There are some improvements that could be made (while using legacy FM demodulators).
A common mistake is to tightly band limit the input the the modem. The problem with this is that the FM clicks- spikes spread out and cover perhaps several bits of information . The clicks have to be removed before bandwidth reduction. since the source is ~ 700bps, and the bandwidth available is large, you might as well up the deviation to at least +/-2kHz. Keeping the symbol rate low is essential for reduction of the effect of long delay multipath. DMR and P25 with their 4800 sps are quite sensitive to long delay multipath and there are many instances where there is tons of signal, maybe 40dB CNR and no communications due to say 20uS multipath (6.5km round trip), and this is easy to get in the suburbs on UHF. With the low symbol rate of 700 sps , and very low SNR requirement, this would provide good multipath performance. anyway a few ideas. Of course performance all inferior compared to QPSK / FSK when optimally demodulated... -glen On 14/09/2017 8:28 AM, Adrian Musceac wrote: > Hi Glen, > > We don't need to use the same chain for digital demodulation, since we > are not software constrained. Using Gnuradio, we can simply reuse a > flowgraph and connect/disconnect blocks on a click of a button to > switch from one mode to another. What I don't want to do, is write a > custom Gnuradio out-of-tree block like the gr-ysf. I would like to > create the flowgraph using only the default blocks, and do that > efficiently enough to have lots of room to spare on an ARM CPU like > the one I have on my phone. That's because there will be other stuff > coming, including a TDMA scheduler which is very sensitive re. the CPU > frequency scaling implemented by the (Linux) kernel, as well as H264 > decoding - which I'm not sure I can do on the dedicated metal because > of proprietary stuff which depends on phone brand. Please don't get me > started on reverse engineering Qualcomm qmux device communication :) > > I guess what I'm asking for is one volunteer for the 4FSK demodulator, > which I would like to use soft symbol decoding and Viterbi. Any > takers? Also, for people who like to test even buggy things, I might > have an SD card image ready to pop into a phone and play with. > > Cheers, > Adrian > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > Freetel-codec2 mailing list > Freetel-codec2@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/freetel-codec2 > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Freetel-codec2 mailing list Freetel-codec2@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freetel-codec2