Hi all,

I agree with some of this.

> <Harlan>
> I agree with Phil (J) that concerns about "classification" are
> entirely unnecessary to the creative process. It's all about the
> ears - if it sounds good, do it.

<me>
I think that is aim of most people in this group, to produce something
that sounds good, or least interesting. Not as a mathematical proof.

<Harlan>
>
> The interesting part for me (aside from the listening), is figuring
> out what aspects of music are actually "fractal" in the same sense
> that we measure visual fractals. We take it for granted that we
> often "see" the way in which an image or object is self-similar.
> However, whether we see the fractal structure or not, it is
> mathematically demonstrable.

<me>
I think it would be a much easier chore with a graphic fractal. With
music, the waters are further muddied by the fact that in practice, a
lot of people do a good bit of editing and combining the output of
several things together these days. (Much to Shawn's chagrin ;-)






<Harlan>

>
> Assuming that music derived from a fractal is automatically fractal
> is something like taking a poem by Kipling, transliterating it using
> the Cyrillic alphabet and then presenting it as Russian poetry. The
> point here is simply that the desired outcome is by no means
> guaranteed.
>
> As I wrote earlier, it is possible to insure fractality by being
> careful with mapping. It is also possible that fractal structure
> exists without intent, but to be certain of this it is necessary to
> make some sort of measurement relative to pitch, interval, duration,
> dynamics, etc.
>

<me>
I would think the mapping important, or rather how you arrive at the
mapping. Graphically, you might iterate over sets of points in the
real/complex planes and paint a pixel a certain color depending upon how
many iterations of a complex equation occur before either an "escape
value" is reached or N iterations occur and you determine it will never
reach "escape velocity". Music-generating software would mostly follow
that same model I would think, though there are others. I think that
there was software at one time that produced notes at each iteration, so
not sure about that...

<Harlan>
> Often researchers try to link some concept of fractal dimension to
> genre or "likeability," though this is not my focus. Some use MIDI
> renditions while others use audio files. The common thread is that
> all such efforts require identifying specific attributes for
> measurement.

<Me>
Interesting. If they are looking for "likeability" does that mean that
they are analyzing the "fluff" on top-40 radio stations? Surely not.

<Harlan>
>
> AND... none of this matters if one simply wishes to make music!
>

<me>
Quite true! I am sure that your research is quite challenging, but what
if you used your acquired knowledge thus far on this subject to actually
write a fractal music composition tool that was 100% aligned to your
ideal of what fractal music should be like?


Phil J.


>