Skip to main content

šŸ¤“ Fractal Music, even NERDS make music

  • August 30, 2026
  • 3 replies
  • 38 views

CharlyMX
Superuser

Ā 

Spoiler alert: This post is super NERDY, proceed at your own risk 😊

If you thought mathematicians were only good for splitting the bar bill or failing high schoolers, think twice. Turns out the biggest geeks in class have spent centuries messing around with music, but with fractal music, they took things to another level. There's no broken-heart inspiration here; it's pure equations turned into rhythm.

What on earth is a fractal?

A fractal is a geometric figure that repeats itself all the way to infinity. If you zoom in on a corner, it looks exactly like the full image. Think of a head of broccoli: if you snap off a tiny floret, it looks like a miniature head of broccoli. Nature is full of them, from snowflakes to ferns.

And how does music fit into this?

Simple: it translates those infinite-repetition formulas into notes, rhythms, and melodies.

Instead of composing by ear, a mathematician takes a numerical pattern and sets up rules: "every time the equation outputs a 2, play a C; when it goes back up, speed up the tempo." The result is a structure where the main melody repeats on both a macro and micro scale at the same time.

The DNA of Math on the Sheet Music

The relationship between numbers and sound isn't new. Pythagoras was already measuring lyre strings way before it was cool. However, fractal music uses specific concepts to build tracks:

  • Self-similarity: A short musical motif repeats inside the larger melody, but at a different speed or across different octaves.
  • 1/f Noise (Pink Noise): The sweet spot between predictability and chaos. Commercial music follows this pattern without even realizing it: not boring enough to put you to sleep, but not chaotic enough to sound like an old blender.
  • Recursive Algorithms: Math formulas that feed on their own output to generate the next measure.

Examples to Train Your Ears

So this doesn't just stay as chalkboard theory, here are a few clear cases where fractals and math turned into audio tracks:

  • Johann Sebastian Bach – Canon by Augmentation: Bach was a geek of his era without even knowing what a fractal was. In his canons, one voice plays the main melody while another plays the exact same sequence, but at half the speed. It's pure self-similarity written on the staff.
  • Harlan Brothers – Mandelbrot Music: Modern composers who directly use the coordinates of the Mandelbrot Set (one of the most famous fractal equations) to map sound frequencies. It's literally like listening to the graph of a function.
  • Gyƶrgy Ligeti – Piano Etudes: This Hungarian composer was explicitly inspired by fractal geometry and African rhythms to create hyper-complex piano pieces where multiple rhythms move forward simultaneously at different scales.
  • Generative Sound Algorithms (MusiNum): Software tools where you feed in a mathematical equation and the program spits out a full ambient fractal piece.

Next time you hear a song with a rhythm that hooks you without you knowing why, remember there's probably a fractal structure behind it tickling your brain.

Ā 

Personal note: NERDS RULE!

3 replies

Nina Nebo
Superuser
Forum|alt.badge.img+17
  • Superuser
  • August 30, 2026

Ā 

Ā 

Let's clarify a bitĀ šŸ˜šŸ¤“šŸ˜œ

Ā 

Pink noiseĀ Definition

Ā 

Within the scientific literature, the term "1/fĀ noise" is sometimes used loosely to refer to any noise with a power

spectral density of the form

Ā 

{\displaystyle S(f)\propto {\frac {1}{f^{\alpha }}},}

where f is frequency, and 0 < α < 2, with exponent α usually close to 1. One-dimensional signals with α = 1 are usually called pink noise.

Ā 

1D signal

For pink noise constrained to continuous frequencies fromĀ kminĀ toĀ kmax, the autocorrelation coefficient isr(d)=Ci(2Ļ€kmaxdN)āˆ’Ci(2Ļ€kmindN)log⁔kmaxkmin,{\displaystyle r(d)={\frac {{\textrm {Ci}}\left({\frac {2\pi k_{\textrm {max}}d}{N}}\right)-{\textrm {Ci}}\left({\frac {2\pi k_{\textrm {min}}d}{N}}\right)}{\log {\frac {k_{\textrm {max}}}{k_{\textrm {min}}}}}},}whereĀ Ci(x)Ā is theĀ cosine integral function.

Ā 

Ā 

If instead, the pink noise is approximated by a discrete sum of frequenciesĀ k, the Pearson autocorrelation coefficient isr(d)=āˆ‘k1kcos⁔2Ļ€kdNāˆ‘k1k.{\displaystyle r(d)={\frac {\sum _{k}{\frac {1}{k}}\cos {\frac {2\pi kd}{N}}}{\sum _{k}{\frac {1}{k}}}}.}

Ā 

Ā 

Ā 

2D signal

Ā 

The Pearson's autocorrelation coefficient of a two-dimensional pink noise signal comprising discrete frequencies is theoretically approximated as:r(d)=āˆ‘kJ0(2Ļ€kdN)kāˆ‘k1k,{\displaystyle r(d)={\frac {\sum _{k}{\frac {J_{0}({\frac {2\pi kd}{N}})}{k}}}{\sum _{k}{\frac {1}{k}}}},}whereĀ J0Ā is theĀ Bessel function of the first kind.

Ā 

Ā 

DefinitionĀ Self-similarity

Ā 

A compact topological space X is self-similar if there exists a finite set S indexing a set of non-surjective homeomorphisms {fs:s∈S}{\displaystyle \{f_{s}:s\in S\}} for which 

X=ā‹ƒs∈Sfs(X){\displaystyle X=\bigcup _{s\in S}f_{s}(X)}

IfĀ XāŠ‚Y{\displaystyle X\subset Y}, we callĀ XĀ self-similar if it is the onlyĀ non-emptyĀ subsetĀ ofĀ YĀ such that the equation above holds forĀ {fs:s∈S}{\displaystyle \{f_{s}:s\in S\}}. We call.

L=(X,S,{fs:s∈S}){\displaystyle {\mathfrak {L}}=(X,S,\{f_{s}:s\in S\})}

aĀ self-similar structure. The homeomorphisms may beĀ iterated, resulting in anĀ iterated function system. The composition of functions creates the algebraic structure of aĀ monoid. When the setĀ SĀ has only two elements, the monoid is known as theĀ dyadic monoid. The dyadic monoid can be visualized as an infiniteĀ binary tree; more generally, if the setĀ SĀ hasĀ pĀ elements, then the monoid may be represented as aĀ p-adicĀ tree.

The group ofĀ automorphismsĀ of the dyadic monoid is theĀ modular group; the automorphisms can be pictured asĀ hyperbolic rotationsĀ of the binary tree.

Ā 

Ā 

Ā 

Ā 

Recursive Algorithms

Ā 

Ā 


CharlyMX
Superuser
  • Author
  • August 30, 2026

Let's clarify a bitĀ šŸ˜šŸ¤“šŸ˜œ

Ā 

​@Nina Nebo,Ā  And here I thought I was the nerd 😁😁😁😁😁


CharlyMX
Superuser
  • Author
  • August 30, 2026

Ā