AudioCleanUpTool (Crisp)

Free and open source. Use it, modify it, share it — no restrictions. MIT licensed.

A self-contained desktop audio cleanup tool for voice recordings, podcasts, livestreams, and anything that needs to sound better. Built with Python + PySide6, ships as a single .exe — no Python installation required.

---

Download

Grab the latest installer from Releases → AudioCleanUpTool-x.x.x-Setup.exe. Run it, done.

---

Features

• Auto Clean — one button. Analyses your audio and applies the right settings automatically.
• 10 audio processors, each independently toggleable:
  • Noise Reduction (hum, hiss, fans, AC)
  • Dereverb (room tail / reverb reduction)
  • Plosive Removal (breath pops / p-pops)
  • EQ Balance (rumble cut, de-box, air lift)
  • Voice Clarity (presence lift + gentle leveling)
  • Loudness Normalise (ITU-R BS.1770 / LUFS)
  • Trim (manual gain)
  • Chorus (LFO-modulated depth effect)
  • Widener (M/S stereo widening)
  • Panner (constant-power stereo positioning)
• A/B comparison — toggle between original and cleaned in real time
• Live waveform display — before and after, x-linked
• Recording from any input device with a live level meter
• Batch processing of entire folders
• Export formats: WAV (16/24-bit, 32-bit float), FLAC, MP3, AAC, OGG, OPUS, AIFF
• Export presets: YouTube, Spotify, Podcast, Apple Podcasts, Broadcast (EBU R128), Broadcast TV (ATSC A/85), Church/Livestream, Voice-Over/ACX, Archival
• 7 themes: Dark Default, Midnight Blue, Charcoal Orange, Deep Purple, Hacker Green, Light Clean, Warm Cream
• Save/load cleanup settings as .crisp.json files
• Settings dialog (Ctrl+,) for playback volume and theme preferences

---

User Guide

See docs/AudioCleanUpTool_User_Guide.pdf for full documentation.

---

Project layout

src/crisp/
  core/                 # no-GUI audio engine (unit-testable, scriptable)
    audio.py            # AudioClip value object (float32, frames×channels)
    audio_io.py         # load/save via soundfile + pydub
    recorder.py         # sounddevice input + device enumeration
    dsp.py              # shared biquad filters / dynamics
    engine.py           # CleanupSettings + pipeline runner
    analyser.py         # AutoAnalyser + AnalysisReport
    presets.py          # formats + delivery presets
    exporter.py         # format/bit-depth/bitrate encoding
    batch.py            # folder processing
    ffmpeg.py           # locates the bundled ffmpeg
    processors/         # one module per cleanup stage
  gui/                  # PySide6 UI (main_window, waveform, player, workers)
installer/              # Inno Setup build script
tests/                  # 173 tests, all passing

The core package never imports Qt — fully testable and scriptable headlessly.

---

Develop

python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -e ".[dev]"
python -m crisp          # launch the GUI
pytest                   # run the test suite

ffmpeg ships with imageio-ffmpeg — no system install needed.

Build the executable

pip install -e ".[dev]"
pyinstaller crisp.spec
# -> dist/Crisp.exe

# Then package as installer:
& "C:\Program Files (x86)\Inno Setup 6\ISCC.exe" installer\crisp.iss
# -> dist/AudioCleanUpTool-x.x.x-Setup.exe

---

Licence

MIT — free for personal and commercial use, modification, and redistribution.