hello.
i use puredata on a bela-mini to capture sensor-data for audio-synthesis running through some machine-learning algorithms (ml-lib).
additional to that, the board sends the sensor-data out via wifi.
dependent on the the ml-object, training-data, audio-synthesis and sending-speed used, i get:
Underrun detected: 1 blocks dropped
seems data-processing and network tasks are too much for the audio to keep up.
the pd-patch basically looks like this:
adc~ -> snapshot~ -> ml -> synth
|-> netsend
i wonder how it would be possible to detach the data-analysis and data-sending part from the pd audio-thread or run these in a separate process. pd~ object seems not to exist on bela...
is it possible to do this with pd?
or would it be better to run the pd-patch from within a c++ project?
thanks for any advice.