From: Replicant00@aol.com
Date: Wed Dec 29 1999 - 10:57:40 MST
In a message dated 12/27/1999 9:42:09 AM Pacific Standard Time,
jamesr@best.com writes:
From an information theoretic standpoint, the fact that
computers can decompose musical styles so easily and so well is a rather
obvious indication that music is not only algorithmic, but that the
fundamental algorithms aren't particularly complex.
>>
OK from that standpoint yes, yawn, yes, you can *call* them algorithms, and
you would be technically correct, but that is not the sum total of the
experience of a human being involved in the process of composition. Nor
should it be a desired set of tools. That's like saying we talk in wave
forms, we should visualize sawtooth and sine waves when we call mom to say hi.
<< Algorithms do not imply an extremely boring composer. The *complexity* of
the algorithms used (and how different algorithms are used together) would
be a much better guage of interestingness. >>
Unless the person is starting from that mathematical standpoint, they are
not "thinking" of algorithmic calculations, but rather a set of heuristics,
based on the emotional response desired. We like to call them "licks" 'cause
they are tasty.
A computer, or a person, starting the whole thing from the math formula,
will, I stand by it, create boredom music a la Philip Glass (not that there's
anything wrong with Philip Glass).
If you want to be a stickler you can call my thoughts electrical impulses
too, but that misses the point when describing what a thought is to a human
being. You can call rocks molecules etc.
I can also say an animation I create is "highly made of pixels"but that
doesn't give much insight.
me:
> spiked with intersections where interesting things *sometimes*
> happen. Normal music composition doesn't often take this approach.
james:
>>Actually, you've defined all types of music as far as I am concerned.
Even works by the great masters have areas that are weak at best. You
don't need a computer to generate boring music. But then again, good
music is in the ear of the beholder. >>
me again:
Yes, but the "meaningless" factor is not as large. Music "contains"
algorithms, and "contains" much math too, but hopefully, in it's inception,
it is not "based" on thoughts about such things....
What is meaningless to the listener, and analyzed as "simply" algorithms to
you, may be cathartic expression to a composer? At least hopefully it plays
a role in our life that is less abstract than calculation.
James:
(rant begins)
>>Emotional content in music is *highly* algorithmic, and perishingly simple
at that. >>
The post you made is brilliant and too long to post back, and I agree with
you that musicians learn harmonic tricks that stimulate learned and
subconscious associations, which hit the ear in certain dissonance/resonnance
combinations to generate easful or uneaseful feelings and that these are
mathematical in *theory*. Anyone who has ever studied musical theory knows
that.
HOWEVER: regarding your claim of it being trivial to program a machine that
*creates* emotionally engaging music:
***** At this time - if your program consistis of a bunch of heuristics that
make "feelings" - then IMO: ***you*** are the composer, not the program.
You have the brain that knows what's what inside a human's head. Not the
machine.
When an AI - on it's own - is able to put "meaning" into it's simulations,
and express it's own set of "feelings", perhaps I will form a different
opinion.
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:06:13 MST