@audio/mir-melody

Melody contour — frame-level F0 track via YIN with voicing flags (MIREX melody-extraction shape).

npm install @audio/mir-melody

import melody from '@audio/mir-melody'

let { times, f0, voiced } = melody(data, { fs: 44100 })
// times: Float32Array (s, frame centers) · f0: Float32Array (Hz, 0 = unvoiced) · voiced: Uint8Array (0|1)

One frame per hop; f0[i] is only meaningful where voiced[i] === 1 (unvoiced frames are left at 0, not NaN).

Options: - fs — sample rate (default 44100, Hz) · frameSize — YIN analysis window, samples (default 2048) · hop — frame hop, samples (default 512) · threshold — passed through to @audio/pitch-yin's YIN threshold (default from @audio/pitch-yin)

Use when: tracking a single predominant pitch line (vocal/lead) over time; for chords or polyphony use @audio/mir-multif0.

Also exported as an audio.js stat manifest (./audio).

Part of @audio/mir.