zigify-msaf

Music structural segmentation for the Zigify pipeline. A thin CLI wrapper around MSAF that pins to the olda boundary detector and scluster labeler — the combination that scored best in evaluation against hand-annotated ground truth.

Install

uvx zigify-msaf <audio>          # ephemeral, recommended
uv pip install zigify-msaf       # into a project

uvx resolves and caches a dedicated environment on first run; subsequent calls cold-start in ~100 ms.

Use

zigify-msaf path/to/track.mp3
zigify-msaf track.mp3 --out track.segments.json
zigify-msaf track.mp3 --feature mfcc --verbose
zigify-msaf track.mp3 --bpm 118        # skip detection, lock all segments to a known tempo
zigify-msaf track.mp3 --viz             # write track.png + track.html viewer next to the mp3
zigify-msaf track.mp3 --min-segment 5  # absorb segments shorter than 5s (default: 3.0; 0 disables)

stdout is newline-delimited JSON: progress events first, then a single final result line. stderr carries human-readable logs from msaf/librosa (silenced by default; pass --verbose to surface them).

Output schema

Each line on stdout is one JSON object. The shape of the final result line:

{
  "type": "result",
  "source": "path/to/track.mp3",
  "duration": 357.98,
  "tempo": 117.45,
  "tempoPrior": 117.45,
  "bpmOverride": null,
  "beatCount": 700,
  "accentCount": 88,
  "feature": "pcp",
  "boundaryAlgo": "olda",
  "labelAlgo": "scluster",
  "nSegments": 12,
  "nClusters": 5,
  "loudness": -18.4,
  "peakLoudness": -3.1,
  "segments": [
    {
      "start": 0.0,
      "end": 18.1,
      "duration": 18.1,
      "cluster": "S4",
      "bpm": 117.62,
      "beats": [0.51, 1.02, 1.53, 2.04],
      "beatCount": 36,
      "accents": [0.51, 4.6, 9.2, 13.8],
      "accentCount": 4,
      "topAccent": 9.2,
      "onsetCount": 22,
      "onsetRate": 1.215,
      "loudness": -23.7,
      "peakLoudness": -8.9,
      "dynamicRange": 14.8,
      "energy": 0.21,
      "brightness": 1840.5
    }
  ],
  "elapsed": 12.3
}

Per-segment fields beyond start/end/cluster describe musical character useful for downstream light-show or visualization generation:

Earlier lines look like:

{"type":"stage","name":"loading","message":"reading track.mp3"}
{"type":"stage","name":"onsets","message":"computing onset strength envelope"}
{"type":"stage","name":"accents","message":"detecting onsets and accents"}
{"type":"stage","name":"energy","message":"computing loudness and brightness"}
{"type":"stage","name":"features","message":"extracting pcp"}
{"type":"stage","name":"boundaries","message":"olda"}
{"type":"stage","name":"labels","message":"scluster"}
{"type":"stage","name":"tempo","message":"estimating global bpm prior"}
{"type":"stage","name":"beats","message":"tracking beats per segment"}

On failure the tool emits a single {"type": "error", ...} line and exits non-zero.

Segment cleanup

MSAF (olda boundaries) often emits a sub-second leading segment and occasionally a tiny trailing or mid-track sliver. The CLI absorbs any segment shorter than --min-segment (default 3.0s) into a neighbor: leading slivers merge into the next segment, trailing into the previous, and mid-track ones into the same-cluster neighbor (or the longer of the two if neither matches). Pass --min-segment 0 to keep every raw boundary.

Beat-sheet visualization

Pass --viz to render a horizontal beat sheet alongside the JSON output: waveform with RMS overlay, segment bands colored by cluster (with per-segment BPM), beat ticks (dark) and accent ticks (red). Useful for sanity-checking that detected beats actually line up with the audio.

Two files are written next to the audio file, sharing its basename:

• track.png — the beat sheet itself.
• track.html — a self-contained HTML viewer with a native <audio> element, the PNG below it, and a vertical playhead cursor that tracks currentTime. Click anywhere on the strip to seek. Open it directly in a browser.

Requires the viz extra:

uv pip install "zigify-msaf[viz]"
uvx --with matplotlib zigify-msaf track.mp3 --viz

Calling from Node / TypeScript

import { spawn } from 'node:child_process'
import { createInterface } from 'node:readline'

const proc = spawn('uvx', ['zigify-msaf', audioPath], { stdio: ['ignore', 'pipe', 'pipe'] })
const rl = createInterface({ input: proc.stdout })

let result: SegmentResult | undefined
for await (const line of rl) {
  const evt = JSON.parse(line)
  if (evt.type === 'stage') console.log(`[${evt.name}] ${evt.message ?? ''}`)
  if (evt.type === 'result') result = evt
  if (evt.type === 'error') throw new Error(evt.message)
}

Algorithm choice

Evaluated against a 14-boundary ground truth on Michael Jackson — Thriller (tolerance ±5 s):

olda wins precision and ties for recall. Scluster labels group the segments into ~5 clusters that align with verse/chorus/outro structure on test tracks.

License

MIT — see LICENSE.