audipart 0.0.1

Extract and separate audio signals by pitch from `.mp3` or `.wav` files.

audio-pitch-separation

Extract and separate audio signals by pitch from .mp3 or .wav files. Harmonic, percussive, and note-wise components are visualized and saved as .wav files automatically.

Features

• Supports both mp3 and wav audio input
• Separates harmonic & percussive components using HPSS
• Extracts note-wise waveforms via Constant-Q Transform (CQT)
• Visualizes all audio tracks as waveform plots
• Saves each pitch as an individual .wav file
• Easy GUI file picker (via PyQt5)

Installation

git clone https://github.com/Hepheir/audio-pitch-separation.git
cd audio-pitch-separation
pip install -r requirements.txt

Requirements

• numpy
• matplotlib
• librosa
• PyQt5
• soundfile
• pydub

You can install all dependencies via:

pip install -r requirements.txt

Usage

python -m audipart

1. A GUI window will prompt you to select an .mp3 or .wav audio file.
2. From the selected audio, the following will be generated:
   • harmonic.wav and percussive.wav using HPSS
   • .wav files for each pitch from C2 to C8
   • A waveform plot saved as figure.png
3. All output files will be saved in a new folder named after the input file (excluding the extension).

you can also import audipart like this demo.

Example

For example, if you select song.mp3, the following directory will be created:

song/
├── harmonic.wav
├── percussive.wav
├── C2.wav
├── C#2.wav
├── ...
├── C8.wav
└── figure.png

TODO

• Extract note-wise MIDI representation

License

MIT License.

Acknowledgements

• librosa
• HPS