@audio/pitch-yin 1.0.0

package/audio-module.js

@@ -0,0 +1,31 @@
+// audio-module manifest — YIN as an analyzer: audio in, no audio out, pitch/clarity
+// emitted per analysis frame via ctx.emit.
+
+import yin from './yin.js'
+
+const FRAME = 2048
+
+export const pitch = (ctx) => {
+	const win = new Float32Array(FRAME)
+	let fill = 0
+	return (inputs, outputs, params) => {
+		const x = inputs[0] && inputs[0][0]
+		if (!x) return
+		for (let i = 0; i < x.length; i++) {
+			win[fill++] = x[i]
+			if (fill === FRAME) {
+				const r = yin(win, { fs: ctx.sampleRate, threshold: params.threshold[0] })
+				if (r) {
+					ctx.emit('pitch', r.freq)
+					ctx.emit('clarity', r.clarity)
+				}
+				fill = 0
+			}
+		}
+	}
+}
+pitch.channels = { inputs: 1, outputs: [] }
+pitch.events = { out: { pitch: 'number', clarity: 'number' } }
+pitch.params = {
+	threshold: { type: 'number', min: 0.05, max: 0.5, default: 0.1 },
+}

package/package.json

@@ -0,0 +1,43 @@
+{
+  "name": "@audio/pitch-yin",
+  "version": "1.0.0",
+  "description": "YIN pitch detection (de Cheveign\u00e9 & Kawahara, 2002)",
+  "type": "module",
+  "sideEffects": false,
+  "main": "yin.js",
+  "exports": {
+    ".": "./yin.js",
+    "./package.json": "./package.json",
+    "./audio-module": "./audio-module.js"
+  },
+  "files": [
+    "yin.js",
+    "audio-module.js"
+  ],
+  "keywords": [
+    "audio",
+    "dsp",
+    "pitch",
+    "pitch-detection",
+    "f0",
+    "yin"
+  ],
+  "license": "MIT",
+  "author": "Dmitry Iv. <dfcreative@gmail.com>",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/audiojs/pitch.git",
+    "directory": "packages/pitch-yin"
+  },
+  "homepage": "https://github.com/audiojs/pitch/tree/main/packages/pitch-yin",
+  "bugs": {
+    "url": "https://github.com/audiojs/pitch/issues"
+  },
+  "engines": {
+    "node": ">=18"
+  },
+  "publishConfig": {
+    "access": "public"
+  },
+  "audio-module": "./audio-module.js"
+}

package/yin.js

@@ -0,0 +1,54 @@
+/**
+ * YIN pitch detection (de Cheveigné & Kawahara, 2002).
+ * Difference fn → cumulative mean normalized difference → threshold → parabolic interpolation.
+ *
+ * @param {Float32Array | Float64Array} data - audio samples (single window, eg. 2048 samples)
+ * @param {{fs?: number, threshold?: number}} params
+ * @returns {{freq: number, clarity: number} | null}
+ */
+export default function yin(data, params) {
+  let fs = params?.fs || 44100
+  let threshold = params?.threshold ?? 0.15
+  let len = data.length
+  let half = len >> 1
+
+  // step 1-2: difference function
+  let d = new Float64Array(half)
+  for (let tau = 1; tau < half; tau++) {
+    let sum = 0
+    for (let i = 0; i < half; i++) {
+      let diff = data[i] - data[i + tau]
+      sum += diff * diff
+    }
+    d[tau] = sum
+  }
+
+  // step 3: cumulative mean normalized difference
+  let cmndf = new Float64Array(half)
+  cmndf[0] = 1
+  let running = 0
+  for (let tau = 1; tau < half; tau++) {
+    running += d[tau]
+    cmndf[tau] = running > 0 ? d[tau] * tau / running : 1
+  }
+
+  // step 4: absolute threshold — find first dip below threshold
+  let tau = 2
+  while (tau < half - 1) {
+    if (cmndf[tau] < threshold) {
+      // find local minimum
+      while (tau + 1 < half - 1 && cmndf[tau + 1] < cmndf[tau]) tau++
+      break
+    }
+    tau++
+  }
+  if (tau >= half - 1) return null
+
+  // step 5: parabolic interpolation
+  let s0 = cmndf[tau - 1], s1 = cmndf[tau], s2 = cmndf[tau + 1]
+  let denom = s0 - 2 * s1 + s2
+  let shift = denom !== 0 ? (s0 - s2) / (2 * denom) : 0
+  let period = tau + shift
+
+  return { freq: fs / period, clarity: 1 - s1 }
+}