So I’ve played around a bit with the Raspberry Pi embedded computer and a USB sound card (Soundblaster) in an attempt to use the soundmodem software TNC for APRS tracking. My initial impressions are that my current implementation (Raspbian “wheezy”) is not entirely up to the task. It appears that soundmodem is not able to reliably demodulate incoming AFSK data – perhaps due to an overloaded processor.
I’m working to optimize the current distribution I am using in the hopes that I can trim as much fat as possible – freeing up precious clock cycles for software decoding. In reality, this software solution should work without a hitch as the Raspberry Pi has more than enough processing power to FFT the audio and decode it that way. The reality is mostly missed frames at this point.
I have no doubt that the Raspberry Pi is a fantastic platform using an external stand-alone TNC, acting as an iGate or digipeater. My hope is to have a purely software implementation (aside from the sound card for audio input). I will keep plugging away at the kernel and peripheral modules – there has got to be a way to reliably decode packets in software!