[<<][staapl][>>][..]
Mon Oct 1 05:48:46 CEST 2007

structures for music

this is more of a tutorial pre. i saw aymeric was using the stack to
store sequences, which is not a good idea.. i see 2 other ways: flash
and ram. i kinda like the x / . approach for pattern synths. the trick
is to do multiple voices, so i really need some kind of multitasking.

say i have 3 patterns

: bd  o . . . o . . .  bd ;
: sn  . . . . o . . .  sn ;
: hh  o . o . o . o .  hh ;

what do o and . do ?

let's assume that recursion is not allowed in these patterns. what can
we hide in a single invocation? a simple trick is to use the
dictionary shadowing: the words could call some fixed word, which is
re-implemented later.

: instrument   do something ;

: bd   o . . o . .  bd2 ;
: bd2  . . o . . o  bd ;

we could have:

: o instrument yield ;
: . yield ;

hmm.. it's probably better to directly use names instead of this
name-capture thing.

if recursion is disallowed, it should be possible to store each thread
in a single byte, so a lot of threads are possible. in that case, an
explicit interpretation and automatic looping might be better, using
routing macros.






[Reply][About]
[<<][staapl][>>][..]