemingly finite width of t-1? The mathematics tells you not to ask this question. Recall that the frequency domain has a scaling factor of Γ(α+1). For α = -1, the gamma function is undefined.

Now we come to a feature that I find absolutely fascinating. In general, we have seen that a power law in one domain corresponds to a power law in the other domain. Further, there is an inverse relationship; if the time domain decays faster, then the frequency domain decays slower, and vice-versa. This means that there must be a certain decay rate that is unique, where both domains are equal. This occurs for α = -0.5, where the time domain is t-0.5 and the frequency domain is ω-0.5. Interesting, but what does this mean?

Now look at the figure below, a graph of the measured noise that originates within a common electronic amplifier. The flat section above 100 Hz is called white noise , and is well understood. However, the sloping portion below 100 Hz is not well understood at all. This is 1/f noise, a mystery that has resisted explanation for over 80 years. 1/f noise has been observed in the strangest places- electronics, traffic density on freeways, the loudness of classical music, DNA coding, and many others.

1/f noise example

Many of the properties of noise are directly related to the amount of power in a signal, that is, to the square of the amplitude. Accordingly, most of those working with noise think in terms of power spectra, not amplitude spectra. 1/f noise gets its name because its power spectrum has a shape that is close to 1/f. However, if we look at the amplitude spectrum for 1/f noise it has a shape of 1/f1/2. As can be seen above, on a log-log plot of amplitude, 1/f noise has a slope of -0.5. Now you can see where I’m going.

At least in a limited sense, 1/f noise is its own Fourier transform, with ω-1/2 in the frequency domain, and t-1/2 in the time domain. For instance, a single pulse given by u(t) t-1/2 has a 1/f power spectrum. Likewise, a randomly occurring sequence of such pulses has a 1/f power spectrum, at least over a wide range frequencies. Further, 1/f noise can be created by passing white noise through a filter with an impulse response of u(t) t-1/2. Unfortunately, none of these scenarios seems to have a physical interpretation that explains the widespread observation of 1/f noise. It’s clear that something is still missing.

There is also another issue: u(t) t-1/2 is the transform pair of Mag = ω-1/2 , φ = -π/4. However, no one knows what the phase of 1/f noise is, or even if it has a defined phase. If the phase happens to be different from φ = -π/4, then the corresponding time domain signal will also be different. That is, it may be that the characteristic time domain waveform of 1/f noise is simply not t-1/2.

Nevertheless, the idea that 1/f noise is its own Fourier transform is very compelling. Consider the Gaussian curve, the most important waveform associated with random events. The Central Limit Theorem tells us why the Gaussian is so commonly observed. However, it is also true that the Fourier transform of a Gaussian is a Gaussian. This seems more than coincidence– I think it is a critical clue in solving the mystery of 1/f noise.

Comments are certainly welcome! I especially appreciate suggestions for new directions in research and (god forbid) errors in math.

Steve Smith email


Next post by Steve Smith:
   Waveforms that are their own Fourier Transform


[ - ]
Comment by inhaheMay 21, 2009
i'm not very strong in math, so here's my simplistic take on 1/f noise: it's already well-known that lower frequencies travel and permeate better than higher frequencies. take sound, light, or whatever. for example that's why elephants can communicate at sub-audial (for humans) frequencies over ~100 miles. so my theory is that if you have a system whose different parts, and/or even the integrated whole are generating waves at varying frequencies, the lower the frequency the better it'll permeate throughout the system, and hence at any point in the system all those lower frequencies generated anywhere within it will add up, and the higher ones won't. might this principle lead to a 1/f spectral envelope?
[ - ]
Comment by JBBNovember 21, 2010
I am reading a book about paleoclimatology, Ice Ages and Astronomical Causes, which goes into considerable detail about analyzing various climate proxy time series for spectral evidence of astronomical influence. The authors make extensive use of "Monte Carlo" Fourier analysis using Matlab. One interesting claim they make is that if you run a bunch of MC studies, creating "spectra" by simply choosing a series of random numbers, the resulting power spectra come out "pink". They don't exactly say the spectra are 1/f, but they don't rule it out, either. It seems to me that if this is true, then it should be a theorem. That is, it should be possible to show that, given an ensemble of all possible spectra of a certain length and resolution, the most probable power spectrum has certain characteristics. Sort of a thermodynamics approach to Fourier Transforms. I don't have Matlab. I am going to throw something together in C++ and see what I can discover experimentally, before I try to approach it analytically.
[ - ]
Comment by nonanonFebruary 11, 2016
Just came across this interesting blog! The convolution of the power law with a Gaussian has similar properties. We call it the Q function. It is essentially a parabolic cylinder function. It describes with great precision the dose profile of a proton beam, with a so-called Bragg peak. This has applications in cancer medicine, where proton beam therapy is becoming more common. I am curious if Q functions have many other applications, they probably do.
[ - ]
Comment by gerryharpApril 18, 2016
This blog is very interesting, as is this comment. If I understand what you're saying, then the FT of a power law is a power law, the FT of a gaussian is a gaussian, and the FT of their product is also a function of the same form. That is pretty neat!

Does anyone have a reference for the result that the FT of a power law is a power law? I have been hunting for one and can't find it. Thanks
[ - ]
Comment by bubble41January 29, 2008
Dear Steve, Thank you for a very interesting discussion on 1/f noise. Time series data in the real world cover a limited time span. So a time series signal that looks like a straight line with a unit slope, when truncated, becomes a sawtooth. The amplitude of the frequency spectra of a sawtooth is 1/ω, and not the 1/ω2 for that of a linearly increasing line. I guess my point was that real life time data which had been used to demonstrate 1/f noise so far were always truncated in time. Would that consideration present any problem for your consideration of the Fourier transform of 1/f noise? Please advise. Woo, a student new to fft
[ - ]
Comment by rajen.iitdApril 21, 2008
Dear Dr. Smith, I am a working on a paper which deals with fractional Fourier transform. Can you please provide me some material about FrFT, specially I am interested in knowing the physical/geometrical significance of FrFT, something which can explain me how FrFT of an image looks like and Inverse FrFT will look like? Sincerely - Rajendra Uppal
[ - ]
Comment by SteveSmithApril 21, 2008
Hi Rajendra, I've never worked with FrFT, so I don't think I can give you much help. It sound interesting-- if you have a chance I'd like to see some information on it. Thanks, Steve
[ - ]
Comment by ayodeiMarch 13, 2009
I a generated a signal using fourier series and measured it at three different sensors by simulation of a model. I now want to reconstruct back my original signal. I carried out an fft on the three signals from the sensors and know if I can extract their amplitudes and frequencies, I can use that to get back my original signal using the fourier series. My problem now is how to extract this frequencies and also, if I should analyze all the signals from three sensors or one will be ok. Thank you. Ayodei
[ - ]
Comment by SteveSmithMarch 13, 2009
Hi Ayodei, Fee free to contact me directly at Steve.Smith AT Regards, Steve
[ - ]
Comment by gshekharJune 4, 2009
Dear all, I have a doubt though it is not linked to the current discussion. How do we calculate the snr of a signal only from its sampled observations and when the noise variance is not known a priori. Assume y[n] = x[n] + noise where x[n] is the original exponential signal and y[n] is the noise corrupted signal. Now is there a way to estimate snr only from observations y[n]. Is there a way to measure noise variance and power of signal x[n] just from y[n].
[ - ]
Comment by BruecknerJune 24, 2009
Not sure if this helps the thought process but, maybe it is good to consider photons. A photon is its own antiparticle. Of course, once you get onto that road, you would have to think four dimensionally! :)
[ - ]
Comment by dspchandNovember 11, 2016

Thanks Steve. This is a wonderful and stimulating article. I like this kind of observations/interpretations. This is what makes math 'beautiful'. The moment I read this, the first thing that strikes my mind is the theory of eigenvalues and vectors. So many problems become much easier when though-off as matrices. When the signal is sampled, FT is replaced with DFT which can simply be denoted by the multiplication of matrix and vector. If F is the DFT matrix and x is a vector, then xf=Fx, and xf is the DFT of x. We are looking at vectors (signals) x that satisfy Fx = \lamda x where \lamda is an eigenvalue of the transformation x. We know from matrix theory that there could be an infinite number of eigenvectors for a matrix A but then no of non-zero eigenvalues cannot exceed the rank of matrix A, and that the space spanned by the eigenvectors is the null space of A and its dimensionality is determined by the rank of A as well. In your other article 'waveforms that are their own fourier transform' , you clearly show that the no of eigenfuncions (functions who shape remains the same under FT) is infinite. The same conclusion can be achieved through linear algebra as well. For DFT, the eigenvectors would span a finite-dimensional space N and for continuous FT, the eigenfunctions would span an infinite dimensional space and so on.

Anyways, coming back to the interpretation of the 1/f noise's PSD being an eigenfunction of FT - since Gaussian function can be expanded into polynomials, it is not very surprising that power law functions satisfy the same property. I think the uniqueness of flicker noise is the following. While the probability distribution itself is an eigenfunction (Gaussian), it time correlation is also an eigenfunction. Dont know if this would help solve the mystery. But enjoyed it.

[ - ]
Comment by GerhardKJanuary 16, 2017

Hi Steve,

"Unfortunately, none of these scenarios seems to have a physical interpretation that explains the widespread observation of 1/f noise. It’s clear that something is still missing."

Some thoughts about that:

Electrical current is represented by moving electrons in some conductor or semi-conductor. This materials all have a internal cristallin structure. And as an universal rule told us, nothing is perfect, this cristallin structures contains damages effecting the flow of electrons somehow.

Guess we have an artificial watercourse with really smooth edges. Now we put leaves on it, normal distributed, and count the leaves comming along per time unit some where down the stream. We will see, the recorded numbers will follow more or less a normal distribution.

Guess we have a natural watercourse with rough edges (Banks). We do the same experiment. This time some leaves will stuck at a point along the banks. After a while some other leave will stuck at the same position, and we will see a growing collection of leaves stucking at this position. The recorded numbers of leaves is less now. One more leave and the collection break free. In opposite to its 'construction' this collection breaks free and swim down as a collection of leaves. This produces a high peak in our counts of leaves per time unit. (Some times flicker noise is also called Popcorn noise, especially by Audio engineers)

And it needs time for the collection to grow. So, lager collections (peaks) will be observed very seldom relative to smaller collections ==> 1/f.

Unfortunately, electrons carrying a negative charge and so want to maximize her distance, leaves do not. I am not that good in physics or mathematics to overcome that problem, its just an idea.

Sorry for my english, hope you get the big picture ....

With best regards


To post reply to a comment, click on the 'reply' button attached to each comment. To post a new comment (not a reply to a comment) check out the 'Write a Comment' tab at the top of the comments.

Registering will allow you to participate to the forums on ALL the related sites and give you access to all pdf downloads.

Sign up
or Sign in