Add to Chrome
Add to Firefox
Add to Edge
Share this with someone who'll enjoy it:
Use a parametric representation of audio to train a generative model in the interest of obtaining more flexible control over the generated sound.