finalize initial SpatLib implementation
|Target version:||(Jamoma Platform) - CNMAT 2012 workshop|
_I've basically thought that the SpatLib DSP extension would accept m inlets, and mix them to n outlets. It would have a simple API for adding support for SpatDIF descriptors to a spat unit, and the message space would be extendable so that parameters specific to one spatialisation method could be implemented. And finally the mixing method might draw on all other DSP units available, including gaining, matrixing, delays, filtering, etc.
An initial AudioUnit wrapper for SpatLib would wrap it using multichannel input and output signals. Possibly the multichannel signals would be extended to be able to carry meta-information on what kind of signal we have (e.g. 1st order B-format, stereo, 2 or more mono signals, etc).
The one major issue that I haven't had a solution to earlier has been how the AudioUnit wrapper would deal with the parameter space changing depending on what unit one is using. It might be that the upcoming Max 6 added support for dictionaries might make this a whole lot easier. Anyway we already have dictionaries implemented in Jamoma Graph.