harmonic-entropy

Compute harmonic entropy of musical intervals.

Installation

npm i harmonic-entropy
Copy

Documentation

• API docs (TypeDoc): https://xenharmonic-devs.github.io/harmonic-entropy
• Generate docs locally:

npm run doc
Copy

What this package provides

This package exposes:

• precalculateRatios(options): precomputes rational pairs used by the entropy algorithm.
• harmonicEntropy(options, ratios): computes a tabulated entropy curve as [cents, entropy] pairs.
• EntropyCalculator: convenient class for evaluating intervals (ofCents, ofFraction) with cached tables.

Options reference

HarmonicEntropyOptions fields (all optional):

• N: max rational complexity.
  • default for series: 'tenney': 10000
  • default for series: 'farey': 1000
• s: Gaussian frequency deviation (default 0.01)
• a: Rényi order (default 1, evaluated numerically near Shannon entropy)
• series: 'tenney' | 'farey' (default 'tenney')
• minCents: lower tabulation bound (default 0)
• maxCents: upper tabulation bound (default 2400)
• res: tabulation step in cents (default 1)
• normalize: divide by Hartley entropy term (default false)

Example: entropy table for a range

import {
  type HarmonicEntropyOptions,
  harmonicEntropy,
  precalculateRatios,
} from 'harmonic-entropy';

const options: HarmonicEntropyOptions = {
  N: 10000,
  s: 0.01,
  a: 1,
  series: 'tenney',
  minCents: 0,
  maxCents: 2400,
  res: 1,
  normalize: false,
};

// Compute the set of rational numbers to consider.
const ratios = precalculateRatios(options);

// Compute the table of [cents, entropy] pairs. Entropy is measured in natural (base e) units.
const table = harmonicEntropy(options, ratios);

// This would be replaced by passing the table your favorite plotting library.
console.log(table);

/*
[
  [0, 2.465367706139234],
  [1, 2.4695232705775982],
  [2, 2.481975530994572],
  [3, 2.5026079139822173],
  [4, 2.5312627678840913],
  [5, 2.5676902925403278],
  ...
  [2399, 3.900196510219676],
  [2400, 3.898697709259859],
]
*/
Copy

Example: evaluate individual intervals

import {EntropyCalculator} from 'harmonic-entropy';

const entropy = new EntropyCalculator({maxCents: 1200});

// Evaluate by cents.
const perfectFifth = entropy.ofCents(700);

// Evaluate by ratio.
const pureFifth = entropy.ofFraction('3/2');

// Numeric input is interpreted as a frequency ratio.
const sameFifth = entropy.ofFraction(3 / 2);

// Tablulated values are linearly interpolated
const majorSixth = entropy.ofFraction('5/3');

console.log({perfectFifth, pureFifth, majorSixth});
/*
{
  perfectFifth: 4.126342260377048,
  pureFifth:    4.121900707092091,
  majorSixth:   4.42017406844399,
}
*/
Copy

Serialization and revival

import {EntropyCalculator} from 'harmonic-entropy';

const calc = new EntropyCalculator({N: 1000});
const serialized = JSON.stringify(calc);

const revived = JSON.parse(serialized, EntropyCalculator.reviver);

console.log(revived.ofCents(700));
// 4.126342260377048
Copy

Notes and caveats

• Constructing EntropyCalculator performs precomputation immediately.
• Lower res (finer step) usually improves interpolation fidelity but costs more memory/time.
• ofCents rejects values outside [minCents, maxCents].

Background reading

• XenWiki article on harmonic entropy and consonance: https://en.xen.wiki/w/Harmonic_entropy