issue07

EMUSIC-L Digest                                      Volume 68, Issue 07

This issue's topics: Algorithmic composition
	
	Algorithm Composition
	Composition?
	algorithmic composition list
	Composition?
	Algorithm Composition (5 messages)

Your EMUSIC-L Digest moderator is Joe McMahon .
You may subscribe to EMUSIC-L by sending mail to listserv@american.edu with 
the line "SUB EMUSIC-L your name" as the text.
 
The EMUSIC-L archive is a service of SunSite (sunsite.unc.edu) at the 
University of North Carolina.


------------------------------------------------------------------------
Date:         Mon, 12 Sep 1994 15:29:24 BUE3
From:         Alejandro Feiges 
Subject:      Algorithm Composition

Hi all!

I'm developing a programm that composes music, based on the information
aquired from MIDI files.
I want to know what is Algorithm Composition, which techniques it has, and
whatever that makes this theme.

Are there papers talking about this?

Thanks!
    Alejandro

----------------*>        <*---------------
|    Alejandro Feiges                     |
|    Tel      :  54-1-381-4624            |
|    Internet :  ale@chopin.satlink.net   |
|                feiges@colort.edu.ar     |
----------------*>        <*---------------

------------------------------
Date:         Tue, 13 Sep 1994 15:54:23 -0400
From:         "L Mccoy (GD 1999)" 
Subject:      


TO WHOM


I have just now joined the list.  This is my first list.  I am whelmed,
but not overly so.  Could someone tell me more about what goes on?  Are
AI topics frequently discussed here?  Other computer-assisted composition?
Thanks.

Scott McCoy

------------------------------
Date:         Wed, 14 Sep 1994 00:51:00 -0700
From:         Nate Waddoups 
Subject:      Re: your mail

On Tue, 13 Sep 1994, Joe McMahon wrote:

> > I have just now joined the list.  This is my first list.  I am whelmed,
> > but not overly so.  Could someone tell me more about what goes on?  Are
> > AI topics frequently discussed here?  Other computer-assisted composition?

> Welcome, Scott! We haven't discussed AI and computer-assisted composition
> much because we tend to have more studio types. However, many of us (myself
> certainly included) love to discuss such topics. Please feel free to post
> a new message and get something started.

With respect to AI and composition, there is a list devoted to
algorithmic composition techniques of all kinds that might interest you.

I don't remember exactly how I subscribed to it (it was a while ago), but
try sending a message with the word 'HELP' in the body (and maybe the
subject too, for good measure) to: listserv@heinous.isca.uiowa.edu

It will probably tell you to send a message with 'SUBSCRIBE ALGO-COMP' or
something like that...

Hope this helps.

(And welcome here, by the way)
+-------------------+-------------------------------------------------------
| natew@halcyon.com | You write Windows code?  Mac code?  OS/2 PM code?
+-------------------+ Want a great job in the pacific northwest?
                      Send me email, we'll talk...

------------------------------
Date:         Fri, 16 Sep 1994 10:58:39 +0200
From:         Terje Winther 
Subject:      Re: Composition?

>TO WHOM
>
>
>I have just now joined the list.  This is my first list.  I am whelmed,
>but not overly so.  Could someone tell me more about what goes on?  Are
>AI topics frequently discussed here?  Other computer-assisted composition?
>Thanks.
>
>Scott McCoy

No, I can't say. I would very much welcome more composition-related topics.


(:  tw.

Terje Winther
Produksjonskonsulent/Producer
Ny Musikk / I.S.C.M.
(Norwegian section of the International Society for Contemporary Music)
Ny Musikks Komponistgruppe / New Music Composers' Group
Tollbugata 28
N-0157  OSLO
Norway

E-mail: terje@notam.uio.no
Telephone: +47 2233 7090
Telefax:   +47 2233 7095

------------------------------
Date:         Fri, 16 Sep 1994 10:59:23 +0200
From:         Terje Winther 
Subject:      Re: Algorithm Composition

Hello, Alejandro.

>Well. I want that the computer be able to compose a classical music piece.
>I input, for example, 20 pieces of Bach, inventions for 2 and 3 voices,ie.
>So, i would like that the computer makes a new piece, keeping inventions
>style.

Gee,this was not what I had in mind. Algoritmic COMPOSITION means that you
(or someone else) makes an algorithm (the process itself) that will be used
as a basic foundation for making music. Remember that the process
inprinciple doesn't have do be done on a computer, but this is usually the
case, since algorithms tend to be rather complex. An extremely simple
algorithm can be like:
- Take a note (this is your input)
- The nest note should be higher (note that it doesn't say anything about
HOW much higher)
- When the range reaches the defined high limit, repeat the process from start
- Do this until the defined number of cycles have been made.

Speaking of complex; an algorithm that can recreate Bach's inventions,
would be extremely complex. When I did my counterpoint exam, I didn't do a
high score, just because I had studied Bach's OWN works, and not the rules
that had been EXTRACTED from his works. What I mean to say, is that Bach
didn't make music that fitted into his rules, because he had a firm grip on
things, and could be very free within the framework he was working in.

>Well, you said that i change the algorithm if i don't like what i obtained,
>but how should i do it?

In my example, if you don't like the result, you can change how much higher
the next note sould be.

>I should do it?

Yes, definitely! If you don't, you don't compose!

>The computer must to find the correct algorithm?

The essential question.

1. I don't know any program that you can feed musical examples, and have
the program extract the rules in order to recreate a new piece in the same
genre. (Anybody?) If this is what you're working on Alejandro, boy, you do
have a large project ahead of you. But please, if you do reach you goal,
let me (and the rest of us) know.

2. There are a lot of programs that mimics several genres, with a sort of
"intelligent" auto-accompament function - this is automatic (but simple)
composition. I don't like to call this to compose, because if you use the
raw material coming out of these programs, someone else have really
composed the material (OK, the rules then), and even though there are a lot
of random functions, it doesn't make interessting music. You CAN however
use the material as a starting point, and then develop the piece further.
THAT, would be composing.

>Now, that you know what i want to do,Do you know which kind of information
>i  must to look for?

You can buy the commercial programs, or get the freeware/shareware
programs, read textbooks et al. on the subject, og get an Ph.D. in music,
art, computerprogramming et al.

>Thanks for your time,
>    Alejandro Feiges.
>
>must to look for
>
>----------------*>        <*---------------
>|    Alejandro Feiges                     |
>|    Tel      :  54-1-381-4624            |
>|    Internet :  ale@chopin.satlink.net   |
>|                feiges@colort.edu.ar     |
>----------------*>        <*---------------



(:  tw.

Terje Winther
Produksjonskonsulent/Producer
Ny Musikk / I.S.C.M.
(Norwegian section of the International Society for Contemporary Music)
Ny Musikks Komponistgruppe / New Music Composers' Group
Tollbugata 28
N-0157  OSLO
Norway

E-mail: terje@notam.uio.no
Telephone: +47 2233 7090
Telefax:   +47 2233 7095

------------------------------
Date:         Fri, 16 Sep 1994 11:33:42 -0500
From:         Arne Claassen ISE 
Subject:      Re: Algorithm Composition

> >Well. I want that the computer be able to compose a classical music piece.
> >I input, for example, 20 pieces of Bach, inventions for 2 and 3 voices,ie.
> >So, i would like that the computer makes a new piece, keeping inventions
> >style.

Bakc in the May 1993 issue of Electronic musician was an article about a
program that seemed to do waht you require. It was an ars combinatoria
algorithm, based on tridas. Since it appears that a number of classical
composers used recombination in their style. The program written in
common lisp was supposed to be able to extract the syle from some examples and
create new measures witht eh original signature. IT was an interesting article.

They said that the program is available on PAN in the EM SIG (at least back
then). The auther was David Cope. If anyone has ever downloaded this program
or could still download it from PAN. Please let me know.

Thanks..

--
Arne F. Claassen             
On the Web:           
-----------------------------------------------------------------------------
"Somebody's got his finger on a DX7!"                             

------------------------------
Date:         Fri, 16 Sep 1994 11:10:19 -0600
From:         Bet my hair is longer than yours! 
Subject:      Re: Algorithm Composition

>Arne Claassen wrote:
>Bakc in the May 1993 issue of Electronic musician was an article about a
>program that seemed to do waht you require. It was an ars combinatoria
>algorithm, based on tridas. Since it appears that a number of classical
>composers used recombination in their style. The program written in
>common lisp was supposed to be able to extract the syle from some examples and
>create new measures witht eh original signature. IT was an interesting article
>
>They said that the program is available on PAN in the EM SIG (at least back
>then). The auther was David Cope. If anyone has ever downloaded this program
>or could still download it from PAN. Please let me know.

I'd like to see the actual program too.  David Cope wrote a book about this
program.  I think the title is Computers and Musical Style and it was
published in 1991.

-Doug

------------------------------
Date:         Fri, 16 Sep 1994 14:06:11 -0500
From:         Siberian Khatru 
Subject:      Re: Algorithm Composition

        With respect to the idea of a computer extracting style from a
composer and composing works in that style, check out David Cope's book
on the subject.  He has spent a vast amount of time doing research and
creating algorithms for that whole thing.  The book (I think) is called
_Computers and Musical Style_ (or some such thing).

Cheers,

Jon Southwood
jsouthwo@keller.clarke.edu

******************************************************************************
"I once played poker with a deck of tarot cards. I got a full house and
four people died."-Steven Wright
"If I had only known, I would have been a locksmith."-Albert Einstein
"Truly great madness cannot be achieved without significant intelligence."
                                                        -Henrik Tikkanen
******************************************************************************

------------------------------
Date:         Sun, 18 Sep 1994 01:00:31 -0700
From:         Nate Waddoups 
Subject:      Re: Algorithm Composition

On Fri, 16 Sep 1994, Terje Winther wrote:

> 1. I don't know any program that you can feed musical examples, and have
> the program extract the rules in order to recreate a new piece in the same
> genre. (Anybody?) If this is what you're working on Alejandro, boy, you do
> have a large project ahead of you. But please, if you do reach you goal,
> let me (and the rest of us) know.

I've done a little bit of work towards this end, actually.  The program
uses a neural network, trained on a few bars of 'flight of the
bumblebee,' and, with a little work, I'll have the outputs fed back to
the inputs.  The results should be interesting.

Unfortunately, I was interrupted in this project, and haven't yet gotten
back to it, but if I get it working, I'll be sure to post here.

I have a 'gut feeling' that neural nets are well suited to this kind of
task, and before long, I'll hopefully be able to back that up...
+-------------------+-------------------------------------------------------
| natew@halcyon.com | You write Windows code?  Mac code?  OS/2 PM code?
+-------------------+ Want a great job in the pacific northwest?
                      Send me email, we'll talk...

------------------------------
End of the EMUSIC-L Digest
******************************