That looks like a stand-alone STK-based app.
In principle a new Bela backend could be added to RtAudio.cpp/RtAudio.h, so this will help with porting this and other STK. However, there is also lots of GUI stuff in there (see displayFunc() in rt_lpc.cpp), that would not work on Bela, so there is stuff to be edited out there as well.
It seems that all the audio thread is doing (see cb()) is simply copying input samples into g_audio_buffer[] and then output samples from g_another_buffer[] into outBuffy[], while the actual LPC computation is performed in the GUI thread (displayFunc()), where (lines 930 to 952) the calls to lpc_analyze() and lpc_synthesize() take place.
I think that you could just add the relevant files (possibly just lpc.cpp and lpc.h to a Bela project (whatever you do, make sure you don't add a function called main(), or it will take the place of Bela's one and nothing will work), call initialize_analysis() in setup(), lines 930to952 of rt_lpc.cpp in an AuxiliaryTask, and what is needed to pass data to the auxiliary task in render(). Not straightforward, but doable.
