Animal Kingdom: new experiments of computer aided composition

In the past weeks, I’ve already written about some funny experiments that i made first with music and machine learning, then with music composed using pseudo-random algorithms.

In both cases, results are some interesting, but strange, melodies that I finished with simple EDM arrangements.

For the third iteration of my project, I restarted the training of my musical AI (based on Magenta [1]), using along with old samples [2], also the composition created with pySchö (pseudo-random melodies) [3].

This time, generated melodies are less dissonant, probabily due the melodies generated with pySchö [5], that are strictly diatonic.
As in previous experiments, I arranged the generated melodies with sinthetic sound, with a very “ambient” flavour: a very good way to stay in training with musical theory and arranging!

The EP is called “Animal Kingdom“, you can listen it on major music streaming services, including Spotify [4]:


References

  1. Magenta
  2. Spotify – A.I. (Art-typical Intelligence)
  3. Spotify – Elements
  4. Spotify – Animal Kingdom
  5. https://github.com/andreafortuna/pyscho

Related posts

  1. “Change Of Season”: four songs, entirely written by human
  2. pySchö: algorithmic music composition
  3. “Elements”: an experiment on algorithmic composition
  4. APT: Advanced Persistent Theme
  5. My experiments with music and machine learning: A.I. (Art-typical Intelligence)