pitchfinder

A compilation of pitch detection algorithms for Javascript. Supports both the browser and node.

Provided pitch-finding algorithms

YIN - The best balance of accuracy and speed, in my experience. Occasionally provides values that are wildly incorrect.
AMDF - Slow and only accurate to around +/- 2%, but finds a frequency more consistenly than others.
Dynamic Wavelet - Very fast, but struggles to identify lower frequencies.
YIN w/ FFT (coming soon)
Goertzel (coming soon)
Mcleod (coming soon)

Installation

npm install --save pitchfinder

Usage

Finding the pitch of a wav file in node

All pitchfinding algorithms provided operate on Float32Arrays. To find the pitch of a wav file, we can use the wav-decoder library to extract the data into such an array.

const fs = require("fs");
const WavDecoder = require("wav-decoder");
const Pitchfinder = require("pitchfinder");

// see below for optional configuration parameters.
const detectPitch = Pitchfinder.YIN();

const buffer = fs.readFileSync(PATH_TO_FILE);
const decoded = WavDecoder.decode.sync(buffer); // get audio data from file using `wav-decoder`
const float32Array = decoded.channelData[0]; // get a single channel of sound
const pitch = detectPitch(float32Array); // null if pitch cannot be identified

Finding the pitch of a WebAudio AudioBuffer in the browser

This assumes you are using an npm-compatible build system, like Webpack or Browserify, and that your target browser supports WebAudio. Ample documentation on WebAudio is available online, especially on Mozilla's MDN.

import * as Pitchfinder from "pitchfinder";

const myAudioBuffer = getAudioBuffer(); // assume this returns a WebAudio AudioBuffer object
const float32Array = myAudioBuffer.getChannelData(0); // get a single channel of sound

const detectPitch = Pitchfinder.AMDF();
const pitch = detectPitch(float32Array); // null if pitch cannot be identified

Finding a series of pitches

Set a tempo and a quantization interval, and an array of pitches at each interval will be returned.

const Pitchfinder = require("pitchfinder");
const detectPitch = Pitchfinder.YIN();

const frequencies = Pitchfinder.frequencies(detectPitch, float32Array, {
  tempo: 130, // in BPM, defaults to 120
  quantization: 4, // samples per beat, defaults to 4 (i.e. 16th notes)
});

// or use multiple detectors for better accuracy at the cost of speed.
const detectors = [detectPitch, Pitchfinder.AMDF()];
const moreAccurateFrequencies = Pitchfinder.frequencies(
  detectors,
  float32Array,
  {
    tempo: 130, // in BPM, defaults to 120
    quantization: 4, // samples per beat, defaults to 4 (i.e. 16th notes)
  }
);

Configuration

All detectors

sampleRate - defaults to 44100

YIN

threshold - used by the algorithm
probabilityThreshold - don't return a pitch if probability estimate is below this number.

AMDF

minFrequency - Lowest frequency detectable
maxFrequency - Highest frequency detectable
sensitivity
ratio

Dynamic Wavelet

no special config

Note

If you'd like a version that uses compiled C++ code and runs much faster, check out this repo. However, it will not work in the browser.

Todo

Integrate with teoria or another music theory tool to add more intelligent parsing.
Note-onset algorithms.
Enable requiring of single detectors.

Thanks

Several of these algorithms were ported from Jonas Six's excellent TarsosDSP library (written in Java). If you're looking for a far deeper set of tools than this, check out his work on his website or on Github.

Thanks to Aubio for his YIN code

peterkhayes / pitchfinder

readme