soundsync VS aubio

Maybe try THIS. It supports local speaker as an output (I'm pretty sure) and allows many inputs.

Aubio is a good library for working with audio and midi: https://aubio.org/

How does this compare to Aubio?

Beyond that you'll need pitch detection/pitch tracking. It's not the most difficult thing in the world but you need basic understanding of digital signal processing, and the math behind it. Otherwise I think this library can do the heavy lifting for you.

aubio looks like an option. It contains the console program aubiopitch with a “jack” option for live audio. HTH.

I dunno, that article is paywalled. ES-HyperNEAT uses what's referred to an an evolvable substrate, which is the internal architecture of the ANN being involved. This provides for arbitrary neuronal wiring which is interesting and a departure from conventional GPU-based training architectures based on layers. For classifying audio stuff typically spectrogram analysis with CNNs are used, to pick up on latent features from a visual perspective. Probably getting your audio sources into an appropriate format would be the challenging part of it. Maybe experiment with the output from Aubio to get it into a representable format that an ES-HyperNEAT ANN could consume.

Hey, glad that you like the idea. For the audio-analysis, I'm using aubio (https://aubio.org/), great library, once you managed to install it :D Was a bit tricky in my case.