[sc-users] VstPlugin.sc - a class to load VST plugins in SC

[christof.ressi@xxxxxx]

hi, I've been developing a Pure Data external for loading VST plugins (https://git.iem.at/pd/vstplugin) and it has been working quite well so far. Now here's my first attempt at porting it to SuperCollider:

https://git.iem.at/pd/vstplugin/tree/supercollider

see the README for a quick overview and how to build the project.

Here are binaries you can test in case you don't want to build from source:

Windows (64 bit): https://drive.google.com/open?id=1nIFdK0oArsCUBoXfH6mP548ZCagsDmsD
macOS: https://drive.google.com/open?id=1KZJHwWd-sLpA1JG7TieHX488as2yqHuU

for more detailed information and examples checkout the VstPlugin class help file.

---

Here's a short explanation of the current implementation:

The VstPluginUGen class is responsible for loading VST plugins on the server but it's only used internally. The VstPlugin class provides the actual high level interface by wrapping a VstPluginUGen in a synth, hiding all the necessary messaging between server and client behind instance methods.

Right now, the class method 'makeSynthDef' is responsible for creating the necessary SynthDef for all VstPlugin instances which share a certain number of inputs and outputs (+ whether output is replacing or accumulating). In the SynthDef's UGen graph function, an instance of VstPluginUGen.ar takes its input from In.ar and outputs to either Out.ar or ReplaceOut.ar, depending on the 'replace' argument. Using a single Node for each plugin helps tremendously with the OSC messaging (because I need to know the SynthIndex of VstPluginUGen within the UGen graph function), also you can freely move individual plugins within the graph, changing your FX chains after the plugins have been loaded.

I *could* get rid of 'makeSynthDef' if VstPluginUGen would directly read and write to audio busses (without the help of UGens like In and Out). In this case we'd only need a single SynthDef for all VstPlugin instances regardless of number of input and output channels and it could be created internally. If you think the current 'makeSynthDef' solution is too messy/confusing, I could totally do that!

---

I have to say that my expertise is really Pd and I'm only a casual SuperCollider user, so I'm open for any kind of suggestions!

Best, Christof



_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: https://listarc.bham.ac.uk/marchives/sc-users/
search: https://listarc.bham.ac.uk/lists/sc-users/search/