Squeak MIDI and Mac Serial Ports

johnm at wdi.disney.com johnm at wdi.disney.com
Mon Feb 8 19:47:50 UTC 1999


Mark Guzdial <guzdial at cc.gatech.edu> wrote:
> ...
> - The tweaks that I posted last week allow Siren's MIDI Port (based on OMS)
> to be used as if it were one of Squeak's SimpleMIDIPorts with things like
> the ScorePlayerMorph.  Thus, you can get the nice features of Squeak's MIDI
> support (like playing a MIDI file from the FileList, seeing the piano roll
> scroll by, etc.) with the nice sound of QuickTime through OMS.

Using the new Mac VM (2.3b), you do not need either the Siren VM nor
the tweak's Mark posted last week.

> ...
> - Several of us are interested in and working at improving the performance
> of Squeak's internal MIDI synthesizer.

I'd love to see folks built some great-sounds using Squeak's
built-in synthesis! This approach has a number of benefits:

  1. It's truely portable--you can get the exact same sounds on
     everything that has a sound output port from a Mac to
     a palmtop device like to Compaq Itsy.

  2. It may enable avenues of expression that are simply not
     available through MIDI.

Incidentally, using a commerical sample library (Peter
Siedlaczek's Advanced Orchestra), we've been able to get
some spectacularly realistic orchestral sound. Unfortunately,
commercial sound libraries have very restrictive copyrights,
so we can't just post these sounds for you all to use.

If anyone knows of a free-ware sample library, or would be
interested in creating one, please speak up. It seems possible
that a large music school might have such a thing, but I
haven't yet found one in my web searches...

One other thing we're very interested in is physical-modeling
synthesis. Modern computers have enough horsepower to
do this in real time (at least a few voices), and the examples
I've heard have been quite compelling. One real challenge
here is in the area of control--to get a realistic sound, you need
to continuously vary a number of performance parameters such
as breath and lip pressure, tongue position, etc. You basically
need a model of the performer that maps between the score and
performance gestures. This area would make a great masters
project or even a Ph.D. disseration.

	-- John





More information about the Squeak-dev mailing list