[Octal-dev] a little milestone for OCTAL :-) and more to come

[Dave O'Toole]

I spent the night working on OCTAL (I have shifted my sleep schedule in
anticipation of my marathon NYC trip this weekend) and there is some new
stuff.

First, the signal engine has been given a healthy rough draft; the core
of OCTAL is now functional. I now have a "add machine, connect machines"
interface going, and I have connected two squaregen machines to a single
delay, and heard their output reverbed and mixed together. Harmony
sounds nice :-).  There is a routine "do_block" that traverses the
signal network and makes sure everyone gets done in the right order. I
am probably going to remove the recursion from this (offloading it to a
topological sort routine that lets me store the sort) but the
proof-of-concept is there. 

(please note I'm talking about a C function call interface to the parts,
not a GUI human interface :-). but this is coming real soon anyway)

Also, the mixing algorithm right now is temporary. It's not dynamic---
this is the "divide by the number of inputs" thing that wouldn't work
for professional applications. But it's just a placeholder until the
heavy-duty dynamic mixer is finished. 

I was also working on some of the first GTK+ interface elements; namely,
the signal network editor window that lets you drag and drop the little
machines and draw lines between them. This isn't done yet, but will be
much easier now that the engine interface is there.

As always, revisions will probably be needed. I am starting to get a
weird feeling about the interface to the idea of "signal buffer", or
rather, lack of an interface. :-) I'll probably fix this very soon. 


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