I was wondering if one could use DireWolf with the LimeSDR/LimeSDR Mini/HackRF. The bandwidth would be much more acceptable than even the highest rate sound cards, which rarely reach even 192 khz.
Can this be done?
The direwolf
executable on Linux can take streaming audio on STDIN (or UDP). So, if you can find a utility or combination of utilities that can control the LimeSDR and output demodulated audio to STDOUT (or to a UDP port) it should work just fine for monitoring.
Monitoring is easy, though. Can we make a digipeater or an Internet Gateway with this hardware? I think it's possible, but it may take some clever hacking.
There are Dire Wolf utilities that can generate audio for AX.25 frame data. Again, we need some app or utilities that can accept the audio data and bridge to the transmitter.
Linux virtual audio connections can mimic traditional TNC wires, making it possible to route audio in to rig control software. I don't know what sort of rig control is available on various platforms for this hardware, but it might be an opportunity to add supported hardware to a project like Quisk, for example. Quisk can take audio input from a few sources and send it to transmitter hardware it knows about.
HackRF has its own utility that can be used to push a WAV file to a HackRF transmitter so it can be sent. Which might be a good test "sink" at least.
sure, if you do the audio demodulation in software and feed that through a virtual sound loopback device. That's been done before.
However, direwolf seems to be simply the wrong tool then, as it assumes you get demodulated audio baseband instead of RF baseband. I'm almost certain there's SDR-based AX.25 solutions out there.
So, direwolf: use when you have a physical receiver that outputs analog audio. If you have an SDR device, that just gives you the raw RF signal without any of the losses incurred by doing the audio demodulation optimized for human intelligibility instead of for digital data communications, it's the wrong tool.