I appreciate all replies!
Reason for Pi is I have a bunch of them left from various projects over the years and they're super embeddable and modular without having to design custom circuits. Same with arduinos. Ideally a custom FPGA ASIC would be righteous. As per the paper though "FPGA-accelerated Real-Time Audio in Pure Data" they were successful doing physical modeling on a Terasic DE10-Nano but tooling for any realtime interactivity with the core was essentially non existent at the time of writing (2022). I don't understand FPGAs well enough to come up with a good solution yet, but parallel programming is somewhat comparable to writing an orchestral piece. The logic circuits can be rearranged so you could have different setups for different heavy synthesis types, just not on the fly iirc.
I could see having two FPGA cores to switch between heavy instrumentations w an MCU handling the rest of the audio functionality. Will need to hire ssomeone to write this. Accepting applications!
Thankfully modern FPGA's allow for c++ and python.
The Box im building is for patch recall and midi mapping with realtime programming built into the workflow. I want to get away from the laptop .with. PD.
@alexandrous exactly. Laptops are wonderful, absolutely, but looking at the screen for hours building sound machines and then using the same thing to improvise emotionally and then for email, web etc is wearisome and counterproductive. Keeping form factor and desk space small is crucial as well. Yes of course a screen is still necessary but w/a workflow more like an Elektron machine hence the desire to know the computational expense of atoms pushed to their limit (which is obviously a super lengthy and nearly impossible venture). Industrial 3" screens are are kind of better than capacitive or resistive touch screens
So maybe HiFiBerry is at the top of Pi DACs. Patchbox feels like a toy, definitely could be wrong. If latency between platforms can be solved a Pi/arduino cluster of sorts can make pD experimenting much more fun than sitting with the laptop ALL day. I don't think the human brain feels the same about things we can't immediately aurally detect. The vibe of a violin vs that of a synthesized violin is different. Embracing the difference is one thing but having standards in the architecture of the synth is like having standards in the making of a Stradivarius vs an amazon basics factory production. We're living in the decline of culture bc we accept the 'just enough' philosophy of building without functional aesthetics. Starting to ramble..
Bela looks really good and agree with @old about Elk and the site etc, lol, I perceive they have things going on behind the scenes, buyouts, integration, who knows... they may surprise us. Either way I've found a good route so I guess it's the HifiBerry or have a custom built to spec.
Flux-ai could probably design one flawlessly. BTW very interested to see @old 's large language patch mentioned in previous discussion.
Keep Calm and Build Community