Attention gmfsk users. I am putting a new PSK31 mode in gfmsk. I need some test guinea pigs. In this case, there will be one, two, ...., N orthogonal trellis coded PSK31 signals. Each tone will have 4 or 4.5 dB of coding gain over the PSK31 but will otherwise work the same.
When Peter did the QPSK Viterbi decoder on a constraint length five code, he didn't realize that you were going to realize essentially zero coding gain over PSK31 with that arrangement. Not to mention the fact that to decode with anything like full gain, you need to wait 15-20 bit times to get the job done. With this Trellis code, it is 8 states or possibly 16 states (testing will tell) and will decode in 1/2 or 3/4 as many bits. By going to N parallel tones, we can choose between higher speed or more robustness. More robustness will come at the cost of decoding delay for the interleaver with concatenated TCM and Reed Solomon interleaved data. I will first release the single tone TC-PSK31 and then go to N in parallel. The thing that Peter got so right was the source coding with the varicode. Please send me private email if you are or quickly can be an gmfsk user on Linux. This will be one of my talks at DCC in Hartford. Bob N4HY -- AMSAT Director and VP Engineering. Member: ARRL, AMSAT-DL, TAPR, Packrats, NJQRP, QRP ARCI, QCWA, FRC. ARRL SDR WG Chair "If you're going to be crazy, you have to get paid for it or else you're going to be locked up." Hunter S. Thompson
