4 ms·
Thanks! Very interesting approaches, especially AlgoMotion's. Hudson (author of the first link) doesn't really tell what algorithm exactly do they use.
by vova_hn2 7mo ago
Thanks!
Very interesting approaches, especially AlgoMotion's. Hudson (author of the first link) doesn't really tell what algorithm exactly do they use.
- vunderba 7mo agoYeah I think there's a lot of room for experimentation in this area like locking the grid to musical modes (mixylodian, phyrgian), setting up intervals, etc.
- vova_hn2 7mo agoMyy experiment kinda supports every mode. For example, you can pick C major scale and it will have exactly the same notes as G Mixolydian or you can pick E Phrygian. Although, due to the chaotic nature of the simulation and also because the algorithm to pick notes that I chose, there is no tonal center, so strictly speaking we can't determine mode.
- vunderba 7mo agoyeah. One thing I've done in the past with experiments playing around with genetic algorithm derived music is that I would setup a recurrent tonic that would play every X beats to musically anchor to a signature and/or mode.