stubbe63 Since Gen~ offers sample-level processing I wonder if Bela can too?
you can go down to I/O audio buffers of with two samples per block. The internal processing of
[gen~]-generated code is always sample-by-sample.
stubbe63 Are there any latency benchmarks of Belas audio input to audio output?
Plenty,e.g.: http://www.eecs.qmul.ac.uk/~andrewm/mcpherson_aes2015.pdf, https://qmro.qmul.ac.uk/xmlui/handle/123456789/12479 . Ultimately, pass-through audio latency is ~40 samples (converter internal filters) + 2 * blockSize, so with a blockSize of 2, you get the baseline value of ~44 samples (< 1ms).
stubbe63 Ultimately I want to run several functions assembled in one Gen~. How would I compare/monitor CPU usage on a laptop vs Bela?
This is very difficult to compare. Expect that Bela will be able to run much fewer audio processing than your laptop. In particular,
[gen~] - exported code does not seem to be especially fast on Bela.
stubbe63 Will the new Audio Expander Cape run with the BeagleBone Black Wireless board? Seems very similar to the BeagleBone Black.
It tends to work fine with the Wireless, although a number of people have had issues with power: depending on the combination of power supply/cables/chance sometimes it may be a bit complicated to start it. I'd recommend a safer (and possibly cheaper) approach: get a BeagleBone Black (or Green) and add a supported USB wifi dongle.
stubbe63 Are the audio-adapter cables stereo?