1. Getting started

To install Djalgo, run

pip install git+https://github.com/essicolo/djalgo.git

Djalgo comes battery excluded. Although optional, the following packages will be very useful in your music composition workflow with Djalgo.

  • pip install jupyterlab to interact with your code in a notebook

  • Musecore and pip install music21 to render scores in the notebook

  • pip install pretty-midi to fine-tune and export your music to midi objects

  • pip install scamp to play music with sound fonts right from your notebook

  • Although not ready yet, Djalgo’s AI is based on Tensorflow, so you will need to pip install tensorflow to train an AI on your music

Djalgo is designed to generate musical pieces as generic Python objects.

  • A note is defined as a tuple of (midi pitch, duration time, time offset from the start). A rest is a note with a pitch defined as None. A rhythm is the same thing of a note, but without pitch, i.e. a (duration, offset) tuple.

  • A track is a list of notes.

  • A piece is a list of tracks.

Such objects can be converted to several music packages in Python, like Music21, Pretty-Midi, Mido and SCAMP.

To start with djalgo, install it then lauch your session by importing the package. The alias dj will be used through the documentation.

[1]:

import djalgo as dj

Djalgo offera a range of functionalities designed for music composers and enthusiasts. Here’s a snapshot of what Djalgo brings to the table:

  • Analysis: Located in analysis.py, discover a suite of indices for dissecting tracks—whether it’s pitches, durations, or offsets. These metrics serve not just for analysis but also as benchmarks for the evolutionary algorithms found in genetic.py.

  • Conversion: conversion.py is your gateway to integrating Djalgo with popular music packages. Transform notes and compositions into formats compatible with Music21 for notation, Pretty-Midi for MIDI refinements, and SCAMP for sound production. Installing these packages is a prerequisite for conversion.

  • Fractals: fractal.py delves into the beauty of mathematics, extracting music from the intricate patterns of cellular automata and Mandelbrot fractals.

  • Genetic Algorithms: Use genetic.py to evolve your music, steering it towards specific analytical targets defined in analysis.py.

  • Harmony: harmony.py equips you with tools to enrich compositions with scales, voicings, and ornamental touches.

  • Loop Visualization: loop.py helps visualize musical loops with radar plots, offering a new dimension to beat creation.

  • Minimalism: Explore minimalist techniques in minimalism.py, from additive and subtractive processes to Arvo Pärt’s tintinnabuli, and craft music with a minimalist ethos.

  • Rhythm: The Rhythm class in rhythm.py is designed for crafting and experimenting with complex rhythmic structures.

  • Utilities: utils.py provides essential tools for fine-tuning: repair, tune, and quantize your compositions to perfection.

  • Random and Kernel Walks: In walk.py, let music wander through algorithmic paths, guided by random and kernel-induced walks.

  • Machine learning: DJai is a work in progress, and does not behave as expected. djai.py is aimed at using machine learning to generate new music pieces, all powered by TensorFlow. DJai was designed to learn from any MIDI file, but if art your aim, you’d better create a machine that learns from your own compositions.

↳ To Harmonies