Polymath: Revolutionizing Music Production with AI
Polymath is an innovative AI tool that transforms any music library into a music production sample-library. It offers a seamless and efficient workflow for music producers, DJs, and ML audio developers.
Key Features:
- Automatically separates songs into stems (beats, bass, etc.).
- Quantizes songs to the same tempo and beat-grid.
- Analyzes musical structure, key, and other important infos.
- Converts audio to MIDI.
Use Cases:
- Easily combine elements from different songs to create unique compositions.
- Quickly create a polished, hour-long mash-up DJ set.
- Simplify the process of creating a large music dataset for ML developers.
How It Works:
- Music Source Separation is performed with the Demucs neural network.
- Music Structure Segmentation/Labeling is performed with the sf_segmenter neural network.
- Music Pitch Tracking and Key Detection are performed with Crepe neural network.
- Music to MIDI transcription is performed with Basic Pitch neural network.
- Music Quantization and Alignment are performed with pyrubberband.
- Music Info retrieval and processing is performed with librosa.
Requirements and Installation:
- Requires ffmpeg and python version >=3.7 and <=3.10.
- To install, follow the steps: git clone , cd polymath, pip install -r requirements.txt. If you encounter an issue with basic-pitch, run pip install git+.
- Most libraries used come with native GPU support through cuda. Follow the steps on to setup tensorflow for use with cuda.
Docker Setup:
- If Docker is installed, use the provided Dockerfile to build a polymath docker image. Create the necessary directories for input and output file exchange.
Running Polymath:
- Add songs to the Polymath Library, including from YouTube or local audio files.
- Quantize songs in the library to a specific tempo or the tempo of the song.
- Search for similar songs in the library and optionally quantize them.
- Convert processed audio files and stems to MIDI.
Audio Features Extracted:
- Stems: bass, drum, guitar, other, piano, vocals.
- Features: tempo, duration, timbre, pitch, intensity, volume, loudness, beats, segments boundaries, segments labels, frequency.
Polymath is released under the MIT license and is a powerful tool for those working in the music and AI fields.