Re: info on voice architecture (was re: simple modelling)

[David O'Toole]

> Oh.  This all makes sense now.  All that happens is that new notes are
> assigned the first available channel, and the channel is marked as free
> when a note off is received ... right?


Almost. Your machine has an ADSR that will sometimes cause a note to
play out for a short time after the note off is played. It would suck if
Octal presumed you went to absolute silence upon note-off, because the
next note could cut off that pretty synth decay. Since only the machine
knows when it's really done with a channel, you have to let it know by
calling a teensy little wrapper function like pkg->channel_freed() with
the channel number. 

> I really should have gotten this earlier....  :/


Oh, don't worry :-) I think I made it more confusing by giving too much
implementation info about the host. 

@@@ david o'toole
@@@ address@hidden
@@@ www.gnu.org/software/octal