I am trying to build a project with my rtl-sdr dongle. The idea is to process the radio audio signal through a machine learning model. My model currently takes 10 seconds wav files as input and converts them to numpy arrays that are further processed and then classified.
I would like to classify 10 seconds audio chunks directly when they are generated by my dongle. I figured I would probably need to take a look at PyAudio to generate my chunks, however I don't really get how i could define my r820T dongle as a source.
Does anyone know how to access the stream generated by an rtl-sdr dongle directly as python numpy array ? Or maybe how to pipe the stream from rtl_fm reliably to a python progam ?
Thanks in advance !
Need help installing or figuring out something about your SDR? Ask here.
2 posts • Page 1 of 1
The stream generated by an RTL-SDR (which can be accessed by librtlsdr) is baseband IQ data, not audio. So your best bet is the latter: a pipe from rfl_fm.