Dave Phillips:
Greetings:
I'm putting finishing touches to a very long chapter on software sound
synthesis languages in which I've profiled current versions of Csound,
RTcmix, and SuperCollider3. IMHO I'd still recommend Csound to a novice,
particularly to a programming novice. I prefer the more modern language
elements in RTcmix and SC3, but the one is heavily C-like (fine with me)
and the other is a Smalltalk derivative. Csound also wins in its
abundance of helper applications, e.g., FJenie, nGen, Common Music,
blue, Cecilia, Csound Blocks, and others. However, RTcmix will see a new
release Real Soon Now, and there may be some more GUI stuff included.
SC3 for Linux lacks the neat graphic elements of the OSX version, but
it's still quite an interesting environment. I should also note that
Csound and RTcmix have no special requirements WRT editors, whereas with
SC3 you definitely want to learn how to use emacs. (Note that both emacs
and vi have Csound editing modes available to them.)
I'm currently accessing the supercollider synth fromwithin pure data
using the python PD external. I'd say its a supergreat combination
where you get easy/clean programming with python, functional programming
with the guile pd-external, sound processing power with SC3, and
GUI with pd. Check out the supercollider module in the pure-data
CVS.
Finally I would urge the beginner to make a real
study of some other
language, i.e., C/C++, Java, whatever, along with learning a sound &
music programming language. That assumes the time for such study, but
consider it time well spent, you'll learn a lot by the inevitable
comparisons.
Hmm, I would rather reccomend common lisp/scheme or python than
c/c++ and java for music purposes. C are fine for making hardware drivers,
c++ can be fine if you need OO to C, and java is horrible in all possible
ways. My opinion. :)
--