| Author |
Message |
Inventor
Stream Operator
Joined: Oct 13, 2007
Posts: 6221
Location: near Austin, Tx, USA
Audio files: 267
|
 Posted: Sat May 10, 2008 5:03 am Post subject:
David Cope's AI Music Composition
Subject description: If the Turing test were musical, would it pass? |
  |
|
I recently got word of the following from the ChucK Users Group mailing list (it's not about ChucK). Unlike my silly experiments with under 50 neurons, whatever David Cope has going sure seems amazing to me. The following is from the email:
If you are not familiar with David Cope's work, it represents a landmark
moment in computer creativity (similar to Kurzweil's AARON application
in the visual sphere). Cope's Experiments in Musical Intelligence set
the bar for generative music by analyzing scores and writing new pieces
in the style of the composer analyzed. If the turing test was a musical
one, experiments in musical intelligence would be likely to pass the
test. Those who encounter this work are often excited or even frightened
by its implications. I encourage you to come and engage in this
important work directly. You can listen to realizations of Cope's
experiments in musical intelligence scores here (listen to "After Bach",
"After Beethoven", etc. - there might be a lag before playing):
http://arts.ucsc.edu/faculty/cope/mp3page.htm
I had trouble loading the music, perhaps your system will do better. I feel that the limited bandwidth of music (20 kHz or so) compred to vision or a full brain model makes artificial intelligence advances of this type more likely to emerge in the music field than in most any other field, except for those areas already solved/enhanced by AI. Anyway, I thought some folks might find the work interesting so I posted it. Comments?
_________________
"Let's make noise for peace." - Kijjaz |
|
|
Back to top
|
|
 |
bachus
Joined: Feb 29, 2004
Posts: 2922
Location: Up in that tree over there.
Audio files: 5
|
Posted: Sat May 10, 2008 5:19 am Post subject:
Re: David Cope's AI Music Composition
Subject description: If the Turing test were musical, would it pass? |
|
|
| Inventor wrote: | | Unlike my silly experiments with under 50 neurons, whatever David Cope has going sure seems amazing to me. The following is from the email: |
As I recall Cope does not make any use of neural net techniques. He has a number of excellent books in print if you are interested. One of his programs is available as freeware for the Mac -- but I aint got a Mac -- the only time I ever really wanted one 
_________________
The question is not whether they can talk or reason, but whether they can suffer. -- Jeremy Bentham |
|
|
Back to top
|
|
|
x_x
Joined: May 05, 2008
Posts: 215
Location: mother earth
Audio files: 4
|
| Posted: Sat May 10, 2008 10:55 pm Post subject:
|
|
|
I've heard his work and I've read one of his books. David Cope has been working with his EMI (experiments in musical intelligence) for about 20 years, about 20,000 lines of lisp code.
The idea is to take the style of a composer (with two pieces of the author) then his program does an analysis and "extracts" the style this is taking what the two pieces have in common. After that you have a piece of music.
This is an interesting article that you can read by Douglas Hofstadter:
http://www.unc.edu/~mumukshu/gandhi/gandhi/hofstadter.htm
I'll quote a part of the article:
| Quote: | EMI's central modus operandi, given a set of input pieces, is:
(1) chop up; (2) reassemble.
There are, of course, significant principles constraining what can be tacked onto what, and these principles are formulated so as to guarantee coherence. I summarize these two principles as follows:
(1) Make the local flow-pattern of each voice similar to that in source pieces;
(2) Make the global positioning of fragments similar to that in source pieces.
These could be likened to two types of constraints that a jigsaw-puzzle solver naturally exploits when putting together a jigsaw puzzle:
(1) The shape of each piece meshes tightly with those of neighboring pieces;
(2) The stuff shown on each piece makes sense in the context of the picture.
The former of these constraints might be characterized as "syntactic meshing", or meshing based solely on "form", while the latter could be characterized as "semantic meshing", or meshing based solely on "content". In isolation, perhaps neither of them would be too impressive, but when used together, they form a powerful pair of constraints. |
Cope's work is truly amazing, his EMI can even imitate the style of Shoenberg. That's why I admire Stockhausen, I don't think EMI could replicate Stockhausen. |
|
|
Back to top
|
|
|
Low Note
Joined: Jul 20, 2007
Posts: 146
Location: New Jersey
Audio files: 2
|
| Posted: Sun May 11, 2008 12:52 pm Post subject:
|
|
|
| I would love to get my hands on that and feed the program two pieces by very different composers and see what happens. |
|
|
Back to top
|
|
|
|
Forum index » Discussion » Composition
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
