DarwinTunes.org

http://www.sciencedaily.com/releases/2012/06/120618153716.htm

http://darwintunes.org/pnas-paper

http://www.pnas.org/content/early/2012/06/12/1203182109.full.pdf+html

Tunes evolve by "genetic algorithm" where "fitness" is determined
by human listener-ratings on the internet. It certainly appears to have worked to
produce interesting music different from any other I've heard. I would not say
it is superior to Mozart, etc, but better than plenty of professional music.

--- In tuning-math@yahoogroups.com, "WarrenS" <warren.wds@...> wrote:
>
> http://www.sciencedaily.com/releases/2012/06/120618153716.htm
>
> http://darwintunes.org/pnas-paper
>
> http://www.pnas.org/content/early/2012/06/12/1203182109.full.pdf+html
>
> Tunes evolve by "genetic algorithm" where "fitness" is determined
> by human listener-ratings on the internet. It certainly appears to have worked to
> produce interesting music different from any other I've heard. I would not say
> it is superior to Mozart, etc, but better than plenty of professional music.

participate in their new experiment here:
http://darwintunes.org/evolve-music-with-beats

listen to their 3000-generation music from their old experiment here:
http://darwintunes.org/3000-generations

--- In tuning-math@yahoogroups.com, "WarrenS" <warren.wds@...> wrote:

> Tunes evolve by "genetic algorithm" where "fitness" is determined
> by human listener-ratings on the internet. It certainly appears to have worked to
> produce interesting music different from any other I've heard. I would not say
> it is superior to Mozart, etc, but better than plenty of professional music.

I want to be able to fit xenharmony into it. The easiest way would be to fix a scale and stick with it, rather than trying to mutate that too, but you'd probably need to restrict the voters to the xenharmonically sophisticated.

Warren: Perhaps a more sophisticated genetic algorithm could logically discourage changing the scale by giving the voter something else to change. And perhaps, the algorithm should test how the new tuning affects the overall "goodness" and weight the change appropriately (although the whole point of changing something may be to create a new model of "goodness"....)