[PD] Flute synthesis

Frank Barknecht fbar at footils.org
Fri Apr 27 19:34:25 CEST 2007


Hallo,
Nicolas Montgermont hat gesagt: // Nicolas Montgermont wrote:

> Andy Farnell a écrit :
> > That's good. A solo or featured instrument might use an expensive
> > physical model, but cheaper methods for the backing instruments.
> > Maybe some kind of namespace could be started now [orc-flute],
> > [orc-cello] and so on. A coherent performance interface would
> > be a great start, something simple akin to MIDI {/ducks}
> >   
> I agree with the namespace, simple and clear. I also think that midi 
> notes combined with a few continuous parameters should be sufficient.

I thought so, too, however instead of midi notes I'm not in favour of
midi notes anymore and would strongly vote for frequency inputs for
all instruments. It's more flexible in the long run, and in the short
run will make alternate tunings easier.

> > Actually the structure isn't as important as the flexibility
> > and documentation. One should be able to query an instrument with
> > a [help( message and get a print out like
> >
> > /string/position/ 
> > /string/unitweight/
> > /bow/pressure/
> >
> > all with normalised scales in OSC style so you know what is addressable
> >   
> Definitely.

Note that you'll probably want to have instance-specific OSC targets
like "/guitar1/string/position". At least I found these immensly
useful in Memento/RRADical. You can still set all instances with a
wildcard like "/*/string/position". 

While we're at Memento: It automatically creates OSC-targets for
everything. Getting feedback about available targets is on the TODO
list. What will probably be done this weekend is a massive reduction
in the number of externals needed by Memento. My realistic goal is to
get down to just [pool] and [OSCroute], and later [pool] will be
factored out as well and become optional. Then a bit later on Memento
will marry SSSAD and they will have a child.

Ciao
-- 
 Frank Barknecht                 _ ______footils.org_ __goto10.org__




More information about the Pd-list mailing list