Two programs used as part of an installation built by Cat Toren and Andrew Yoon which was up on May 10, 2016 at Purchase College.

A microphone was set up in front of a printed booklet of algorithmically built text. Viewers of the installation were encouraged to read from the text into the microphone. A series of digital sine wave oscillators spontaneously played tones while audio captured from the microphone was remembered and recalled over time.

The seed text which formed the basis of the algorithmically built text was written by Cat Toren while Andrew Yoon wrote the programs which transformed the text and which processed and generated the sound.

You can listen to a recording of the installation online here, and you can view an example of the algorithmically generated text [here] (example/example_book.pdf).

Cat Toren is a jazz pianist, improviser, and composer; to hear her other works or get in touch, you can find her online here.

Running the programs in this project requires Python 3, if you do not have it installed you can do so here.

To get started, download this repository onto your local machine. Use your system's command line to navigate to the project location and install its dependencies using pip.

cd path/to/yoon-toren-installation
pip install -r requirements.txt

To generate the algorithmic poem booklet, run python build_book.py. This will generate a PDF document, place it in bookmaker/output, and open the file.

To run the audio processing program, run python run_electronics.py.

Depending on the audio environment of your system, you may need to modify the I/O device indices used by the program. You can temporarily set these with the -i your_input_device_index and -o your_output_device_index arguments. To change the default indices used, open electronics/config.py and edit DEFAULT_INPUT_DEVICE_INDEX and DEFAULT_OUTPUT_DEVICE_INDEX.

To view a list of available devices and their indices, run the program with the debug flag -d or --debug ie. python run_electronics.py --debug.

This project was originally hacked together in May 2016, and contained lots of poor decisions. Extensive work is currently underway refactoring and rewriting the code, but in the meantime please forgive any bugs or ugliness.

The run_electronics.py script has only been tested so far on Ubuntu using PortAudio. Modifications may need to be made to get it working on your system. See "Troubleshooting" above for more details.

• Continue improving docstrings and code legibility
• Continue refactoring to fix the various quick hacks used throughout
• Consider possibly splitting the two sub-projects into separate repos
• Better explanation for how to configure the audio environment
• Test audio configurations on other OS's and audio drivers.
• Implement automatic level adjustments to prevent clipping when the output becomes too loud.