This is a configurable tool that generates music written in the twelve-tone technique. In short, this means that the output pieces are based on different principles than the vast majority of classical and popular music. If you have never listened to the twelve-tone works, you may find them not ear-pleasant, but, given enough experience, you can change your mind. Twelve-tone music is logic-oriented and has many symmetries, so it is a good choice for combinatorial optimization.
Each run of the tool results in creation of a fragment which can be used as a part of a larger composition. These fragments are stored in separate directories containing:
- MIDI file
- WAV file
- Events file in sinethesizer TSV format
- PDF file with sheet music and its Lilypond source
- YAML file that can be copied to a config for a derivative fragment
- Text file with meta information (such as path to config and evaluation scores)
With this tool, I released a 26-minutes album on Spotify and many other streaming platforms. However, main melodies for the album were written manually by myself and a lot of hard work was done with configs, but, nevertheless, the package really helped to generate background melodies. To read more about the album, please visit its official page.
To install a stable version, run:
pip install dodecaphonyTo create a new musical fragment, run:
python -m dodecaphony [-c path_to_config]Default config is used if -c argument is not passed. To create your own config, please look at this example with detailed explanations. Also, you can read the source code — it is structured and has built-in documentation.
If you are on macOS, please make sure that parallelism is enabled.