On Thu, 2004-12-02 at 14:02 +0100, Dave Griffiths wrote:
> I haven't ever played non-livecoding :) but I seem to plan things out in a
> very broad sense, with ideas to follow during a performance - ie
Seems similar to the approach adrian and I developed out of necessity,
of making a pictorial score up to half an hour before a performance.
Even though they are at times nonsensical, we do at least have something
to point at when we're agreeing upon where we are and where we're going.
> I don't think I like the idea of memorising a script and just typing it out
> live (I don't know if that was what you were getting at alex) I like to think
> things could be a little more fluid than that...
Agreed - that isn't what I meant to suggest. However through practice
an improv musician might build up a set of progressions and expressions.
But where you're playing a progression that you've played before, you'll
do it differently, in reaction to what your co-performers are doing.
When I program to an idea, and then program to it again, the two results
can look and sound very different. For example one idea might be "tween
between a random and regular rhythm and back again." I might practice
various ways of fulfilling that objective but would never come up with
the same thing twice, and then hopefully in a live situation would be
able to come up with something that worked well with whatever else was
going on.
Where I find I'm typing in almost exactly the same thing over and over,
then I think that should be library code.
However, after saying all that, I am coming to the conclusion that "live
coding from scratch" is an extreme position, and one I've been finding
very difficult to live up to while producing music I'm happy with.
To start with, live coding from scratch was great, truly liberating.
But once I'd gone through all the obvious things that can be programmed
in two minutes or less, I found it quite limiting.
So right now I'm looking into developing well constructed but incomplete
starting points to improvise away from. Here's a live recording I made
last night for example:
http://yaxu.org/yaxu-20041201-test.mp3
I had an idea for a compromise between from-scratch coding and using
pre-prepared scripts, that I'm hoping to find time to research soon...
That is, while pre-preparing a script (in a livecoding environment of
course), the hacker records many 'snapshots' of the code while it's
being built. Then when it comes to performing the code, the performer
can move backwards and forwards between these snapshots, as well as
doing live edits.
Therefore the audience gets to hear the code grow as it was developed in
the studio, but sped up (or slowed down, or reversed). Development time
meets performance time.
Does that make sense?
I've been taking these snapshots of my livecoding for some time now
(using RCS), but haven't got around to putting in commands to jump
between snapshots...
alex
--
alex <alex_at_slab.org>
Received on Thu Dec 02 2004 - 13:58:38 GMT