Audio File Splitter

Vinyl & Tape Audio File Splitter is a single-file browser tool for:

• splitting recordings into individual tracks,
• tagging them with metadata, and
• exporting them as WAV files or FFmpeg scripts.

No installation. No uploads. Runs in the browser. Optional song recognition requires third-party connection.

1. Add an audio file.

(MP3, WAV, FLAC, OGG, M4A). Large VBR MP3s handled via chunked decoding.

2. Waveform viewer and split editor.

• A zoomable view of the audio file.
• Splits can be added, deleted, and moved.
• Audio can be played/stopped from any location with right-click.
• Optional view as spectrogram.

3. Detection.

Automated detection using a simple process.

Looks for "silence" of a minimum length below a certain volumn threshold. Default theshholds based upon album recording type.

Automated detection using a more complex process.

Additional analysis for detecting tracks where typical silence is less reliable. See below for more details.

Apply splits from a manual list.

Paste a list of tracks and times using one of several formats.

Apply splits based upon information of individual song tracks.

Search for an album and apply metadata and split information from MusicBrainz database or Discogs database.
Results are often off by a few seconds. Users can choose to "click-to" the promising detection and manually adjust each split in the waveform.

4. Review and tagging of identified tracts.

• Ability to look-up albums using MusicBrainz or Discogs.
• Ability to manually add track information.
• Ability to play each tract.
• Ability to review each proposed split location and confirm.

All metadata search uses the MusicBrainz public API (CC0 licence). No key required. Requests identify themselves as VinylSplitter/1.0 per MusicBrainz's usage guidelines.

5. Identify unknown tracks and add metadata.

Using your own API key, sending an audio sample to Shazam via RapidAPI, AcoustID, or AudD Add BPM metadata.

6. Export

Formats

• FFmpeg Script .sh (macOS/Linux): Place in the same folder as your audio file, run chmod +x script.sh && ./script.sh. Produces fully-tagged MP3s. Get FFmpeg ↗
• FFmpeg Script .bat (Windows): Place in the same folder as your audio file, then double-click or run from a command prompt. Requires FFmpeg on your PATH.
• WAV (Lossless): Exports each track as a WAV file. Use Mp3tag or Kid3 to add ID3 tags afterwards.
• Cue Sheet: A single .cue file describing all split points — use with your audio player or ripper.
• Metadata JSON: A structured JSON file with all track metadata for use in custom workflows.

Opening page:

Dark Mode with splits detected:

Two-columns optional:

Automated detection using a more complex process.

A more detailed description.

After identifying cadidates, applies audio analysis signals. Takes longer but may handle drum solos, fades, and ambiguous silences more reliably than simple threshold detection.

• Silence / Energy — Detects track boundaries by finding quiet regions where the audio amplitude drops below a fixed threshold. Gaps with deeper silence and lengths matching the expected inter-track gap score higher; a built-in beat guard suppresses false positives caused by drum solos or sparse sections.
• Spectral Flux — Measures how much the overall frequency content of the audio changes across a candidate boundary. A large spectral shift between the seconds before and after a gap is strong evidence that two different songs surround it.
• High-Freq Drop — Tracks the level of high-frequency content on either side of a gap. Track endings typically roll off treble content as a song fades, while new tracks open with a brighter sound.
• BPM / Rhythm — Estimates tempo using autocorrelation over 10-second windows and compares the dominant beat before and after a gap. A significant tempo change across the boundary raises the boundary score.
• Hi-Pass BPM — Works like BPM detection but first filters the audio above ~1 kHz to focus on snare and hi-hat transients, which often give a cleaner and more stable rhythm estimate than the full-band signal.
• Onset Density — Counts the rate of musical note events (onsets) in the audio and looks for a drop or surge across a candidate boundary. A song ending produces fewer and fewer onsets; a new song starting brings a burst of new activity.
• Spectral Centroid — Measures the average "brightness" of the audio on each side of a gap. A significant shift in the centre of spectral mass suggests a change in instrumentation or production style between two songs.
• Key / Chroma — Analyses the 12-note pitch-class distribution (chroma) of the audio on either side of a boundary. A change in tonal centre or key is a reliable indicator that a different song has begun.
• Dynamic Range — Measures the crest factor (peak-to-RMS ratio) of the audio. A fade-out at the end of a track compresses dynamics before the gap; a new track typically opens with greater dynamic contrast.