Hello Moe, Thank you very much for your comments.
1/ Yes, as I mentioned in the blog post I also wonder about the practical advantages of the coherent GMSK demod. In my HF work (greatly helped by your fine PathSim software) coherent PSK has resulted in a big jump in performance. The first plot in this post: http://www.rowetel.com/blog/?p=3700 Shows a 4dB improvement in QPSK over DQPSK. Much to my surprise, the HF channel results I obtained (using PathSim) are very close to the theoretical results for a Rayleigh channel, e.g.: http://www.dsplog.com/2008/08/10/ber-bpsk-rayleigh-channel/ Both the HF multipath and VHF mobile channels can be modelled as Rayleigh so I expect a similar jump in performance for VHF when using a coherent demod. It also gives us is head room for other stuff like TDMA or FEC overhead. Costs us nothing if implemented in software, so the engineer in me can't leave that performance on the floor! 2/ I haven't thought about higher layers yet, like FEC, interleaving, or protocols. Starting with the basics and I'm a bit of a physical layer guy myself. I'll take a look at those refs. 3/ I am forming similar conclusions about latency and PTT (simplex) radio. No one is going to notice a few 100ms of delay, as long as the sync algorithms are fast, recover quickly from fades, and don't drop out for seconds at a time. 4/ Yep I had to do some optimisation for RAM for the SM1000 although it was pretty straight forward. Yet to see if it's an issue for larger frame sizes. Ultimately I am chip agnostic, it's all going to be portable software in a few years. 5/ Yes it might be worth playing with C4FSK. I am assuming its QFSK, or 4 freq levels. I think this has the same BER versus Eb/No perf as GMSK and BPSK. Perhaps there are some other advantages. That Yaseu PDF quotes identical rx sensitivities for DSTAR/C4FM FDMA, and they are only 4dB better than FM. In the tx specs it appears that they are using FM style modems "variable reactance" even for DV. Cheers, David On 22/12/14 01:19, ae4jy wrote: > Hi David, > From a practical standpoint I wonder how much of an advantage a > theoretical limit synchronous GMSK demodulator would be in the VHF/UHF > mobile environment where signals rarely hover around the Shannon limit. > They are either well above or well below due to multipath/Doppler etc. > My gut feeling is an interleaver and FEC would do much more in that > environment. I know folks are paranoid about turn around delays but in > reality hundreds of mSecs is not that big a deal and not having to > listen to squawks from a bit starved vocoder might be worth it. > > For whatever reason, the commercial world seems to have moved to C4FM > instead of GMSK so might be worth a look as to why. Here is a(of course > biased) report but I think the arguments at the low level modulation > level are valid, the system level arguments not so much. > > http://www.yaesu.com/downloadFile.cfm?FileID=7146&FileCatID=151&FileName=DigitalCommunicationsGuide%5FE%5B1%5D.pdf&FileContentType=application%2Fpdf > > One could also leverage some of the existing signalling technology > instead or re-inventing the wheel. > Google for ts_102658v010201p.pdf will find an ETSI TS 102 658 spec for > dPMR which has a section on the physical layer that might be of use. > > As a side note it looks like the STM32F7 series will be pin for pin > compatible with the stm32f4 and give us a little more of everything > especially ram which is what I seem to run out of first when doing DSP. > > 73, > Moe ae4jy > > > ------------------------------------------------------------------------------ > Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server > from Actuate! Instantly Supercharge Your Business Reports and Dashboards > with Interactivity, Sharing, Native Excel Exports, App Integration & more > Get technology previously reserved for billion-dollar corporations, FREE > http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk > _______________________________________________ > Freetel-codec2 mailing list > Freetel-codec2@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/freetel-codec2 > ------------------------------------------------------------------------------ Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server from Actuate! Instantly Supercharge Your Business Reports and Dashboards with Interactivity, Sharing, Native Excel Exports, App Integration & more Get technology previously reserved for billion-dollar corporations, FREE http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk _______________________________________________ Freetel-codec2 mailing list Freetel-codec2@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/freetel-codec2