Hi Doug, I have some code that you could download that attempts (rather unsuccessfully) to use neural networks to compose new chant-like melodies in the style of those it is trained with. It is mostly pure Common Lisp but there is a CM function for translating note lists to Midi. (Would be easy to output something else like Csound though). You can obtain the code and a paper that I wrote explaining it here:
http://briefcase.yahoo.com/bc/[EMAIL PROTECTED]/lst?.dir=/NNChanter or try just http://briefcase.yahoo.com/bc/[EMAIL PROTECTED]/ and go to the NNChanter folder if the first link does not work for you. Even though it does not work well, it might be an interesting stepping off point. I also have some experimental code that builds on the NNChanter project and uses markov chains instead of neural networks. I think that this gives better results for what I was trying to accomplish. In addition, I have code for several Csound projects using CM to algorithmically generate scores. Most of this code is my private compositional material and is not online to download as yet, but I would be happy to share it directly with you if you understand that it represents my "works in progress" and therefore comes with little or no documentation. Anthony Kozar anthonykozar AT sbcglobal DOT net Doug Geers wrote on 3/11/06 4:56 PM: > Hi. I've been lurking for awhile but have only recently begun > exploring CM. I'd love to see some example code beyond those on the > website, especially functions that export to Csound, CMN, > Supercollider, Music XML, and similar--Basically, anything beyond MIDI > files. Also especially interesting would be further examples of > algorithmic generation using Markov chains, neural nets, genetic > algorithms, etc, and interfaces to OpenMusic and/or Audiosculpt. _______________________________________________ Cmdist mailing list [email protected] http://ccrma-mail.stanford.edu/mailman/listinfo/cmdist
