[chuck] Audicle

Ge Wang gewang at CS.Princeton.EDU
Tue Aug 10 04:44:00 EDT 2004


>>> I wonder how you display the six-sided cube on a screen or do you 
>>> use 6 screens for performances?
>> The idea is to be able to fully engage one face at a time, but you 
>> can move to other
>> faces at will.
>
> so you can switch between the faces on your screen?

Yep.  There is a command line that you can bring up (and hide) from 
where you
can change to another face (using code or UI).  Also there are keyboard 
shortcuts.

>>> Are 3D interfaces (on 2D screens) expedient for computer music? (i'm 
>>> a little bit sceptical)
>> Some of the face will be 2D, others 3D, and also there is the 
>> tabula-rasa face where the
>> user can display anything (2D or 3D).  3D can be effective for a lot 
>> of things, but can be
>> bad if everything was conformed to be in 3D.  Hopefully having both 
>> the 3D and the 2D
>> will be an effective combination.
>
>> We shall all soon see.
>
> :-) I never did anything with OpenGL. Do mobility video cards in 
> recent notebooks support OpenGL in hardware? (ATI 9200, 9600, etc).

The vast majority of recent notebooks (in the last 3 years) support all 
the basic rendering
in hardware for things the Audicle needs to do.

>>> Did I understand it correctly, that there is one dsp-server that 
>>> renders  the audio in a networked setup with several 
>>> people/computers? does this mean that one can bog down the 
>>> dsp-server with a resource hungry patch?
>> The simplest setup is the one-server/listener, in which case 
>> misbehaving code can
>> hang the server.  However, the model can include a conductor (using 
>> the on-the-fly
>> programming / code as instrument idea), who can remove/fix bad 
>> shreds, and also the musicians/ChucKists can learn to play well 
>> together, which would also fit this model.
> >
>> Multi-server setup will also be possible, though we still haven't 
>> worked out the timing
>> semantics.
>
> it gets complicated, when you try to synchronize different sound cards 
> with own master clocks, but a ChucK rendering farm with a sound card 
> in one of the machines would be cool!

This is an intriguing idea - especially since we have a large research 
group here at
Princeton doing shared high performance computing with very low latency 
(Kai Li's
shared virtual memory, and the Princeton display wall - synchronized 24 
cpu).  This
could very well lend itself to this.  Also, we can use multiple hosts, 
and use network
to synchronize or to otherwise compensate for the different clocks on 
the sounds cards.

Best,
Ge!



More information about the chuck mailing list