/// Track wash station between your download folder and the next-gig CDJ

Find thedead tracksbefore thebooth does.

Name: DigWash
Author: DigWash

Drop in a fresh digging folder. DigWash reads every file, explains what will not load and what may sound rough, then previews booth-ready copies and a plain playlist you can import into rekordbox — originals never touched.

Get the free beta → Read the guide →

macOS app + cross-platform CLI
fully offline
free during the beta · pay what you want

DigWash does not make bad tracks sound good. It makes risky files visible before the gig, and prepares safer copies when the issue is compatibility.

§ How you use it

Three steps between dig and gig.

01
Drop your digging folder
Bandcamp, SoundCloud downloads, promos, old rips.
02
See what's risky
Won't load / too hot / suspicious encode / weird format.
03
Export a clean gig folder
Originals untouched. Clean copies + a plain playlist you can import into rekordbox.

Where it sits

Download folderfresh dig, promos, rips
DigWashcheck · wash
rekordboxyour grids, cues, analysis — DigWash never touches them
USBexport as usual

§ See it work · one truthful scenario

The same facts, from one track to the whole dig.

Eight tracks in: two need conversion for older decks, two carry real clipping and get flagged to listen first, the other four ride as-is. Out: a clean folder + playlist, originals untouched.

Pick what you want to check and how you like to work — every panel below is the shipped tool looking at the same eight-track dig folder.

Real output, not a mockup: eight real tracks under fixture names, run through the shipped CLI and app. Terminal captures are verbatim, colors included; every number in the app views comes from the same runs.

One suspicious download: a 24-bit / 96 kHz FLAC. Clean source — and three USB deck families still can't load it.

digwash check 02_hires.flac

$ digwash check 02_hires.flac
Checking 1 track(s) — level, true-peak, clipping, channel balance, spectral roll-off, fake-lossless
  tip: add --no-spectral to skip the slowest (frequency) analysis

[DW≋] DigWash check — read-only

→ 02_hires.flac

Verdict
  Booth: 3 deck families won't play
  Source: lossless
  Loudness: ok

Do
  Convert for compatibility
  Conversion can make it compatible, but it cannot restore missing detail.

File
  flac in flac · 96 kHz · stereo · 2.5 Mbps · 7:01.9

Audio
  loudness -11.9 LUFS · true peak +0.0 dBTP · LRA 0.8 LU
  sample peak -0.7 dBFS · hard clipping evidence: none

No files changed — analysis is read-only.
Next: `digwash convert <track>` or `digwash wash <folder>`
Add --verbose for the full spectral, dynamics, and provenance readout.

digwash check ~/digs/2026-06

$ digwash check ~/digs/2026-06
Checking 8 track(s) — level, true-peak, clipping, channel balance, spectral roll-off, fake-lossless
  tip: add --no-spectral to skip the slowest (frequency) analysis

  1. 01_ready.aiff
  2. 02_hires.flac
  3. 03_old-flac.flac
  4. 04_honest.mp3
  5. 05_vinyl.wav
  6. 06_quiet.aiff
  7. 07_balanced.wav
  8. 08_live.mp3

[DW≋] DigWash check — read-only

Checked: 8

Action  ⚠ listen 2 · → convert 2 · ✓ keep 4
Main reasons  2 sustained clipping

⚠ Listen-check
  05_vinyl.wav
    samples are pinned at digital full scale for sustained runs; clipping distortion is likely audible when played loud
  08_live.mp3
    samples are pinned at digital full scale for sustained runs; clipping distortion is likely audible when played loud

→ Convert for compatibility
  02_hires.flac
    levels and spectrum look usable
  03_old-flac.flac
    levels and spectrum look usable

✓ Keep
  4 track(s) look usable

No files changed — analysis is read-only.
Run `digwash check --verbose <playlist>` for per-track levels, spectra, and vectorscopes.

digwash convert · digwash crate

$ digwash convert 02_hires.flac
Convert preview — any-booth preset
  1 to convert, 0 to skip
  clean copies land next to the originals

  → 02_hires.flac → 02_hires.aiff
      ⚠ true peak +0.0 dBTP — this copy preserves an existing clipping risk — reported, not repaired

  Level policy: no gain, no limiter — decoded samples above full scale saturate at 0 dBFS in the PCM copy; the risk is reported, not repaired.
  Preview only — no files were written.
  Use --out-dir <folder> to choose where the copies land.

Nothing was written. Re-run with -y (or --commit) to convert.

$ digwash crate ~/digs/2026-06 --database ~/crate
Planning 8 track(s) — reading levels, clipping, channel balance, spectral roll-off, fake-lossless…
  tip: add --no-spectral to skip the slowest (frequency) analysis

Crate preview — any-booth preset
  source:    2026-06
  database:  ~/crate-db
  playlists: ~/crate-db

  8 track(s): 6 to copy, 2 to convert
  0 duplicate(s), 0 renamed, 0 already in database
  quality: 6 clean, 2 hot
  do:      2 convert · 2 listen · 4 keep

  files:
    ✓ 01_ready.aiff  clean
       action: copy -> 01_ready.aiff; .aiff kept as-is (any-booth); current preset keeps this source
    ✓ 02_hires.flac  clean
       action: convert -> 02_hires.aiff; .flac -> AIFF (any-booth)
    ✓ 03_old-flac.flac  clean
       action: convert -> 03_old-flac.aiff; .flac -> AIFF (any-booth)
    ✓ 04_honest.mp3  clean
       action: copy -> 04_honest.mp3; .mp3 kept as-is (any-booth); current preset keeps this source
    ⚠ 05_vinyl.wav  hot
       action: copy -> 05_vinyl.wav; .wav kept as-is (any-booth); current preset keeps this source
       issue: hard clipping detected (0.037% of samples at full scale); check the source before playing loud  (explain: hot)
    ✓ 06_quiet.aiff  clean
       action: copy -> 06_quiet.aiff; .aiff kept as-is (any-booth); current preset keeps this source
    ✓ 07_balanced.wav  clean
       action: copy -> 07_balanced.wav; .wav kept as-is (any-booth); current preset keeps this source
    ⚠ 08_live.mp3  hot
       action: copy -> 08_live.mp3; .mp3 kept as-is (any-booth); current preset keeps this source
       issue: hard clipping detected (0.049% of samples at full scale); check the source before playing loud  (explain: hot)

  quality warnings:
    ⚠ 05_vinyl.wav: hard clipping detected (0.037% of samples at full scale); check the source before playing loud  (explain: hot)
    ⚠ 08_live.mp3: hard clipping detected (0.049% of samples at full scale); check the source before playing loud  (explain: hot)

  playlists:
    2026-06.m3u (8 track(s))

  Preset note: conversion is preset-driven. The default `any-booth` makes the broadest booth-safe copies; run `digwash presets` to compare rules.
  Quality note: warnings are advisory; transcoding does not repair clipping or low-quality sources.
  Preview only — no files were changed. Re-run with -y (or --commit) to build the crate.

Nothing was written. Re-run with -y (or --commit) to build the crate.

digwash check --verbose 02_hires.flac

$ digwash check --verbose 02_hires.flac
Checking 1 track(s) — level, true-peak, clipping, channel balance, spectral roll-off, fake-lossless
  tip: add --no-spectral to skip the slowest (frequency) analysis

[DW≋] DigWash check — read-only
→ 02_hires.flac  clean — 3 deck families won't play · lossless · loud: ok
  do: Convert for compatibility
  call: levels and spectrum look usable
  ran: rig, dynamics, spatial, codec, salience, spectral — cache hit

  container/playability:
    flac in flac; 96 kHz; 24-bit; stereo; 2.5 Mbps; duration 7:01.9; declared lossless yes

  booth compatibility:
    plays on   : Modern FLAC/ALAC up to 96 kHz (CDJ-3000, CDJ-2000NXS2, OPUS-QUAD, XDJ-AZ)
    WON'T play : Pre-FLAC USB CDJ/XDJ (MP3/AAC/WAV/AIFF only) (CDJ-2000NXS, CDJ-900NXS, CDJ-850, XDJ-700, …+4) — FLAC not supported
    WON'T play : Modern FLAC/ALAC up to 48 kHz (XDJ-1000MK2, OMNIS-DUO) — sample rate too high (96.0 kHz > 48 kHz)
    WON'T play : FLAC (no ALAC) up to 48 kHz (XDJ-XZ, XDJ-RX3) — sample rate too high (96.0 kHz > 48 kHz)
    wash fix   : `digwash convert 02_hires.flac` → 24-bit/48 kHz AIFF copy — plays on every USB deck family above

  ── Provenance and integrity ──

  codec/integrity:
    codec flac; real bitrate 2.5 Mbps; declared lossless yes

  ── Rig and dynamics safety ──

  levels/dynamics:
    peak -0.7 dBFS; true peak +0.0 dBTP; RMS -13.2 dBFS
    true-peak over: yes (+0.0 dBTP); hard clipping: none — flat-top 0.0%
    loudness -11.9 LUFS; LRA 0.8 LU; PLR 12.0 LU; crest 12.5 dB; flat-top 0.0%
    envelope full track 7:01.9  [▇▇▅██▇██▇█▇▅███▇██▇▁]  loudest 2:29.0

  ── Spatial translation and spectral fatigue ──

  spectrum/transcode:
    rolloff 20.9 kHz; transcode guess no; bandwidth score 0.0; HF sparsity 0.0
    sharpness 0.4; roughness 0.3; tilt -3.9 dB/oct
    spectrum    31 ▆▇████▇▇▆▆▅▅▅▆▆▆▆▆▆▆▆▆▆▅ 20k

  stereo/spatial:
    stereo  balance -0.0 dB  corr 1.0
    side/mid -27.6 dB; mono-sum loss -0.0 dB; LF mono loss -0.0 dB; negative correlation 0.0%

    stereo field (L·R vectorscope):
              ▒▓▒
              ▒▓▒░
              ▒▓▒░
              ▒▓▒
              ▓█▓
              ▒▓▒
              ▒▓▒
              ▒▓▒
              ▒▓▒
      mono = vertical spine · wide = filled · out-of-phase = horizontal

  warnings:
    none strong enough to flag

No files changed — analysis is read-only.
Next: `digwash convert <track>` for one-file compatibility, or `digwash wash <folder>` for a dig folder.
Use `digwash explain <term>` for peak, RMS, roll-off, transcode, or channel balance.

§ What every file goes through

Reads · tests · calls · acts.

01
Reads
Codec, container, bitrate, sample rate, bit depth and channels — the honest file facts.
02
Tests
Level, clipping, dynamics, stereo and codec clues, without pretending every warning can be repaired.
03
Calls
A plain verdict, the reason, and which USB deck families play the file as-is.
04
Acts
Copy, convert, warn or skip — only after you review the plan and choose Run wash.

Safe by default✓ analysis and preview are read-only✓ never overwrites silently✓ originals kept

§ Why this exists

Built by a digger. Honest, not flattering.

Measured signal, cited sources

Real measurements and cited evidence, not an AI taste score. A bad source stays a bad source even when its copy becomes playable.

Explore the research and bibliography →

Free during the beta

The full offline wash is free during the beta, with no track-count game. Pay what you want — or nothing.

Yours, end to end

Runs offline. Plain audio files and playlists in, plain files out. Your music and metadata do not become a hosted library.

Back to the scene

After fixed costs, a published share is intended for associations defending free-party culture in France.

§ Where to next

/beta · start here

Find thedead tracksbefore thebooth does.

Three steps between dig and gig.

Drop your digging folder

See what's risky

Export a clean gig folder

The same facts, from one track to the whole dig.

Looks usable for a set

Provenance and integrity

Rig and dynamics safety

Spatial translation and spectral fatigue

2026-06

8 tracks → clean copies + one playlist

The instruments behind the verdict.

Reads · tests · calls · acts.

Reads

Tests

Calls

Acts

Built by a digger. Honest, not flattering.

Measured signal, cited sources

Free during the beta

Yours, end to end

Back to the scene

Get the beta

Wash your first track

Tell us how you dig

Find thedead tracksbefore thebooth does.

Don't let one bad file ruin the set.

Three steps between dig and gig.

Drop your digging folder

See what's risky

Export a clean gig folder

The same facts, from one track to the whole dig.

Looks usable for a set

Provenance and integrity

Rig and dynamics safety

Spatial translation and spectral fatigue

2026-06

8 tracks → clean copies + one playlist

The instruments behind the verdict.

Reads · tests · calls · acts.

Reads

Tests

Calls

Acts

Built by a digger. Honest, not flattering.

Measured signal, cited sources

Free during the beta

Yours, end to end

Back to the scene

Get the beta

Wash your first track

Tell us how you dig