|
From: Richard A. S. <rs...@bi...> - 2002-11-06 21:01:25
|
On Wed, 6 Nov 2002 17:46:34 +0000, Steve Harris wrote: > > For real-time control DMIDI would be good as it allows hardware control > > and the syntax checking is minimal. This is where using XML instead of > > CC or SYSEX messages would be too heavy. > > Argh, yes! I think we were talking at cross purposes. I was just thinking > of the service discovery phase, obviously the control protocol wants to be > be binary nad lightweight. > > > I think it would be good if we could find a split point between what can > > be controlled by a GUI and what can be controlled by MIDI hardware, even > > though there's overlap to an extent. What kind of real-time GUI control are we talking about? Perhaps I'm showing my sampler ignorance here, but seems to me that control via MIDI and control via socket are 2 sperate entities in 2 seperate worlds with very different manners of operation. So why try to handle them with the same system? If you wanted to send the engine say something like DMIDI data then shouldn't that be on a seperate socket? I'm not sure that the one-socket-do-it-all approach makes any sense. Most of the GUI stuff is all patch uploading, layering control, asignment of channels, loop points, graph setup, etc, all non-real time stuff. I suppose if you are controlling it via a software sequencer then there would be a good bit of real-time type data but I would think thats much better handled via the midi system rather than a general purose GUI control port. -- Richard A. Smith Bitworks, Inc. rs...@bi... 479.846.5777 x104 Sr. Design Engineer http://www.bitworks.com |