Re: Virtual Concerto


Subject: Re: Virtual Concerto
From: Eliot Handelman (eliot@generation.net)
Date: Tue May 18 2004 - 19:38:53 EDT


Michael Rempel wrote:

>I think we may be dancing around the same issue from different perspectives.
>I have answered you below but this is my over all impression.
>
>

My impression is that I don't understand what it is you want to talk
about. You took
exception to my statement that pilfering and recombining databases was not,
to my mind, what AI in music might be all about. I proposed instead the
formulation
of conceptualizing models. I haven't asserted that there's some
particular method
that could work, and I suggested that imaginative approaches are
probably what's
best for the moment. I was reflecting in this my sense of the analytic
poverty of models
employed to date in any of the music retrieval research I'm aware of. In
the absence
of powerful methods for guessing at "what is going on the music" it's
premature to
speak of the utility of databases. As for values and such, of course the
whole idea
is to represent human-level conceptualization or listening in proxy. I
suggested that
here too that imaginative approximations or interpretations, perhaps
alien but
provocative, could be more interesting than something that aims for
general validity. As for composing something "interesting,"
and weighing that interest somehow, I think Harold Cohen's advice is
best taken
here -- model process, not outcome. The process here is the
composing/listening mind,
rather than some sort of specifable external results. Or rather, the
only result I care
to specify is that the music should have sense of human (or posthuman)
agency.

-- eliot



This archive was generated by hypermail 2b27 : Sat Dec 22 2007 - 01:46:01 EST