[Audacity-devel] OSC for Audacity

[Jean-Baptiste T. <jb...@dc...>]

Hi,

I have recently added OSC (Open Sound Control) to my personal version of 
Audacity, using Ross Bencina's oscpack library. I am using this so far 
for my own research, so that a Java application can communicate with 
Audacity and create tracks, add sounds and play these directly without 
any GUI control of Audacity.

The code is working well, but it would be great if an Audacity expert 
could have a look at it.

Let me explain first why I think this would be a great feature for a 
future release of Audacity, and what is left to do.

OSC is a protocol that is comparable in many ways to MIDI. It 
facilitates the exchange of data between programs. It consists of a UDP 
(or TCP) platform for message exchanges. Many music programs and 
hardware today have implemented OSC, but none of these are a digital 
audio workstation, as far as I know. The music programs that have OSC 
implemented include the real time applications Reaktor, Max/MSP, Pure 
Data or SuperCollider. Hardware devices include the Lemur, Arduino, and 
other sensor systems.

Adding the OSC feature to Audacity would allow to communicate with other 
programs. But communicate what? Audacity is not a real-time program, 
i.e. it is not meant to modify sound as you go. However, the 
representation of sounds within Audacity could be manipulated and 
controlled from outside, allowing e.g. the construction of musical 
pieces without the constraints of time-based representation. The 
constraints imposed by time-based representations have been described by 
many, e.g.  Roger Dannenberg, a great contributor to this very software 
(Dannenberg, ``Music Representation Issues, Techniques, and Systems,'' 
Computer Music Journal, 17(3) (Fall 1993), pp. 20-30.)

What I have done so far is to create a simple send/receive interface for 
Audacity. This works as a separate thread. It listens to the port 7000 
and writes on the port 7001.
I have implemented simple actions following the reception of messages. 
These include adding a mono sound at a given offset on a new track and 
create a bounce of the whole representation. For some reason, I had 
trouble to control the play/stop commands.

It would be necessary to add new features before this is part of a 
release though, such as adding a stereo sound, adding a sound at a given 
offset to an existing track (as opposed to a new track), play, pause, 
stop, and also add some controls over sound transformations. Changing 
the offset of an existing sound could be useful too, as well as dumping 
a text description of the whole window (i.e. offsets, tracks, and 
canonical pathnames for all the sounds).

I would be happy to provide the code I have developed so far, with the 
list of modifications, as well as a simple java program to test it. I 
can also host a forum and a repository if needed. I would be glad to do 
more work on this, but I'm aware that it will be difficult for me to 
develop bullet-proof C++ and I need a bit of support!

I would appreciate advices regarding the requirements for this to be 
part of a future release, as well as collaborations for the development 
of a more complete OSC interface for Audacity.

Thank you,
Jean-Baptiste Thiebaut
Ph.D. student
London, UK