This a ongoing discussion, but almost surpassed. I
tried both ways and I find the direct transduction
of data into generative music as musically,
conceptually, esthetically speaking, limiting. There
are no practical reasons.
With samples, which I actually call "sound design",
I can control a sort of "pathos", or some
semiological issues... some codified, explicit
symbols for the public. For me, in these
installations, in terms of the public, the most
important is how the people read the trees, beyond
any technological matters. The direct transduction
from data into music is for me, as a composer-sound
designer, more a programming home-work than a
music-making process.
For
example, in our installation in Mexico, in a big
central patio with a fountain, I decided to turn the
fountain off and put some water sounds for the tree.
Then, the tree played the roll of the fountain in
terms of sonic environment in the building.
Conceptually I changed the roll of the tree, from a
moving, green plant, into a fountain. And I decided
in which way I wanted to relate wind-activity levels
with sound-design. In this point come a more
complex, more intriguing extrapolation process, not
a direct, transductive, rational one.
I
tried also with generative synthesis, but after some
time of listening how the tree generates sound, one
begin to think… “How I did it… how it works…the code
is ok?” And this is not, for me, a submersive
process related to art but to technology and
algorhythms. Is more or less the same case when you
read a novel and you think “This guy… how good he
is, he writes great, what a technique!” or…. You
just read the novel as your OWN life experience and
forget about who wrote it, even if the novel is not
technically perfect. Then the novel is an art work,
not a technical work. I am not suggesting that I
succeed in this, but it is the direction I am
looking for.
I
am not judging real-time synthesis, of course, but
for these installations, the samples give me more
tools to find myself totally lost and ask more and
more artistic questions.