Copy Link
Add to Bookmark
Report
Music-Research Digest Volume 5 : Issue 27
Music-Research Digest Sun, 18 Mar 90 Volume 5 : Issue 27
Today's Topics:
Fractal music generation
job posting
Music Education (was: Re: MR Vol. 5, #21)
PD Music editing software anyone ?
Sound Design and Music Publishing Colloquium
*** Send contributions to Music-Research@uk.ac.oxford.prg
*** Send administrative requests to Music-Research-Request
*** Overseas users should reverse UK addresses and give gateway if necessary
*** e.g. Music-Research@prg.oxford.ac.uk
*** or Music-Research%prg.oxford.ac.uk@nsfnet-relay.ac.uk
----------------------------------------------------------------------
Date: 10 Mar 90 17:17:15 GMT
on.cc.bingvaxu>lich <alex%bilver%tarpit%ge-dab%crdgw1%albanycs%leah@edu.binghamt
Subject: Fractal music generation
To: music-research@prg
I need some kind souls to send me some algorithms to generate music based
on fractals.
I don't know much about this subject, but it occurred to me that a fractal
music algorithm might be used as an aid in writing fugues. I have tried
one experiment already, based on the geometrical shape known as a "Koch
curve", where a pitch and duration corresponds to segment angle and length,
respectively.
My suspicion is that I'm going about it all wrong. My experimented generated
a sequence of notes that sounded interesting. The problem is that it was
a single sequence and the notes were all the same length. How can a fractal
generator be made to create overlapping sequences of notes which have
harmonically correct relations to each other?
Music is a serial event that unfolds as time passes. I am only able to
visualize fractals as geometric entities which where all parts exist at once.
Please e-mail me suggestions, no matter how simple or bizarre. They will
be appreciated. Thanks.
--
/// Alex Matulich
/// Unicorn Research Corp, 4621 N Landmark Dr, Orlando, FL 32817
\\\/// alex@bilver.UUCP ...uunet!tarpit!bilver!alex
\XX/ From BitNet use: bilver!alex@uunet.uu.net
------------------------------
Post: 76 of 179
From: alex@bilver.UUCP (Alex Matulich)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Fractal Music Generation (summary)
Keywords: I need help
Date: 9 Apr 90 01:19:56 GMT
Organization: William J. Vermillion, Winter Park, FL
Lines: 73
Several weeks ago I posted an a plea for help in comp.music and
comp.sources.wanted for an algorithm to generate fractal music. I lost the
original text of my posting, but the gist of it was this:
A fugue is a piece of music rich in self-similar structure. J. S. Bach, a
master at writing fugues, was able to maintain up to six instrumental parts
playing a short theme in different ways -- at different pitches, different
speeds, inverted, upside-down, backwards, and so on -- and it all fit
together too!
Fractals also are rich in self-similar structure. By definition, after all,
a fractal IS a self-similar object. The parallels between fractals and
fugues seem so close, I thought, that maybe a MUSICAL fractal generator
could be developed as an aid in writing fugues.
I tried an experiment based on the generation of a Koch curve, assigning
a relationship between note pitch and line angle, and another relationship
between note duration and line length. My experimented generated a
sequence of notes that sounded interesting. The problem is that it was
a single monotonic sequence. How can a fractal music generator be made
to create overlapping sequences of notes which have harmonically correct
relations to each other?
I got 10 replies. Three offered algorithmic advice, and everyone else
wanted the same information I was asking. Apparently there is a fair
amount of interest out there, but little knowlege.
You there, reading this: If you know anything about generating fractal
music, send e-mail or post an article, and quit lurking in the shadows!
Now for the summary [My comments are in brackets]:
(From Kevin Quitt uunet!demott!kdq)
Use 6 dice for the note to be selected and another six for the length.
Roll die 0 for every note, die 1 half as often, die 2 half as often as
die 1, etc., and add then numbers to determine a number for selecting a
note within a predetermined key. Accidentals are also randomly played.
More dice tends to smooth out the music, larger values gives more variation.
[Very interesting, but I was looking for something more deterministic.]
(From Doug Bischoff uunet!psumv.psu.edu!deb110)
A 3-D fractal may be used to control 3 different musical event attributes
plus a fourth if the points are colored. Use the X axis as a time scale.
For each X-axis time point, perform additive synthesis using the Y axis
for harmonics or frequencies and the Z axis for volume, and use the color
of the point on the X axis to determine a fundamental frequency from which
each harmonic is calculated.
[Yes, a more deterministic algorithm, but it seems to me that such an
algorithm would create music having little natural unity and flow since
(depending on the initial 3-D object) unfolding musical events might not
have any real dependence on previous events. I'd like to be able to
give the fractal music generator an initial theme and see where it goes.]
(From Fred Sena uunet!samsung.com!infinet!sena)
Map the numerical values from an iterative fractal generator onto some
harmonic rules. For example, choose notes that have some harmonic relation
to each other (like a blues scale) and let the generator choose the
sequence. Other levels of structure could be added to fractally change
fundamental keys, note lengths, and so on.
[This is very similar to what I was trying to do with the Koch curve
generator.]
Since I posted my original question, the alt.fractals newsgroup has been
created, so I'm also posting this summary there.
--
/// Alex Matulich
/// Unicorn Research Corp, 4621 N Landmark Dr, Orlando, FL 32817
\\\/// alex@bilver.UUCP ...uunet!tarpit!bilver!alex
\XX/ From BitNet try: IN%"bilver!alex@uunet.uu.net"
Post: 77 of 179
From: edgar@shape.mps.ohio-state.edu (Gerald Edgar)
Newsgroups: comp.music,alt.fractals
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 9 Apr 90 12:37:24 GMT
Organization: The Ohio State University, Dept. of Math.
Lines: 28
In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
>
>I tried an experiment based on the generation of a Koch curve, assigning
>a relationship between note pitch and line angle, and another relationship
>between note duration and line length. My experimented generated a
This sounds like something I have done. I used about 10 of the common
"dragon curves" (including Koch). The change in pitch was related to
the angle (360 degrees corresponds to an octave), and duration was
related to line segment length. Of course, the duration should be a
POWER of the line length (the exponent is the reciprocal of the
fractal dimension) in order to achieve true self-similarity.
The curve known as "McWorter's pentigree" uses angles of 72 and 144 degrees,
which correspond to intervals not used in Western music. Peculiar.
If there is some interest I can post the programs. (Logo source code,
or Macintosh executable.)
(By the way, there is some literature on "fractal music", and it is NOT
this!!!)
--
Gerald A. Edgar
Department of Mathematics Bitnet: EDGAR@OHSTPY
The Ohio State University Internet: edgar@mps.ohio-state.edu
Columbus, OH 43210 ...!{att,pyramid}!osu-cis!shape.mps.ohio-state.edu!edgar
Post: 79 of 179
From: george@shumv1.ncsu.edu (George Browning)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 9 Apr 90 15:19:58 GMT
Reply-To: george@shumv1.ncsu.edu (George Browning)
Organization: NCSU Computing Center
Lines: 36
In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
>Several weeks ago I posted an a plea for help in comp.music and
>comp.sources.wanted for an algorithm to generate fractal music. I lost the
>original text of my posting, but the gist of it was this:
>
I have an article from the book Fundamental Algorithms for
Computer Graphics written by Richard F. Voss that talks about fractal
music. Voss says "One of my exciting discoveries was that almost all
musical melodies also mimic 1/f noise." He gives some pictures and
examples, including a couple of "spectral density measurements of the
pitch variations in various types of music showing their common
correlations as 1/f noise" These graphs show such things as Medieval
music up to 1300, Beethoven's 3rd Symphony and the Beatles Sgt. Pepper.
I am not sure exactly how to generate 1/f noise (it doesn't look too
easy) but I will know how to by the end of the semester, as my graphics
project depends on it. I am going to use it to make both terrain maps
and texture maps for water.
You may also want to look at:
Voss, R. F. and Clarke, J. "1/f Noise in Music: Music from 1/f Noise",
J. Accous. Soc. Am. 63, (1978), 258-263.
Voss, R. F. and Clarke, J. "'1/f noise' in music and speech", Nature
258, 317-8 (1975).
- Jeff
--
_____________________________________________________________________
| George Browning North Carolina State University |
| george@shumv1.ncsu.edu Raleigh, NC |
|___________________________________________________________________|
Post: 87 of 179
From: billd@fps.com (Bill Davidson)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation
Keywords: I need help
Date: 10 Apr 90 19:27:13 GMT
Followup-To: comp.music
Organization: FPS Computing Inc., San Diego CA
Lines: 11
In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
[asks for info on fractal music]
I have two references:
Dietrick E. Thomsen, "Making Music Fractally", Science News, Mar 22, 1980
Richard F. Voss, "Random Fractal Forgeries", SIGGRAPH '85 Course Notes
for Fractals: Basic Concepts, Computation and Rendering.
--Bill Davidson
Post: 89 of 179
From: mu298ac@sdcc6.ucsd.edu (Philip Marlowe)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 11 Apr 90 03:28:53 GMT
Followup-To: comp.music
Organization: University of California, San Diego
Lines: 41
e Browning) writes:9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu (Georg
>In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
>
> I have an article from the book Fundamental Algorithms for
>Computer Graphics written by Richard F. Voss that talks about fractal
>music. Voss says "One of my exciting discoveries was that almost all
>musical melodies also mimic 1/f noise." He gives some pictures and
This is an incredibly obvious statement to make. Stepwise
motion is an important attribute of many tonal melodies,and 1/f
noise
generates stepwise motion. So why can't you program 1/f noise to
produce good tonal melodies? Because tonal melody is not random; it
has very strong directionality, and any programmer who wants to
have an algorithm that would produce good tonal melodies has to take
goal-oriented motion into account, which I don't believe is possible
with fractals. Traditional tonal melody is incredibly causal. It
can not be modeled on random procedures. If there is any way for
computers to write good, catchy, tonal melodies, I suspect it must
be through an alogrithm which is contructed on the rules that most
musicians learn in theory class for writing melodies (too much
stepwise motion in the same directionis boring; an upward leap is
usually followed by a downward resolution by step, unless it's
outlining a triad; etc.)
If you really want some insight into how tonal melody works,
and why good melodies *sound* good, try reading Leonard Meyer's
_Emotion_and_Meaning_in_Music_ and _Explaining_Music_.
Previous discussions in this group about fugues being
"self-similar" shows a lack of understanding about just what a fugue
is. Just because something is repeated at the same level, it doesn't
imply self-similarity (or does it?) If you examine a Bach fugue at
the middleground or background level, you will see absolutely no
replication of the subject or countersubject, say. What is
self-similar, perhaps, on these levels will be the movement from
tonic to dominant to tonic, but even this isn't guaranteed, and
besides, it's a self-similarity shared by just about every other
piece of baroque and classical music, as Schenker would have us
believe. I really don't think you can call thematic unity
self-similarity.
Post: 90 of 179
From: err@fibercom.COM (Eric Rubin)
Newsgroups: comp.music,alt.fractals
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 10 Apr 90 12:44:48 GMT
Reply-To: err@fibercom.COM (Eric Rubin)
Organization: FiberCom Inc., Roanoke, Virginia
Lines: 12
-state.edu (Gerald Edgar) writes:zaphod.mps.ohio-state.edu> edgar@shape.mps.ohio
>If there is some interest I can post the programs. (Logo source code,
>or Macintosh executable.)
I'd like to see the Logo source code.
--
Eric Rubin INTERNET: err@fibercom.com
FiberCom, Inc. UUCP: ...!uunet!fibercom!err
P.O. Box 11966 PHONE: 703-342-6700, 800-423-1183 x348
Roanoke, VA 24022-1966 FAX: 703-342-5961
Post: 94 of 179
From: smoliar@vaxa.isi.edu (Stephen Smoliar)
Newsgroups: comp.music,rec.music.classical
Subject: Re: Fractal Music Generation (summary)
Summary: tonal music and random procedures
Keywords: music theory, composition, Meyer
Date: 11 Apr 90 14:34:03 GMT
Reply-To: smoliar@vaxa.isi.edu (Stephen Smoliar)
Organization: USC-Information Sciences Institute
Lines: 170
In article <9613@sdcc6.ucsd.edu> mu298ac@sdcc6.ucsd.edu (Philip Marlowe)
writes:
>In article <1990Apr9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu
>(George Browning) writes:
>>In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
>>
>> I have an article from the book Fundamental Algorithms for
>>Computer Graphics written by Richard F. Voss that talks about fractal
>>music. Voss says "One of my exciting discoveries was that almost all
>>musical melodies also mimic 1/f noise."
>
> This is an incredibly obvious statement to make. Stepwise
>motion is an important attribute of many tonal melodies,and 1/f
>noise
>generates stepwise motion. So why can't you program 1/f noise to
>produce good tonal melodies? Because tonal melody is not random; it
>has very strong directionality, and any programmer who wants to
>have an algorithm that would produce good tonal melodies has to take
>goal-oriented motion into account, which I don't believe is possible
>with fractals. Traditional tonal melody is incredibly causal. It
>can not be modeled on random procedures. If there is any way for
>computers to write good, catchy, tonal melodies, I suspect it must
>be through an alogrithm which is contructed on the rules that most
>musicians learn in theory class for writing melodies (too much
>stepwise motion in the same directionis boring; an upward leap is
>usually followed by a downward resolution by step, unless it's
>outlining a triad; etc.)
>
There have been no end of attempts in this direction, and none have been
particularly successful. The problem is that random procedures are being
applied at the wrong level of granularity. To try to draw an appropriate
analogy, it is sort of like assuming that you could construct sentences
through random selection of syllables. Lejaren Hiller actually tried to
do something like this in his "Computer Cantata," experimenting with Markov
processes with different "prior memory capacity;" and the best he could do
was come up with the occasional coherent word or two. People who have been
interested in random sentence generation know that you get a lot more mileage
out of defining your world in terms of a context-free grammar and then using
random procedures to determine which productions you invoke.
There are a few analogies to this practice in music. If we consider the model
era, which preceded tonality, we can find an example of such a context-free
grammar in Dom Paolo Ferretti's ESTHETIQUE GREGORIENNE. (The French
translation of this book appeared in 1938, so don't expect to find any
of Chomsky's terminology in it.) Ferretti devotes considerable text to
the analysis of CENTONIZATION, a process by which new plainchants were
made up by piecing together fragments (CENTONS, from the French for a
patch in a patchwork quilt) of old ones. Ferretti was astute enough
to realize that one could not put the patches together any old way;
and he offers up a table which, for all intents and purposes, is a
set of productions for centonizing chants in the Dorian mode. It works
rather well; and I implemented a "random sentence generator" based on
this table as part of my doctoral thesis.
There are any number of "dice composers" which apply a similar principle to
tonal music, the most famous being by Mozart. Here, a random procedure is
invoked only for the selection of the terminals. The nonterminal nodes of
the parse tree have been fixed by the "composer." The bulk of his work has
gone into making sure that the choices of terminals for any given node are
interchangeable.
I find it slightly disheartening that people continue to disregard what appears
to be an important lesson from these experiments, which is that composers tend
to work at a higher level of granularity than individual notes. This is not to
say that there are not situations in which choosing a specific note is not
important. Certainly, every writer has situations in which it is critically
important to choose just the right word; but if every writer applied that
attention to EVERY word, very little would get written. Composition is a
matter of working which "musical ideas." None of us may be able to pin down
just what that phrase denotes, but my own intuition tells me that it has a lot
to do with memories of past listening experiences. To some extent, all
composers centonize--picking up materials from past experiences and finding
new ways in which to assemble them. If we are determined to seek out
algorithmic rules, then it would seem that these rules should be directed
at two key questions:
1. How do we identify such units of material?
2. How do we determine how, given a collection of those units,
they may be properly assembled?
> If you really want some insight into how tonal melody works,
>and why good melodies *sound* good, try reading Leonard Meyer's
>_Emotion_and_Meaning_in_Music_ and _Explaining_Music_.
>
Meyer probably deserves due credit for being one of the first to recognize that
a question like "how tonal melody works" is probably as much a matter of
psychology as it is of music theory (if not more so). However, Meyer's
understanding of psychology is rather naive. He seems more interested
in exhibiting the BREADTH of his reading in non-musical subjects than
in trying to apply any of those areas in DEPTH. Anyone interested in
a more serious exposition of how cognitive psychology may provide the
sorts of insights Philip has in mind would do better to turn to a book
like John Sloboda's THE MUSICAL MIND. (I disagree with a good deal of
what Sloboda says in this book, but he DOES know how to lay out the relevant
issues.)
> Previous discussions in this group about fugues being
>"self-similar" shows a lack of understanding about just what a fugue
>is. Just because something is repeated at the same level, it doesn't
>imply self-similarity (or does it?) If you examine a Bach fugue at
>the middleground or background level, you will see absolutely no
>replication of the subject or countersubject, say. What is
>self-similar, perhaps, on these levels will be the movement from
>tonic to dominant to tonic, but even this isn't guaranteed, and
>besides, it's a self-similarity shared by just about every other
>piece of baroque and classical music, as Schenker would have us
>believe. I really don't think you can call thematic unity
>self-similarity.
Again, the issue seems to be one of granularity. What is REALLY important
about Schenker is that he tried to make us acknowledge that analysis must
proceed at many different levels of granularity. Unfortunately, his (German?)
sense of order led him to assume that these granules could be neatly embedded
in a hierarchy; and this assumption has been carried on by both Meyer and
Narmour, on one hand, Lerdahl and Jackendoff, on another, and Yeston, on a
third. (There are probably several more hands lurking out there, but I am
not particularly inclined to catalog them.) Fortunately, Lewin seems to have
broken out of this "dictatorship of the hierarchy" in his recent "Music Theory,
Phenomenology, and Modes of Perception" paper; and my own guess is that he
will benefit from this liberation.
Another question is why we wish to place so much emphasis on "self-similarity."
Do we, as listeners, devote so much of our cognitive attention so simply being
able to recognize that we have heard something before? Let me try sticking my
neck out on a hypothesis here which has been inspired by the work of Marvin
Minsky (who has written about music, as well as artificial intelligence).
Minsky believe that much of understanding is a matter of being able to
recognize, and account for, DIFFERENCES. This is a bit like saying that
much of music is concerned with what we loosely call "variation" and the
fact that, as music history has progressed, we have become more and more
liberal about what constitutes a variation. What makes the game interesting,
however, is that we cannot perceive differences unless we gauge them against
some standard of SAMENESS. For example, in BOLERO, we quickly recognize that
variation is almost entirely a matter of orchestral color (all that parallel
motion is almost like trying to build up new sound spectra) while everything
else stays the same. Thus, we seek out self-similarity not for its own sake
but for the ability to detect differences. Fugues are exercises in how a
melodic motif may be engaged in many different contexts, so that it is CONTEXT
which becomes the basis for variation.
In all fairness, I should point out that Meyer has tried to pursue a similar
line of thought. Much of his writing in music theory is concerned with
EXPECTATIONS. However, he seems to believe that expectations may be grounds
on universal principles, such as those of gestalt psychology. I, on the other
hand, think they are grounded on our ability to perceive self-similarity,
either within the context of a single composition or with respect to our
past listening experiences. In other words, we seek out trying to identify
what we are hearing as being like something we have heard before, because then
we will assume that it will "go the same way." This becomes a basis for our
expectations, and we listen to hear if those expectations are satisfied or if
something different occurs. Thus, the mind is engaged; and we are now
exhibiting the behavior of listening to music.
(One final point: I am cross-posting this to rec.music.classical, since that
bulletin board provides a home for many opinions about both composition and
music theory.)
=========================================================================
USPS: Stephen Smoliar
USC Information Sciences Institute
4676 Admiralty Way Suite 1001
Marina del Rey, California 90292-6695
Internet: smoliar@vaxa.isi.edu
"Only a schoolteacher innocent of how literature is made could have written
such a line."--Gore Vidal
Post: 95 of 179
From: kassover@jupiter.crd.ge.com (David Kassover)
Newsgroups: comp.music
Subject: Re: Fractal Music Generation (summary)
Date: 11 Apr 90 15:11:34 GMT
Organization: Aule-Tek, Inc.
Lines: 38
In article <9613@sdcc6.ucsd.edu> mu298ac@sdcc6.ucsd.edu (Philip Marlowe) writes:
rge Browning) writes:9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu (Geo
| | In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
| |
| | I have an article from the book Fundamental Algorithms for
| | Computer Graphics written by Richard F. Voss that talks about fractal
| | music. Voss says "One of my exciting discoveries was that almost all
| | musical melodies also mimic 1/f noise." He gives some pictures and
|
| This is an incredibly obvious statement to make. Stepwise
| motion is an important attribute of many tonal melodies,and 1/f
| noise
| generates stepwise motion. So why can't you program 1/f noise to
| produce good tonal melodies? Because tonal melody is not random; it
| has very strong directionality, and any programmer who wants to
| have an algorithm that would produce good tonal melodies has to take
| goal-oriented motion into account, which I don't believe is possible
| with fractals.
...
About a year and a half ago, I was at a lecture given by
Mandelbrot. Someone asked him about fractal music. He replied
to the effect that he had heard the output of some experiments in
that area, and that they didn't "sound good".
(Whatever that means)
We in the audience were not given referencer the opportunity
to hear similar musical pieces and thus form our own opinions.
De gustibus non est disputandum. Or as my father would say,
"Sahzeechizone"
--
===================================================
Post: 106 of 179
From: quiniou@calculo.irisa.fr (Rene Quiniou)
Newsgroups: comp.music,rec.music.classical
Subject: Re: Fractal Music Generation (summary)
Keywords: music theory, composition, Meyer
Date: 12 Apr 90 07:20:44 GMT
Reply-To: quiniou@irisa.fr
Organization: Irisa, Rennes(FR)
Lines: 41
Could you post the exact references of the sources cited in your article
as well as your thesis, please?
In article <12859@venera.isi.edu>, smoliar@vaxa.isi.edu (Stephen
Smoliar) writes:
|> In article <9613@sdcc6.ucsd.edu> mu298ac@sdcc6.ucsd.edu (Philip Marlowe)
|> writes:
|> >In article <1990Apr9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu
|> >(George Browning) writes:
|> >>In article <562@bilver.UUCP> alex@bilver.UUCP (Alex Matulich) writes:
|> >>
|> >> I have an article from the book Fundamental Algorithms for
|> >>Computer Graphics written by Richard F. Voss that talks about fractal
|> grammar in Dom Paolo Ferretti's ESTHETIQUE GREGORIENNE. (The French
|> translation of this book appeared in 1938, so don't expect to find any
|> There are any number of "dice composers" which apply a similar principle to
|> tonal music, the most famous being by Mozart. Here, a random procedure is
|> >and why good melodies *sound* good, try reading Leonard Meyer's
|> >_Emotion_and_Meaning_in_Music_ and _Explaining_Music_.
|> >
|> sorts of insights Philip has in mind would do better to turn to a book
|> like John Sloboda's THE MUSICAL MIND. (I disagree with a good deal of
|> in a hierarchy; and this assumption has been carried on by both Meyer and
|> Narmour, on one hand, Lerdahl and Jackendoff, on another, and Yeston, on a
|> third. (There are probably several more hands lurking out there, but I am
|> not particularly inclined to catalog them.) Fortunately, Lewin seems
to have
|> broken out of this "dictatorship of the hierarchy" in his recent
"Music Theory,
|> Phenomenology, and Modes of Perception" paper; and my own guess is that he
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
QUINIOU Rene quiniou@irisa.fr
INRIA / IRISA Phone : +33 99 36 20 00
Campus Universitaire de Beaulieu Fax : 99 38 38 32
35042 RENNES CEDEX - FRANCE Telex : UNIRISA 950 473F
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Post: 107 of 179
From: d88-jwa@nada.kth.se (Jon W{tte)
Newsgroups: comp.music
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 12 Apr 90 15:02:01 GMT
Reply-To: d88-jwa@nada.kth.se (Jon W{tte)
Organization: Royal Institute of Technology, Stockholm, Sweden
Lines: 32
In article <9613@sdcc6.ucsd.edu>, mu298ac@sdcc6.ucsd.edu (Philip
Marlowe) writes:
> In article <1990Apr9.151958.26859@ncsuvx.ncsu.edu>
george@shumv1.ncsu.edu (George Browning) writes:
> This is an incredibly obvious statement to make. Stepwise
> motion is an important attribute of many tonal melodies,and 1/f
> noise
> generates stepwise motion. So why can't you program 1/f noise to
> produce good tonal melodies? Because tonal melody is not random; it
> has very strong directionality, and any programmer who wants to
Actually, try making a plot of baroque music, and compare that to
1/f-squared noise. You'll find some interesting similarities !
(Yes, it's 1/f-squared and not 1/f as the original poster said)
Gregorian music is closer to 1/f-cubed or even to the fourth...
Now, where does that leave acid house ? (oh, sorry...)
> Previous discussions in this group about fugues being
> "self-similar" shows a lack of understanding about just what a fugue
> is. Just because something is repeated at the same level, it doesn't
Look at the mandelbrot set. It is self-similar, but skewed,
rotated, mirrored and transformed in various ways. Actually, I
think you could create reasonable fuge-LIKE music (actually, a whole
new type) that was enjoyable using fractals.
--- Stay alert ! - Trust no one ! - Keep your laser handy ! ---
h+@nada.kth.se == h+@proxxi.se == Jon Watte
longer .sig available on request
Post: 115 of 179
From: music@batman.moravian.EDU (music)
Newsgroups: comp.music,rec.music.classical
Subject: Re: Fractal Music Generation (summary)
Keywords: music theory, composition, Meyer
Date: 12 Apr 90 19:38:04 GMT
Followup-To: comp.music,rec.music.classical
Organization: Moravian College, Bethlehem, PA
Lines: 61
But is it MUSIC? ;-)
I personally believe that all you've discussed regarding the
algorithmic process of producing (or attempting to produce) "good
tonal music" is rather more of a verbose punishment to the reader than
the gleaning of any insight to the process being attempted (no offense
intended!). I have worked around with algorithmic composition off and
on for many years, but "gave up" on attempting to create an artificial
musical learning base from which an algorithm could draw upon to
produce anything more interesting than (and this is a bad example) the
Mozart "Dice Minuet". So I personally decided that the goal of
creating "good tonal music" through "pure math" was a non sequitur to
the nature of the beast known as "tonal music". I therefore treaded
into the teritory of composers such as Xenakis (and Cage from a
philosophical, more than "technical" sense). In certain works of
Xenakis, Herma (piano) for example, the music is about as far from
"tonal" as it is from "12-tone serial" (we limit ourselves here to a
system of 12 notes; if we were to explore beyond to the reaches of
quarter-tones, arbitrary systems (i.e. Partch) we would be streaching
the mind beyond most peoples comprehenshion, something I REALLY WANT
TO DO (but that's another story!)). Xenakis uses algorithms to plot
his pitch classes, tempi, dynamics, etc. in a way that is more or less
"highly organized randomness". Personally, I think his advanced math
has little to do with the ultimate outcome of the music, but I like
what he does regardless. The music produces more of a gestalt
experience than a profoundly complex serial work (like Boulez'
Structures and any number of other works by strict 12-tone
serialists). The music comes from that great unknown: CHANCE. By
carefully controlling the elements of CHANCE on many levels (a grain
of sand to an astroid) we can then begin to produce CHANCE-based
organization, letting "nature control the music" (think of all the
combined chaos and symetry in the universe) and the "composer" guide
nature either via algorithms (serialization of chance structures) or
by subjective reasoning (nurturing nature).
Regardless of the outcome, "tonality" will be replaced by something of
a higher order: music that exists etearnally just waiting to be
"guided into place". This may smack of musical anarchy, but is the
UNIVERSE anarchic? It may seem so on certain levels but ultimately
"GOD" controls the "laws of nature" the way I like to control the
"laws of the music of nature". Strict tonality/serialism is
UN-natural. Only out of cultivating chaos can we deliver the truth of
music. Humans have too long restrained themselves into believing that
man-made rules about tonality (in the Western world at least) they
have unwhittingly enslaved themselves into a very narrow "band" of the
musical spectrum (as visible light is to the entire electromagnetic
spectrum). We must explore the outer limits of sound and learn to
appreciate them as we now appreciate "tonal" music.
(Think (philosophically) of the music of the Krell in the film
"Forbidden Planet" from the '50's. Think of 4'33''. Think of the cosmic
background radiation. Think of the Universe as "music in the making".
And, finally, you'll probably think of ME as a raving lunatic...)
---------------------------------------------------
| Stephen Heller - Music Technician | In transit from
| CSNET -> music@batman.moravian.edu | the center of
| UUCP -> ...!rutgers!liberty!batman!music | Time & Space...
| INET -> music%batman.moravian.edu@relay.cs.net |
---------------------------------------------------
Post: 117 of 179
From: rreid@esquire.UUCP ( r l reid )
Newsgroups: comp.music
Subject: 1/f
Date: 12 Apr 90 15:54:51 GMT
Reply-To: rreid@esquire.UUCP ( r l reid )
Organization: ? !
Lines: 28
Keywords:
e Browning) writes:9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu (Georg
>music. Voss says "One of my exciting discoveries was that almost all
>musical melodies also mimic 1/f noise." He gives some pictures and
The thing to be careful about is that the melodies mimic 1/f
noise, not vice versa. I've got a 1/f program that cranks out
endless (if you want) 1/f melodies, with some nice parameters
(like "swing factor"). It's pleasent enough for a while,
but eventually you find that you are going nowhere. Like,
where's the cadence?
What it is GREAT for is motive generation, if you are feeling uninspired
one fine day and you MUST get a jingle together by 5 pm. In that
case, crank 'er up, and listen for a lovely catchy little
phrase to come by. Then you can pick that out, and start doing
all the normal development kinds of things to it.
The Voss algorithm is a lot of fun to play with and you can get all kinds
of interesting things started with it. To make music, you'll
have to take the output and work with it further.
I can post a description of the algorithm is there is enough
interest.
Ro
UUCP: { uunet | cmcl2 }!esquire!rreid
Internet: rreid@dpw.com -or- rlr@woof.columbia.edu
Post: 118 of 179
From: mcdonald@aries.scs.uiuc.edu (Doug McDonald)
Newsgroups: comp.music,rec.music.classical
Subject: Re: Fractal Music Generation (summary)
Keywords: music theory, composition, Meyer
Date: 13 Apr 90 13:50:58 GMT
Reply-To: mcdonald@aries.scs.uiuc.edu (Doug McDonald)
Organization: School of Chemical Sciences, Univ. of Illinois at Urbana-Champaign
Lines: 9
On a slightly different subject, but related - I have tried to write computer
programs that imitate the paintings of Jackson Pollock - and it is
very difficult. It is probably not impossible, but it would require
essentially coding in the exact style of any painting. I did produce
programs that make nice screen images, quite arty, but I never got close
to the real thing.
Doug McDonald
Post: 122 of 179
From: bdb@becker.UUCP (Bruce Becker)
Newsgroups: comp.music
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 13 Apr 90 16:40:28 GMT
Reply-To: bdb@becker.UUCP (Bruce Becker)
Organization: G. T. S., Toronto, Ontario
Lines: 31
s: article <1990Apr12.150201.12739@kth.se> d88-jwa@nada.kth.se (Jon W{tte) write
|In article <9613@sdcc6.ucsd.edu>, mu298ac@sdcc6.ucsd.edu (Philip
|Marlowe) writes:
|> In article <1990Apr9.151958.26859@ncsuvx.ncsu.edu>
|george@shumv1.ncsu.edu (George Browning) writes:
|[...]
|> Previous discussions in this group about fugues being
|> "self-similar" shows a lack of understanding about just what a fugue
|> is. Just because something is repeated at the same level, it doesn't
|
|Look at the mandelbrot set. It is self-similar, but skewed,
|rotated, mirrored and transformed in various ways. Actually, I
|think you could create reasonable fuge-LIKE music (actually, a whole
|new type) that was enjoyable using fractals.
I know some folks who actually did this.
They seem to have used Scho:nberg's "Principles
of Harmony" (I might have the name wrong) to
translate fractal states into MIDI outputs.
I don't know how they interpreted the text
to produce the results, but it was reasonably
musical, but not particularly melodic.
As the fractal was being generated on an
Amiga, the music would change according to the
part of the M set and depth of recursion...
--
,u, Bruce Becker Toronto, Ontario
a /i/ Internet: bdb@becker.UUCP, bruce@gpu.utcs.toronto.edu
`\o\-e UUCP: ...!uunet!mnetor!becker!bdb
_< /_ "Free your ass and your mind will follow" - Punkadelic
Post: 128 of 179
From: carroll@bcsaic.UUCP (Jeff Carroll)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 13 Apr 90 06:20:07 GMT
Reply-To: carroll@bcsaic.UUCP (Jeff Carroll)
Organization: Boeing Computer Services AI Center, Seattle
Lines: 39
e Browning) writes:9.151958.26859@ncsuvx.ncsu.edu> george@shumv1.ncsu.edu (Georg
>
> I have an article from the book Fundamental Algorithms for
>Computer Graphics written by Richard F. Voss that talks about fractal
>music. Voss says "One of my exciting discoveries was that almost all
>musical melodies also mimic 1/f noise." He gives some pictures and
>examples, including a couple of "spectral density measurements of the
>pitch variations in various types of music showing their common
>correlations as 1/f noise" These graphs show such things as Medieval
>music up to 1300, Beethoven's 3rd Symphony and the Beatles Sgt. Pepper.
Nothing particularly exciting, profound, or metaphysical about that. The
reason that spectral density of music seems to vary as 1/f is that
musical scales are logarithmic in frequency - that is, musical pitch intervals
as measured by a musician are proportional to the logarithm of the
corresponding frequency intervals measured by an engineer.
This, incidentally, is why the slide switches on a
graphic equalizer cover wider bands as frequency gets higher.
The pitches, or "musical states", get farther apart in frequency as
frequency goes up. Therefore the spectral density varies roughly as 1/f.
It's analogous to the relationship between decibels and watts, or
between the Richter scale and the displacement (in inches) of the ground
under your feet.
It's clear that neither medieval chants, nor Beethoven, nor Sergeant
Pepper can be strictly described as "1/f noise". It is not the "1/f"ness
of this music that makes it interesting, great, or even musical. Once
you've built a real 1/f noise generator you'll realize that.
It might be more interesting to know whether the spectral density of
music along the *pitch* axis is normally distributed (approximately, of
course), or skewed one way or the other. In such a case, the spectral
(frequency) density of the music could be said to fit a log-normal
distribution.
Jeff Carroll
carroll@atc.boeing.com
Post: 144 of 179
From: rreid@esquire.UUCP ( r l reid )
Newsgroups: comp.music
Subject: Re: 1/f
Date: 17 Apr 90 13:29:21 GMT
Reply-To: rreid@esquire.UUCP ( r l reid )
Organization: ? !
Lines: 186
When I said recently that I'd be willing to post a description of and
some code using the Voss algorithm, I got deluged with requests.
Some hungry folks out there!
If you missed that thread, there was some discusssion about 1/f
noise as "representing" a lot of music. The fallacy in there
is that while you may be able to take a lot of music and
find that kind of distribution in it, it doesn't follow
that pink noise is sufficient to generate music. In fact,
it isn't. But if you appraoch it as a toy, you can have
some fun with it.
Disclaimer on the code - since this is a toy, the code isn't
as generalized as it could be. And of course you use it
at your own risk.
First, a basic description of What It Is. My note generating program
is called pink, and I wrote this about it:
pink generates to stdout a list of octave point pitch class numbers
as newline-null terminated charater strings (good for sending to
pipes). The numbers are generated using the Voss argorithm
as described by John Simonton in 1970. Our control is to choose
the 16 candidate notes (these are kept in a file), and to a lesser
extent the seed for the random number generator (curerntly hard coded
as 23). We have 5 dice with four sides. We have a five bit counter.
We increment the five bit counter. Any dice whose corresponding bit has
changed as a result of the incrementation gets re-rolled. The
values of the 5 dice are added. This is the index into the
candidate list. As in craps, the middle of the list is favored.
(In fact I have yet to roll all zeros or all threes, so I have yet to use
the very ends of my list). This is derived from the idea of pink
noise, hence the name.
(The reference to Simonton is because I modeled my program after
the one he wrote in 6502 assembler for the PAiA. Unfortunatly this
gave my code a rather assembler-like flavor - bitmasks? Barf!)
As you can see, the idea is pretty simple - you're just favoring certain
notes. Now, my program is generating 8ve.pchclass pitches -
you can use MIDI notes if that's what you need, or you could
use this to select time values, or timbres, or anything else
you might want. What's also obvious to me is that
the choice of candidate notes will very much influence
how pleasent the result is.
Of course, you will need to plug these pitches (or whatever) into
notecards (or whatever) with other information. And if you
want to make Music (capital M), you will not be able to use
this for much more than a jumping off point.
You give this
program 2 arguments: the filename of the candidate list, and the
number of notes you want to generate.
Here's my code:
#define NUMDICE 5
#define DICESIDES 4
int dice_reg[2];
int dice[NUMDICE];
int dice_side[DICESIDES];
int numnotes;
static int dice_vals[DICESIDES] = { 0, 1, 2, 3, };
static char candidates[8][16];
#define SEED 23
#define STATESIZE 256
char state[STATESIZE];
static int mask[NUMDICE] = { 01, 02, 04, 08, 010, };
#include <stdio.h>
#include <math.h>
main(ac, av)
int ac;
char *av[];
{
int i, j, k;
char *c, *index();
FILE *fp;
if(ac != 3)
{
av[0]); fprintf(stderr, "usage: %s 16notefile number_of_notes_to_gen\n",
exit(-1);
}
if((numnotes = atoi(av[2])) < 1)
{
fprintf(stderr, "%s is a strange number of notes\n", av[2]);
av[0]); fprintf(stderr, "usage: %s 16notefile number_of_notes_to_gen\n",
exit(-1);
}
if((fp = fopen(av[1], "r")) == NULL)
{
fprintf(stderr, "Can't open %s\n", av[1]);
av[0]); fprintf(stderr, "usage: %s 16notefile number_of_notes_to_gen\n",
perror("fopen");
exit(-1);
}
for(i = 0; i < 16; i++)
{
if(fgets(candidates[i], 8, fp) == NULL)
{
fprintf(stderr, "input needs 16, not %d notes\n", i);
_gen\n", av[0]); fprintf(stderr, "usage: %s 16notefile number_of_notes_to
exit(-1);
}
if(c = index(candidates[i], '\n'))
*c = '\0';
}
for(j = 0; j < 2; j++)
dice_reg[j] = 0;
/* done initializing dice registers to 0 */
for(i = 0; i < NUMDICE; i++)
dice[i] = random()%DICESIDES;
/* done initializing dice to random */
initstate(SEED, state, STATESIZE);
for(k = 0; k < numnotes; k++)
{
dice_reg[1] = dice_reg[0];
dice_reg[0]++;
for(i = 0; i < NUMDICE; i++)
/* check the low DICENUM bits one at a time */
/* do this by ORing with the right mask */
/* then if the XOR is true, the bits examined*/
/* are different and we need to roll the die */
if((mask[i] & dice_reg[0]) != (mask[i] & dice_reg[1]))
{
#ifdef TEST
fprintf(stderr,"rerolling %d\n", i);
#endif
dice[i] = random()%DICESIDES;
}
/* add up the dice and this is the index into the cand */
#ifdef TEST
fprintf(stderr, "%d %d %d %d %d = ",
dice[0], dice[1], dice[2], dice[3], dice[4]);
#endif
for(j = i = 0; i < NUMDICE; i++)
{
j += dice[i];
}
#ifdef TEST
fprintf(stderr, "%d\n", j);
#endif
printf("%s\n", candidates[j]);
}
}
Here was my candidate list for a nice background stream of notes:
7.00
7.04
7.07
8.00
8.04
8.07
9.00
8.07
8.09
8.05
8.02
7.02
7.05
7.02
7.05
7.09
Have fun. Naturally, this code is supplied without any warrantee
of any kind.
Ro
UUCP: { uunet | cmcl2 }!esquire!rreid
Internet: rreid@dpw.com -or- rlr@woof.columbia.edu
Post: 147 of 179
From: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Newsgroups: comp.music
earch Digest Vol. 5, #36)Handelman quoting Laske quoting Bel (was: Re: Music-Res
Date: 18 Apr 90 05:01:33 GMT
Reply-To: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Organization: Princeton University, NJ
Lines: 78
In article <12935@venera.isi.edu> smoliar@vaxa.isi.edu (Stephen Smoliar) writes:
;In article <15369@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot
;Handelman) writes:
;> If it
;>happens that something tolerably interesting is produced through a random
;>model -- interesting by comparison with standards set by other non-randomaly
;>determined music, early serial, for example -- then one would have to
;>conclude, assuming some sort of homomorphism between the compositional and
;>the cognitive/perceptual, that the random generation model is telling you
;>something about the perception of the world, that perception may have a factor
;>of randomness built into it. Now this conclusion seems absurb and unwarranted,
;>but it parallels, I think, equivalent statements concerning the potential
;>relevance of NN-based music.
;Actually, I would not be so quick to write this approach off as absurd. As I
;suggested in a recent article about using fractals, experimenting with random
;methods, if properly done, may tell us some valuable things about the
;GRANULARITY at which decisions of composition are made. Mozart's dice
;composer "works" (and there are any number of levels at which it DOES
;work) by virtue of invariants among alternatives whose choice is left
;to chance. You can do a Roman numeral analysis which will remain the
;same no matter what the dice choose for you. (I've done this.) You
;can probably also derive a Schenker middle-ground from which you can
;then derive all the different possible foregrounds which the dice may
;provide. (I haven't done that one.) To go back to that old Ligeti analysis
;of Boulez, the study of random processes may tell us when enough decisions have
;been made that it just doesn't matter how the remaining ones come out; and
;that seems like pretty valuable information about both composition and
;perception.
I think there's a false teleology somewhere in here. I can make a not
so bad sounding piece for tom-toms by flipping coins, assuming the piece
doesn't go on too long, and I'm using a constant pulse. Typical sequences
will always have a structural richness unconstrained by the composition
rule, and just so that we know what we're talking about here's an example
of a randomly generated tom-tom piece:
ABBAAABBABAABAAABAABABBAAAAAAAABAAAAABBAAAAABAABABBBAABAAABAABAAABABAABB
I'm not saying this is good or bad, only that this can serve as the basis
for an extended form and analysis seminar that will have nothing to do
with the way in which it was generated. Compare, for example, with part of
the (I should say "a") tom-tom part in Stockhausen's early work Kreuzspiel,
consisting of sequence ABAABAAABAAAABAAAAAB etc. That seems to be a subtext
of my randomly generated sequence. We -- that is, as listeners (or really
here as analysis hacks who learned to talk about music from Allen Forte)
could say things like "the repetition of BB at unit 7 casts a reminder
back to the opening," that the opening begins with the symmetrical
statement of ABB and then enlarges on that symmetry with the pattern AB
then AAB then AAAB which is then folded back on itself through AAB then
AB, the latter cell being pivotally exploited for reintroducing
the opening motive, ABB: the promise of densely composed out material is
then gradually withdrawn by suspending the B tom-tom for eight units, at
which point the listener, on presentation of the B anew followed by
the 5fold inistence on A again (thus introducing the fibonacci series
as a hypothesis of compositional determinism) realizes that the inversion
around the motivic axis AAAB was responsible for introducing the develpmental
idea of diminution, and one could easily continue in this manner.
I'm only claiming for the purpose of argument that this analysis has
something to do with how a resourceful listener, not easily put to
sleep, MIGHT hear this kind of thing -- might, because other analyses can
be constructed. Each of these analyses makes assumptions on the part of
the listener, and none of these assumptions can be traced back to the
composition rule.
Now as to the Mozart example: your concept of "works" is that you get
a conventional set of harmonies out of it. You're not suggesting
that was random (if so, look at the Thomas Atwood notebooks). You say that
music works: I say that music stinks. "Works," in my concept of Mozart,
is something like the slow movement of the f major quartet, and I don't
see a mapping from the technique of dice composition to the technique of
that movement.
As to the grain of decision making in composition: it's whatever you
want it to be. We live in the free world.
I stick to my guns: no inference from autocomposition, including
the formulaic, to perception.
Post: 151 of 179
From: aipdc@castle.ed.ac.uk (Paul D. Crowley)
Newsgroups: comp.music
Subject: Re: "Only Amateurs" Re: Music-Research Digest Vol. 5, #34
Date: 18 Apr 90 16:48:59 GMT
Reply-To: aipdc@castle.ed.ac.uk (Paul D. Crowley)
Organization: Edinburgh University Computing Service
Lines: 6
I'd agree that trying to generate music algorithmically is usually a
pretty poor thing for a musician to be doing - but surely it's a
perfectly good thing for an AI dept. to be doing?
-
\/ o\ "I say we grease this rat-fuck son-of-a-bitch Paul D Crowley
/\__/ right now. No offense." - Aliens. aipdc@uk.ac.ed.castle
Post: 158 of 179
From: andyn@stpstn.UUCP (Andy Novobilski)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Keywords: the 3 M's: math, music, midi
Date: 19 Apr 90 20:18:03 GMT
Reply-To: andyn@stepstone.com (Andy Novobilski)
Organization: The Stepstone Corporation, Sandy Hook, CT 06482
Lines: 19
Somewhere in the 1984-87 time frame, there was an article published
in the proceedings of USENIX (or some UNIX conference) by a research
team at AT&T on the topic of Binary Stocastic Subdivision as an algorithm
for generating music. Included in the article was a number that you
could call to hear a demonstration of the algorithm played on a set of
MIDI controlled instruments.
I know the information is sketchy, but a little time at a technical
library should yield the reference. If anyone is interested and can't
locate the paper in a local library, I'd be happy to try and find it
at home.
Best of luck,
Andy
--
Andy Novobilski | The Stepstone Corp. | The expressed views have been
andyn@stepstone.com | 75 Glen Rd. | approved by a committee of three:
(203)426-1875 | Sandy Hook, CT 06482 | the goldfish, blackfish, and me.
Post: 159 of 179
From: smoliar@vaxa.isi.edu (Stephen Smoliar)
Newsgroups: comp.music
earch Digest Vol. 5, #36)Handelman quoting Laske quoting Bel (was: Re: Music-Res
Summary: teleology at ten paces
Date: 20 Apr 90 00:31:03 GMT
Reply-To: smoliar@vaxa.isi.edu (Stephen Smoliar)
Organization: USC-Information Sciences Institute
Lines: 74
In article <15439@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot
Handelman) writes:
>In article <12935@venera.isi.edu> smoliar@vaxa.isi.edu (Stephen Smoliar)
>writes:
>;In article <15369@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot
>;Handelman) writes:
>;> If it
>;>happens that something tolerably interesting is produced through a random
>;>model -- interesting by comparison with standards set by other non-randomaly
>;>determined music, early serial, for example -- then one would have to
>;>conclude, assuming some sort of homomorphism between the compositional and
>;>the cognitive/perceptual, that the random generation model is telling you
>;>something about the perception of the world, that perception may have a
>;>factor
>;>of randomness built into it. Now this conclusion seems absurb and
>;>unwarranted,
>;>but it parallels, I think, equivalent statements concerning the potential
>;>relevance of NN-based music.
>
>;Actually, I would not be so quick to write this approach off as absurd. As I
>;suggested in a recent article about using fractals, experimenting with random
>;methods, if properly done, may tell us some valuable things about the
>;GRANULARITY at which decisions of composition are made. Mozart's dice
>;composer "works" (and there are any number of levels at which it DOES
>;work) by virtue of invariants among alternatives whose choice is left
>;to chance. You can do a Roman numeral analysis which will remain the
>;same no matter what the dice choose for you. (I've done this.) You
>;can probably also derive a Schenker middle-ground from which you can
>;then derive all the different possible foregrounds which the dice may
>;provide. (I haven't done that one.) To go back to that old Ligeti analysis
>;of Boulez, the study of random processes may tell us when enough decisions
>;have
>;been made that it just doesn't matter how the remaining ones come out; and
>;that seems like pretty valuable information about both composition and
>;perception.
>
>I think there's a false teleology somewhere in here.
If that's your way of saying that we are coming at this with different
objectives, then you're right. I am not interested in whether or not
"I can make a not so bad sounding piece" by any random process. I AM
interested in those random processes which, at some (probably ill-defined)
level, do not SOUND like random processes. (Note that this is a question
of subjective human perception, as opposed to the exegetic skills of your
mathematically-trained theorist.) Such processes probably tell us more about
how one listens than they do about how one composes; but that does not make
them the less interesting (at least in my book, which you are free to return
to the library as long as you have not damaged the spine). (I would argue,
by the way, in response to a remark by Paul Crowley, that the sort of inquiry
I have posed is a "perfectly good thing for an AI dept. to be doing." I am
not so sure about the sort of composition objectives you have in mind.
Composing is a perfectly good thing for composers to do, and I would just
as soon leave it there. If they wish to draw upon the resources of artificial
intelligence, they are as free to do so as if they wish to draw inspiration
from architecture. About the only thing I can think of which would concern
an AI department would be a POST HOC analysis of what such a composer had done
with those resources.)
=========================================================================
USPS: Stephen Smoliar
USC Information Sciences Institute
4676 Admiralty Way Suite 1001
Marina del Rey, California 90292-6695
Internet: smoliar@vaxa.isi.edu
"By long custom, social discouse in Cambridge in intended to impart and only
rarely to obtain information. People talk; it is not expected that anyone
will listen. A respectful show of attention is all that is required until
the listener takes over in his or her turn. No one has ever been known to
repeat what he or she has heard at a party or other social gathering."
John Kenneth Galbraith
A TENURED PROFESSOR
Post: 161 of 179
From: mo@flash.bellcore.com (Michael O'Dell)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Keywords: the 3 M's: math, music, midi
Date: 20 Apr 90 02:24:12 GMT
Reply-To: mo@flash.UUCP (Michael O'Dell)
Organization: The Center for Virtual Reality
Lines: 8
Sorry, folks, it tweren't AT&T, but Bellcore's own Peter Langston
-Mike O'Dell
-Mike O'Dell
"I can barely speak for myself, much less anyone else!"
----------------------------------------
The Center for Virtual Reality --
"Solving yesterday's problems tomorrow!"
Post: 163 of 179
From: mcnamara@vixvax.mgi.com
Newsgroups: comp.music,alt.fractals
Subject: Re: Fractal Music Generation (summary)
Date: 14 Apr 90 03:07:20 GMT
Organization: Management Graphics, Inc.
Lines: 43
In article <562@bilver.UUCP>, alex@bilver.UUCP (Alex Matulich) writes:
> Several weeks ago I posted an a plea for help in comp.music and
> comp.sources.wanted for an algorithm to generate fractal music. I lost the
> original text of my posting, but the gist of it was this:
>
> A fugue is a piece of music rich in self-similar structure. J. S. Bach, a
> master at writing fugues, was able to maintain up to six instrumental parts
> playing a short theme in different ways -- at different pitches, different
> speeds, inverted, upside-down, backwards, and so on -- and it all fit
> together too!
>
> Fractals also are rich in self-similar structure. By definition, after all,
> a fractal IS a self-similar object. The parallels between fractals and
> fugues seem so close, I thought, that maybe a MUSICAL fractal generator
> could be developed as an aid in writing fugues.
>
> I tried an experiment based on the generation of a Koch curve, assigning
> a relationship between note pitch and line angle, and another relationship
> between note duration and line length. My experimented generated a
> sequence of notes that sounded interesting. The problem is that it was
> a single monotonic sequence. How can a fractal music generator be made
> to create overlapping sequences of notes which have harmonically correct
> relations to each other?
>
In 1988 or thereabouts Charles Dodge (_Earths' Magnetic Field_) came
to Mpls. to lecture about computer music. He brought with him a tape of
several pieces of music, one generated using fractal relationships between
the parts of the composition. As I recall, he generated an initial fractal
sequence, and then used fractal relations to generate the other parts from
the original one.
The music was interesting. Sort of like 101 Strings does Phillip
Glass. As he put it: "This is the first computer music I've heard which
sounds like bad music(previous attempts didn't sound like music at all)."
There were several other interesting pieces on the tape. The best
one was by Curtis Braun, titled _Brontosaurus_. It was a child's poem, read
by a computer voice synthesis program, and then modified by the composer into
a sort of self-similar composition.
I think he would send you the tape, and/or provide details of his
algorithms. He is at the Brooklyn College Center for Computer Music.
Phone (718) 780-5582.
Curt McNamara
mcnamara@mgi.com
Post: 165 of 179
From: scott@bbxsda.UUCP (Scott Amspoker)
Newsgroups: comp.music
Subject: Re: "Only Amateurs" Re: Music-Research Digest Vol. 5, #34
Date: 20 Apr 90 15:47:22 GMT
Reply-To: scott@bbxsda.UUCP (Scott Amspoker)
Organization: Basis International, Albuquerque, NM
Lines: 15
:n article <3370@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul D. Crowley) writes
>I'd agree that trying to generate music algorithmically is usually a
>pretty poor thing for a musician to be doing - but surely it's a
>perfectly good thing for an AI dept. to be doing?
I don't 100% agree with this. I have never worked with algorithmic
composers but it seems to me that they could help "suggest" ideas
that a human composer could work with. (Sometimes it takes that
leap of logic to break out of a rut.)
--
Scott Amspoker
Basis International, Albuquerque, NM
(505) 345-5232
unmvax.cs.unm.edu!bbx!bbxsda!scottPost: 166 of 179
From: maverick@fir.berkeley.edu (Vance Maverick)
Newsgroups: comp.music
Subject: Algorithmic composing tools
Date: 20 Apr 90 19:43:13 GMT
Reply-To: maverick@fir.berkeley.edu (Vance Maverick)
Lines: 11
In article <695@bbxsda.UUCP>, scott@bbxsda.UUCP (Scott Amspoker) writes:
> I have never worked with algorithmic
> composers but it seems to me that they could help "suggest" ideas
> that a human composer could work with. (Sometimes it takes that
> leap of logic to break out of a rut.)
Or that leap of illogic. A "random" program doesn't know anything about
your rut, even if you think you've constrained it to do the kinds of
things you're thinking of. I find these tools useful for generating raw
material I could not have invented myself, which I can then polish at my
leisure.
Post: 167 of 179
From: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Newsgroups: comp.music
earch Digest Vol. 5, #36)Handelman quoting Laske quoting Bel (was: Re: Music-Res
Date: 20 Apr 90 22:41:19 GMT
Reply-To: eliot@phoenix.Princeton.EDU (Eliot Handelman)
Organization: Princeton University, NJ
Lines: 17
In article <12988@venera.isi.edu> smoliar@vaxa.isi.edu (Stephen Smoliar) writes:
;In article <15439@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot
;Handelman) writes:
[no inference from music automatons to perception etc.]
; I am not interested in whether or not
;"I can make a not so bad sounding piece" by any random process. I AM
;interested in those random processes which, at some (probably ill-defined)
;level, do not SOUND like random processes.
But Steve, when you come right down to it, what DOES sound like a random
process? My students complained that Pierrot Lunaire, of all things,
sounded like random processes to them -- obviously that changes with a
wee bit of effort. Whereas with over 20 years moderne Musik behind me, not even
the Williams Mix sounds like a random process to me. It's already carrying a
hell of a lot of history.
Post: 168 of 179
From: latta@sting.Berkeley.EDU (Craig R. Latta)
Newsgroups: comp.music
Subject: Re: Algorithmic composing tools
Keywords: Reality is a better source
Date: 21 Apr 90 00:17:35 GMT
Reply-To: latta@sting.Berkeley.EDU (Craig R. Latta)
Organization: Music Department, UC Berkeley
Lines: 30
X-Local-Date: 20 Apr 90 17:17:35 PDT
In article <24264@pasteur.Berkeley.EDU>, maverick@fir.berkeley.edu
(Vance Maverick) writes:
|> In article <695@bbxsda.UUCP>, scott@bbxsda.UUCP (Scott Amspoker) writes:
|> > I have never worked with algorithmic
|> > composers but it seems to me that they could help "suggest" ideas
|> > that a human composer could work with. (Sometimes it takes that
|> > leap of logic to break out of a rut.)
|>
|> Or that leap of illogic. A "random" program doesn't know anything about
|> your rut, even if you think you've constrained it to do the kinds of
|> things you're thinking of. I find these tools useful for generating raw
|> material I could not have invented myself, which I can then polish at my
|> leisure.
|>
I find sources from Real Life (people talking, cats meowing,
jackhammers jacking, etc.) much more useful than some machine
throwing out noises with <insert parameter here>s generated from
<insert method here>.
For me, there is a point when the use of automata in
creating emotion is silly.
-----------------------------------------------------------------------------
Craig Latta "Those who know History are
latta@swindle.Berkeley.EDU doomed to explain it" -- me.
-----------------------------------------------------------------------------
Post: 171 of 179
From: pauld@hpausla.aso.hp.com (Paul Doornbusch)
Newsgroups: comp.music
Subject: Re: "Only Amateurs" Re: Music-Research Digest Vol. 5, #34
Date: 21 Apr 90 01:33:31 GMT
Organization: HP Australian Software Operation
Lines: 16
Well I'm a composer who has been writing for several years. I've been
through many music courses at various universities throughout the
world, and I work with many successful and famous composers.
I am interested in algorithmic composition (and I've never heard a
colleague criticize it) because I'm aware that I use an algorithm when
I write, formalizing that would be advantageous in understanding what
I'm doing. Once the algorithm is formalized it may then be used with
data to produce a composition. Advanced pieces are becoming
structurally more complex, a strong structure as defined by the
algorithm will allow the composer to produce such a piece. Because of
the mass of data involved the process can take extraordinary amounts
of time, computers are the key here as they allow the data to be
manipulated more easily and quickly.
I hope that this explains more of what it's all about.
Post: 172 of 179
From: smasters@gmuvax2.gmu.edu (Shawn Masters)
Newsgroups: comp.music,alt.fractals,comp.sources.wanted
Subject: Re: Fractal Music Generation (summary)
Summary: It was in Science News
Keywords: the 3 M's: math, music, midi
Date: 21 Apr 90 15:53:36 GMT
Reply-To: smasters@gmuvax2.UUCP (Shawn Masters)
Followup-To: comp.music
Organization: George Mason Univ. Fairfax, Va.
Lines: 9
I saw an article a number of years back about something like that. It
was in Science News, and was talking about AI algorithm design this one
team of researchers was doing. Not only did calling this number just
play music, I seem to remeber that it was semi-interactive, and they
wanted the general public to test it. In the end the reponse was to
great, so they shut down the line.
smasters@gmuvax2
smasters@gmuvax
Post: 173 of 179
From: pauld@hpausla.aso.hp.com (Paul Doornbusch)
Newsgroups: comp.music
Subject: Re: Algorithmic composing tools
Date: 22 Apr 90 05:48:18 GMT
Organization: HP Australian Software Operation
Lines: 32
21, 1990 /omp.music / latta@sting.Berkeley.EDU (Craig R. Latta) / 10:17 am Apr
In article <24264@pasteur.Berkeley.EDU>, maverick@fir.berkeley.edu
(Vance Maverick) writes:
|> In article <695@bbxsda.UUCP>, scott@bbxsda.UUCP (Scott Amspoker) writes:
|> > I have never worked with algorithmic
|> > composers but it seems to me that they could help "suggest" ideas
|> > that a human composer could work with. (Sometimes it takes that
|> > leap of logic to break out of a rut.)
|>
|> Or that leap of illogic. A "random" program doesn't know anything about
|> your rut, even if you think you've constrained it to do the kinds of
|> things you're thinking of. I find these tools useful for generating raw
|> material I could not have invented myself, which I can then polish at my
|> leisure.
|>
I find sources from Real Life (people talking, cats meowing,
jackhammers jacking, etc.) much more useful than some machine
throwing out noises with <insert parameter here>s generated from
<insert method here>.
For me, there is a point when the use of automata in
creating emotion is silly.
-----------------------------------------------------------------------------
Craig Latta "Those who know History are
latta@swindle.Berkeley.EDU doomed to explain it" -- me.
-----------------------------------------------------------------------------
----------
Post: 174 of 179
From: pauld@hpausla.aso.hp.com (Paul Doornbusch)
Newsgroups: comp.music
Subject: Re: Algorithmic composing tools
Date: 22 Apr 90 06:00:14 GMT
Organization: HP Australian Software Operation
Lines: 38
[ notes went berserk and posted the previous reply before I'd started
it, apologies to all]
> I find sources from Real Life (people talking, cats meowing,
> jackhammers jacking, etc.) much more useful than some machine
> throwing out noises with <insert parameter here>s generated from
> <insert method here>.
The machine may be throwing out pitches and rhythms that are
permutations of your previous work. Naturally occuring noises are
very usefull if that is what you want to use as a basis for a piece ,a
la Bartok or Kodaly. As a method for (you imply "inspiration") it is
no better or worse than any other method, and substansially less
controllable or dare I say random than some others.
> For me, there is a point when the use of automata in
> creating emotion is silly.
At the risk of starting another notes network storm, what has emotion
got to do with music (I will cite Stravinski, Schoenberg, and Boulez
as references, there are many more)? I write music for other people
to enjoy, whether they have an emotional reaction to a piece is not
under my control, and enjoyment may take many forms. There is no
doubt in my mind that all music is enjoyable in some way to some
people, and equally non-ejnoyable to others. Lyrics may invoke an
emotion as words are a widely understood method of communication, but
pitches and rhythms are extremely diverse both within and between
cultures, and invoke wildly varying responses from people depending on
many things: culture; education; solialization; age; personal history;
and so on. For example there are Indian ragas that invoke fear in
parts of India because people believe that they will cause all water
to burst into flame, I may find these interesting or enjoyable to
listen to, and others may find them repulsive.
The use of automata in creating music is as valid a method as any, and
more flexible than a lot of other methods. Using music to create
emotion is hazardous at best, writing music for people to enjoy
(define that as you wish) seems much more achievable.
Post: 175 of 179
From: maverick@fir.berkeley.edu (Vance Maverick)
Newsgroups: comp.music
Subject: Re: Algorithmic composing tools
Date: 22 Apr 90 18:41:55 GMT
Reply-To: maverick@fir.berkeley.edu (Vance Maverick)
Lines: 43
> > I find sources from Real Life (people talking, cats meowing,
> > jackhammers jacking, etc.) much more useful than some machine
> > throwing out noises with <insert parameter here>s generated from
> > <insert method here>.
>
> The machine may be throwing out pitches and rhythms that are
> permutations of your previous work. Naturally occuring noises are
> very usefull if that is what you want to use as a basis for a piece ,a
> la Bartok or Kodaly. As a method for (you imply "inspiration") it is
> no better or worse than any other method, and substansially less
> controllable or dare I say random than some others.
And "a machine throwing out noises" with random "parameters" sounds like
a pretty good description of a jackhammer....
> The use of automata in creating music is as valid a method as any, and
> more flexible than a lot of other methods. Using music to create
> emotion is hazardous at best, writing music for people to enjoy
> (define that as you wish) seems much more achievable.
I think you're in trouble if you make other people's enjoyment your
single goal and criterion. After all, your access to your own
sensibility is much more direct than your access to others'. It's
satisfying if other people like your music, and vital if you live by
your music, but I think it's a side effect. My definition of what we're
up to in doing music is similar to Craig's, though "creating emotion"
sounds as Romantic as Clynes' seven gestures. Is the experience of
following Stravinsky through the concerto for piano and winds "emotion"?
Perhaps in the broadest definition.
Acoustical events have aesthetic properties, which we apprehend somehow
-- intuitively, mystically, analytically, whatever. We like some of
these, and seek to build acoustical events with properties we care
about. (Maybe we invent the properties of our own pieces, maybe we
refine them from what our environment gives us.) Craig feels the real
world is richer in aesthetic properties than computer-generated sounds;
sure, but once you've found properties you like in a set of parameters
for your computer, the next step is easy to take.
Paul Doornbusch's attitude towards algorithms (expressed in an earlier
posting) shows a confidence that the properties of the output inhere in
the input, which I don't share. Musical structure is a tool, not the
meaning of the music.
Post: 177 of 179
From: latta@sting.Berkeley.EDU (Craig R. Latta)
Newsgroups: comp.music
Subject: Re: Algorithmic composing tools
Date: 23 Apr 90 04:10:44 GMT
Reply-To: latta@sting.Berkeley.EDU (Craig R. Latta)
Organization: Music Department, UC Berkeley
Lines: 38
X-Local-Date: 22 Apr 90 21:10:44 PDT
Paul Doornbusch writes:
"At the risk of starting another notes network storm, what has emotion
got to do with music?"
You can either ignore music, or you can react to it: you can feel
some way about it (have an emotional response). This response can then
lead to others (i.e., a spurring of the intellect, to satisfy a desire to
"understand" what has been heard).
Vance Maverick writes:
"And 'a machine throwing out noises' with random 'parameters' sounds
like a pretty good description of a jackhammer...."
Yes, but do you have to spend several hundreds or thousands of
dollars, or become involved in a research project, to utilize that source?
My point is that the current tools of algorithmic composition are
too cumbersome, expensive, and inaccessible relative to the much richer
array of sources around us all the time, wherever we are.
Vance further writes:
"Musical structure is a tool, not the meaning of the music."
Right on.
-C
-----------------------------------------------------------------------------
Craig Latta "Those who know History are
latta@swindle.Berkeley.EDU doomed to explain it" -- me.
Post: 224 of 225
From: mgresham@artsnet.UUCP (Mark Gresham)
Newsgroups: comp.music
Subject: Re: Fractal Music Generation (summary)
Keywords: I need help
Date: 3 May 90 23:33:12 GMT
Reply-To: mgresham@artsnet.UUCP (Mark Gresham)
Organization: ARTSNET Atlanta, GA USA
Lines: 30
In article <23087@bcsaic.UUCP> carroll@bcsaic.UUCP (Jeff Carroll) writes:
>It might be more interesting to know whether the spectral density of
>music along the *pitch* axis is normally distributed (approximately, of
>course), or skewed one way or the other. In such a case, the spectral
>(frequency) density of the music could be said to fit a log-normal
>distribution.
Do you mean by that
1) how often pitches appear in a piece (like Beethoven's 3rd
Symphony) throughout the audible spectrum? In that case, are you
considering duration or just how many times a given pitch is
initiated?
2) or are you wondering how frequently a pitch occurs within
a given tune? If that, you might like to know that in a large
proportion of tunes from oral traditions and tunes made up by
children exhibit this order of frequency, in terms of the diatonic
scale (in the scale degrees numbered 1 to 8 for 'white keys' C to C:
more often ... less often
5 3 6 2 1 8 7 4
Cheers,
--Mark
========================================
Mark Gresham ARTSNET Norcross, GA, USA
E-mail: ...gatech!artsnet!mgresham
or: artsnet!mgresham@gatech.edu
========================================
