News for hackers. Stuff that matters.
Re:Chord progressions of 25,000 songs by m_eiman on Friday April 17, 2015 @10:30AM (#9394482)
Sounds like something that could be used to generate endless but still decent elevator music with some Markov chains.
Re:Chord progressions of 25,000 songs by smrtinsert on Friday April 17, 2015 @11:07AM (#9394761)
It's funny, whenever I look up algorithmic music, it never seems to be the usual pop cadences. The Phd who make this their field seem to prefer atonal or dissonant generated music.
Re:Chord progressions of 25,000 songs by pierrec on Friday April 17, 2015 @11:25AM (#9394877)
The field of generative music is very under-researched, but there are some great things out there, including tonal and "harmonious" ones.

Try David Cope: https://www.youtube.com/watch?v=CgG1HipAayU

And xoxos' "BreathCube" and "New Blend": http://www.xoxos.net/software/software.html

Re:Chord progressions of 25,000 songs by mjn on Friday April 17, 2015 @01:38PM (#9395882)
I think of classical music as a more "canonical" target than atonal or dissonant music, though it depends on what community of researchers you're looking at. One of the more famous researchers is David Cope, who spent much of his time on algorithmically modeling composition in the style of J.S. Bach. There's also an intermittent stream of papers on jazz improvization, e.g.: http://scholar.google.com/scholar?q=jazz+improvisation+icmc

There is a good deal of avant-garde experimentation as well, in part because computer music has in many places/eras been larger in the field of music, rather than the field of computer science. Contemporary composers like Iannis Xenakis, Karlheinz Stockhausen, Pierre Boulez, etc., picked up and developed new technologies as they became available. But they were avant-garde composers interested in using these technologies as part of their compositional toolkit, not CS researchers interested in automating generation of music per se.

One driving question for a certain group of composers is along the lines of: how can this new technology/technique help us make new, different music? Which leads to different experiments than the question: how can this new technology/technique help us automate the production of existing music? Sometimes that still ends up popularly oriented, e.g. in the '60s and '70s there was quite a bit of crossover between avant-garde synth experimentation and popular music, with Brian Eno as one hub in particular, and Moog synths finding their way onto all sorts of top-charting albums. Other times it's ended up less popularly oriented, like most of the granular-synthesis music.

Re:Chord progressions of 25,000 songs by jader201 on Friday April 17, 2015 @11:22AM (#9394856)
I've actually considered using this approach to build dynamic video game music.
Re:Chord progressions of 25,000 songs by NickNameNick on Friday April 17, 2015 @07:35PM (#9397773)
Have you seen the 'infinite jukebox'?

http://labs.echonest.com/Uploader/index.html

It stretches or compresses existing music by jumping back and forth to parts of the piece that sound similar, which could be handy for dynamically adapting the timing/duration of existing pieces to the players situation.

Re:Chord progressions of 25,000 songs by itsjareds on Friday April 17, 2015 @03:12PM (#9396535)
That is an awesome idea. I can imagine it would be cool in your standard one-on-one fighting games where the music could get more intense when then battle becomes more intense.