Forum

In order for Talkers to be able to execute in a timely manner (ie, to be processed within one audio buffer of being invoked), which is one of the great things about them, they have to effectively be 'pre-processed', so the Belcanto is not actually evaluated at the time of execution, it's parsed and processed when the Talker is edited. Connections to the needed actions in all the Agents that phrase affects are then established in the background, so the phrase can be executed. If you use Belcanto from scratch (as in from the Commander) then it is not intrinsically timely as the phrase must be parsed, all the Agents must be found etc. This is neither speedy or deterministic, making it unsuitable for things where you need them done at that very moment. Sometimes you'll be fine, but sometimes you might miss your timing if the machine is busy or the command affects a lot of Agents, so it is inherently jittery, which is not good for performance. Pre-evaluation prevents this at the cost of introducing difficulty whenever setups are changed as many of the great Belcanto things involve multiple Agents, and they may or may not be there after a change. Hence the issue when a Rig comes and goes. If it has a scaler inside then the Talker must be re-evaluated to 'see' that scaler, and commonly used Belcanto in a Talker like 'all scale to x set' silently and mysteriously stop working. That's the kind of behaviour that makes me shudder as we end up having to field the (justifiably) annoyed user emails that come from it.

When I refer to a 'pluggable rig' I'm actually talking about something a bit more useful than being able to simply import them in Workbench, I'm referring to being able to insert and remove a Rig into a Rig Host while EigenD is running, rather as you can use VST's or AU's in a DAW. I think this is probably a much more useful piece of functionality, and not far off the same level of difficulty as one still has to solve many of the same problems, namely with Talkers.

If you need any help at any time figuring out the source code, please get in touch and we'll do our best to help..

John

ah, i see the talker phrases are pre-compiled with connections, that makes alot of sense.
but still, importing rig is no different from adding a rig and subsequent components in WB... for this operation, the talkers would also (under current implementation) need 'recompiling'

i can see for a pluggable rig changeable at performance time, this would be inefficient/undesirable and you would need to abstract these connections ... and therefore how you would get then to the idea of 'talker ports' or similar. as you first mentioned, though this then brings up a whole raft of other interesting issues.

hmm, definitely something I will take a look into later :)


so do you guys manually duplicate the rigs within the factory setups? or do you something like hack the setup file?

The reason I was asking all this, is I want to create a setup which basically was like the factory setup, but had you mod synth in each rig, I was then going to tailor each mod synth a bit.
I know i can do this all manually in workbench, but then thought 'wouldnt it be nice to copy/paste, import/export rigs'.
anyway will do manually for now, as have some interesting ideas i want to play with :)


source code, im fine thanks, basically along with the videos, seems to be pretty clear, i seem to be able to find my way around it, only occasionally getting a bit lost in the python (which is not really my thing :)), think my next step will be to build a simple agent - but i think I should get back to improving my playing skills!