28 Jul
2005
28 Jul
'05
7:42 p.m.
On Friday 29 July 2005 07:09, David Collins wrote:
Greg Wilder wrote:
FWIW there is a simple approach and some samples locked up
the pat (gus) format here. The idea was to use the file
system as the database and put as much meta info into an
associated (simple) text file as possible. In this case,
using the 1st line as parameters in the auto built timidity
config file was part of a shell script excercise.
http://freepats.opensrc.org/freepats/Tone_000/000_Acoustic_Grand_Piano.pat
http://freepats.opensrc.org/freepats/Tone_000/000_Acoustic_Grand_Piano.txt
--markc
>But what I'm talking about is a bit
different. I'm
talking about a
community built and supported library of acoustic
multi-sampled
instruments. Various dynamics, articulations, mutes,
...
Alternatively - rather than specifying the
characteristics in the name of the file itself - you
could use a database to keep record of it's
characteristics. Specifying duration, dynamic,
articulation, pitch and perhaps other things (e.g.
sample rate, and mono/stereo) all in the name of the
file might become somewhat unwieldy? Instead, you
could use the same parameters as fields in a database,
and be under less constraints with the naming
convention. On the downside, this would only be of use
to people who actually _had_ access to the database.
Just thinking out loud! Your idea's really good
though.