Packets decode using gr-satellites / lilacsat1.py with the message debug block enabled, using raw iq with gqrx in USB at 9.6k offset below, udp to ./lilacsat1.py
(tip of the hat to k4kdr tweets for a clue https://twitter.com/scott23192/status/1172553961984974848)
wanting to record iq spectrum to decode audio. Found this note from: https://amsat-uk.org/2019/09/13/taurus-1-with-codec-2-launched/
" (no response on 436.760 MHz when transmitting on 145.820 MHz)”Note:this was sent before it was realized Codec2 voice and the telemetry were on the same frequency."
converted from short to float as usual with: $ sox -es -b16 -r48k 1026508.raw -ef 1026508_fc.raw
Then used this version grc script derived from k4kdr’s: http://swigerco.com/demod_taurus-1-satnogs.grc
which was changed to use file input instead of udp, removed Plan13 since you already doppler correct, and fixed a file sink name to something generic.
But wait, it gets even uglier - the gr-lilacsat scripts were created with an older version of python ‘Construct’ 2.8.21 or earlier, and gr-satellites now uses version 2.9 - SO I check out an older version of gr-satellites that works with 2.8.21
$ sudo pip install construct=2.8.21
$ git checkout 40319495558fafc910253b68b616235d72b6304c
rebuild and then above script works, and outputs the audio sample above.