I am sure you know this
It is very pedagogical:
http://www.hexahedria.com/2015/08/03/composing-music-with-recurrent-neural-networks/
1 Like
A system able to work on/with ou implicit knowledge.
Articulate sounds together both horizontally and vertically. Something like the old Diphone but generalized.
http://anasynth.ircam.fr/home/english/software/diphone-studio
Demix using NMF (Non-negative matrix factorization):
Sound decomposition using wavelets for instance: using Lastwave:
http://www.cmap.polytechnique.fr/~bacry/LastWave/
ECRIN European project at Ircam:
http://anasynth.ircam.fr/home/english/projects/écrins