Node graph clarifications

nothing strange, indeed you are kind to be interested in this project of mine, thanks!

My product will run basically on windows and mac os (i’m interested also on iOS, not in the short term).

Basically: I born as a musician that needs to manage very big and complex scenarios of concerts and events and about one years ago, when works started to be so many, I started searching some different software solution for basically 1) reduce the hardwares needed for those purposes, 2) reduce my time in programming and preparing those events (where often I waste time only to do stupid and triky things because of softwares limitations), 3) having a smarter and better structured approach to improve my workflow.

In my various studies of a lot of daws and a lot of softwares like main-stage, cantabile, vst-live and so on, I’ve never found any convincing solutions. so I started to do a bit of code by myself to structure something new.

Up to now I’d like to have something that works as soon as possible because I would like to stop having sleepless nights (:sweat_smile:) because of my workflow and also to use in real scenarios to test it ( obviously this “something” to test could non have some secondary features but need to do a stable work because I don’t want to find myself dealing with embarrassing moments with thousands of people in front of me during the show ). I also think that what I’m doing should be really appropriate in various scenarios also for other user, so in bit more long term I like to make it commercial.

Plugins to run are basically third party plugins. My program will be a standalone application and will need to manage really really big projects (so many plugins, routing, automations dmx and midi map), so is something possible only with a multithread app (also for a basically programming approach there will be more performers with more graphs and stuffs in the same app, so, also conceptually, more threads that manage different tracks or graphs). Up to now I don’t benchmarked a single thread system in so analytical way, but I do some test reproducing real scenarios I’m actually working on and the “single thread” fails.

My personal experience with tracktion (not only for the graph) is at beginning really positive, but studying it I founded that it’s really not good at the actual state of my needings… or better: a lot of things tracktion has a beautiful approach and gives me a lot of ideas to do things at better way, but various things of tracktion are not usable without the “edit” and without the “engine” and this is a pity because tons of tracktion sub-modules are “wow” and it should be really pretty to use them separately when needed.

Up to now I think better solution is searching a good strategy to perform some juce::AudioProcessorGraph and node inside in multithread way. It’s really a pity that the graph, that has a fantastic logic, cannot be used as base class to make your custom AudioPluginInstance and to talk with some juce other stuffs without a tones of lines of code.

Sorry for long message !