mnissa
[in fact, Nissa the Kitty]
mnissa is another in a series of algorithmic, semi-(or maybe
even not-so-semi-)ambient automatic music generating apps that
I do to mess around with different music programming languages --
mlooch,
jnissa
and
jlooch
are other examples of this.
jnissa
and
jlooch
were realized using
JSyn,
the java-based digital sound synthesis/sig-processing
language written by Phil Burk.
mlooch
was my first attempt to write a complete Macintosh app using
the snazzy music DSP/synthesis language
Max/MSP
from
Cycling74.
The
jlooch
page also has links to even older code doing similar kinds of things,
including a more chatty "Ancient History" than some of the
other pages. I should really try to consolidate these sometime...
The excuse for mnissa was to try learning how to get around
jitter,
the amazing! astounding! graphics-manipulation package from the
happy people at the aforementioned
Cycling74,
especially after seeing some of the eye-popping marvels that
Luke DuBois
showed off in our
classes
at the
Computer Music Center
and incorporated into his
doctoral dissertation.
Whoa, boy! I sure got carried away with
hypertext
links
in the
above
short
description.
Yikes!
I have to stop doing this!!!!!!!
Sound... Graphics... oh I am So Excited!
click
here
to download the standalone app.
click
here
to download the Max/MSP/jitter patch, if you want.
[NOTE: This only runs on OSX at present -- I don't have a valid jitter
install for OS9 on my home machine. I imagine that the max-patch
above could be made to run under OS9 without too much pain and agony.]
Negotiating the Intensely Complicated User Interface
There are buttons. Click on them. There are menu items, too.
The Ancient History of mnissa
As mentioned above, check out the
Ancient
History of jlooch for a little background information about the
"looching" programs in general.
mnissa makes use of several plucked-string and nonlinear
feedback objects available in the
PeRColate set of
Max/MSP objects. I also stole some algorithmic compositional ideas
from Brad Garton's (hey! that's me!)
early work in "musical style modelling" -- see articles like
Virtual Performance Modelling and
The "Blues-o-Matic" Real-time Interactive Performance Model
in 1992 and 1993
ICMC Proceedings. Of course, this was back in the
paleolithic computer-music era when "musical style modelling" meant
something more than diced-up Mozart...
Although the style-modelling concepts I used in mnissa are rather
crude and approximate, it is possible to get interesting results from
more fully-realized models (or at least I fool myself into thinking
this is true). Here are a couple of mp3 excerpts from pieces I did
using fairly robust models:
The above pieces were 'written' by LISP programs running
on my trusty old NeXT machine. They created data for various
CMIXimplementations
of an extended Karplus-Strong algorithm described by
Charles Sullivan
(published in Extending the Karplus-Strong Algorithm to Synthesize
Electric Guitar Timbres with Distortion and Feedback in the Fall, 1990
issue of the Computer Music Journal). Oh, life was fun
back then.
Those Semi-Silly Answers to Semi-Silly Questions
Q. Why is it called mnissa?
"Nissa" is the name of our cat. "Looch" (as in the earlier
Max/MSP app mlooch) was the name of our dog. Fair is fair.
Looch died awhile ago, but Nissa is still alive and meowing,
daddy-oh.
Q. Why the occasional references to old-fashioned, outdated
'hipster' lingo?
Like, I'm completely hip to this question. I can dig where you're
comin' from. Swingin' on the flip-side. Something about cats, dad.
Gonesville.
Q. Hey, these are identical to the questions and answers on the
jnissa web page!
Yeah, but this one isn't. Is it?
Q. On-the-fly rendering, real-time audio synthesis, do I
need a quad-processor 5 THz machine to run this app?
Actually, no. The jitter video processing
implementation is really good, and ya gotta
love hardware OpenGL. On my 800 Mhz
tiBook, I'm getting nominally 10 frames/second. I even tried it
on my older 400 Mhz iMac. The visuals were a little herky-jerky, but
the music was fine. The audio load is quite low -- physical models are
really efficient. If you want, you can try turning the graphics and/or
sound on and off to see how it goes. The audio should be ok on most
extant Macintoshes. I've set up MSP to run in 'overdrive' mode, and even
though this sounds vaguely
like something that GM would feature on the latest
gas-guzzling luxury SUV, it seems to make for
decently-scheduled sound.
Q. What is it with you and those annoying menu items?
Well, they all have deep meaning for me. I'm just that kind of guy.
Q. Tell me now, are there any known bugs?
No, and here's why: as soon as I discover or hear of any problems
in the code, I immediately begin hitting myself in the head with a 2x4
until I lose conscious recollection of the bug. That way I can
say, with complete honesty, that I don't know of anything wrong.
I believe I can sell this technique to the Bush administration,
but I forgot how to navigate the White House Web Page in order
to send Smilin' George W. e-mail about this innovative breakthrough
in auto-veracity. uh...
what was I saying again?
Contact
Please feel free to send me comments, etc.:
Brad Garton
(garton - at - columbia - dot - edu).
Every once in a great while I randomly answer a bunch of stacked-up
e-mail. Also be sure to visit the
Computer Music Center
home page. It's a pleasantly stimulating place!
From the JSyn guy (Phil Burk):
Random Cool Links