Hi all,
I just dropped the first "beta" releases of the 2400 bps modem online. They are located here: https://github.com/on1arf/gmsk/tree/master/gmskmodem_codec2/development/beta These are both "beta" software. The first version does not apply nor FEC nor interleaving. The 2nd version applies interleaving but not FEC. (althou the interleave matrix is designed to implement FEC). There is a message on my blog that explain some of the things that where needed to create the 2400 bps modem. http://villazeebries.krbonne.net/hamstuff/?p=231 There is also a document that describes the interleaving matrix: http://villazeebries.krbonne.net/hamstuff/?page_id=242 Thanks for everybody who helped out with this, especially David, Bill and Jim. Note that this is still "beta" software. There are a couple of things that need to be sorted out: - of course, implement FEC - there is a small bugs in the "interleave" version that adds one frame at the end of the stream that was not send. This needs to be fixed. Also, one other element is that I have found that even when testing the modem; the GMSK modulation and demodulation code seams to create biterrors. (not that much, just one per some hunderd bits; but they are there). With the 4800 bps modem, I never had this problem. So there seams to be some error in the FIR filters for the 2400 bps modem. Tests that can be done: - create some codec2 file (say "testfile.c2") - convert codec2 file into "c2encap" file. (see source on https://github.com/on1arf/gmsk/tree/master/gmskmodem_codec2/c2encap) c2encap <testfile.c2 >testfile.c2s - create hexdump of c2encap file: od -t x1 testfile.c2s >testfile_before.txt - GMSK modulate codec2 file gmskmodem -v -v -noreceiver -sif testfile.c2s -sof testfile_gmsk.raw - GMSK demodulate GMSK audio fule gmskmodem -dd -v -v -nosender -rif testfile_gmsk.raw -rof testfileout (this will create a file "testfileout-0000.s2c") - create hexdump: od -t x1 testfileout-OOOO.s2c >testfile_after.txt - compair both files: diff -y testfile_before.txt testfile_after.txt I did some tests to determine if there are certain bitpatterns that create this biterror more oftent then others, sofar without luck. Before starting with the FEC, I will, on request of some people, first write the documentation and create the tools so that is possible to recreate the tests done during the GMSK demonstration video. 73 Kristoff - ON1ARF ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_nov _______________________________________________ Freetel-codec2 mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freetel-codec2
