Remove 16 byte AX.25 header, find 1st & last '03' image packets, note sequence # & add missing frames when possible. When 100%, remove INFO bytes only leaving JPG file from 'FFD8' to 'FFD9'
… and to elaborate, obviously this can be done by a programmer but I could Google until I turned blue and not find all the needed code.
Also, while you could certainly do these steps in a hex editor, I find it easier to convert to ASCII & do all the ‘cuts’ and ‘find/replace’ in a text editor or on the linux command line.
Only when I’m finished to I convert back to hex to save the complete .JPG file as binary hex.
Thanks for sharing this IQ recording, it is however quit a challenge to replay the format without information on how it is created. I think this was created with SDR Console, that would explain the WAV64 format.
But at the end I was able to convert it to a usable file.
One final remark, but without any details on the receive chain, I think the RSP is configured with far to much gain and therefor not in its optimal dynamic range.
When you replay the file, you see and experience the gain being changed, this is one of the reasons I stopped using the RSP, maybe it is a hardware setting in SDR Console.
Good description, Scott! I have a single addition: the strips have to be done in the raw data and must not be KISS encoded! If it is a KISS frame, one needs to “unKISS” it first, as there might be bytes escaped!
For anyone who uses GNU Radio, here is the FIRST step that I use when starting with a KSS file saved from a CAS-5A pass. In addition to converting from KSS to normal HEX, it also removes the first 16 bytes (AX.25 header) from each packet. You are left with a file ready to check for missing image packets!
Maybe it is time to let go of my bias towards RSP and give a later model a try to see if things are improved. I had one of the very first models, already many years ago.
