At 06:52 AM 8/31/2009, you wrote: > > Well, over here, the biggest challenges are technical, such as "Will > > the signal survive ionospheric paths?". Only one way to find > that one out. ;) > > > >Digital voice on HF will NEVER catch on because it is a strong >signal mode, and HF is the home of weak signals, well except for 80
Never is a long time, and vocoder technology marches on. Already, the biggest limitation is not technology, but patents and what we hams can get hold of for an affordable price. >DStar is AMBE vocoded voice fed to GMSK modulation while G4GUO/AOR's >HF digital protocol is AMBE vocoded voice fed to OFDM (specificially >36 carrier DQPSK) modulation. Depending on how you measure >performance, GMSK generally requires about twice the bandwidth as >OFDM for the same data and error rate, so there will be additional >regulatory issues in the USA for HF GMSK, at least below 28 mhz. Well, outside the USA, HF D-STAR GMSK is legal. It certainly is in Australia, as the bandwidth limitation is 8 kHz, any mode allowed. The ACMA lists D-STAR's GMSK modulation as having a necessary bandwidth of 6.25 kHz, so a nice comfortable fit within the bandwidth limit. >I've not thought about the s/n required for CW to be intelligible, >but I bet it's just a hair above zero. 10 db s/n generally produces >amateur quality ssb, and 15 db s/n is the minimum for commercial >quality ssb. G4GUO needed 25 db over a 70 km path on 40 meters for >reliable communications (see first link below). When the ARRL >reviewed the AR9800, they needed 18 db s/n to avoid syllable dropout >(see second link below). Yes, I am aware of that. Has anyone measured the S/N required for FDMDV to work? No doubt, SSB is still going to be the more robust of the two modes. :) >Although focusing on satellite communications, Phil Karn (KA9Q) give >an excellent explanation for poor digital performance at low s/n - >"Decrease the power of a SSB signal by 1 dB and the result is simply >a 1 dB reduction in audio S/N. Most might not even notice. (Even the >differences between the simulated SSB signals given here are a >little subtle to me.) But a 1dB change in S/N to an RF demodulator >and FEC decoder can make the difference between perfect operation >and no operation. Being inherently nonlinear, digital modulation has >a threshold effect much like that of wideband FM, only more >pronounced. The stronger (i.e., more efficient) the modulation and coding, That's true. The question is where that threshold is, which is dependent on the bit rate, and the bit rate is dependent on the sample rate, codec/vocoder used (this is the biggie) and the coding scheme used (i.e. FEC, if any, etc). Lower the bit rate, and you lower the amount of signal required. The success of modern narrowband digital modes demonstrates this principle. Of course, we won't get voice that low, but what's possible if we could get hold of a vocoder that can encode speech with useable intelligibility at 600 bits/sec? Of course, it's not going to be hi fi, but the early versions of FDMDV demonstrated what's possible at 1400 bps (and also demonstrated that licensing is as big an issue as technology when it comes to achieving the best performance). With D-STAR, we are fortunate that we can buy AMBE on a chip. It's a pity that we couldn't do the same for the MELP codec used in early FDMDV versions. >the sharper the threshold." (see third link below) Imaging this >effect with QSB right around the threshold... And as Tony says, HF >multipath is a killer for these demodulators. SSB is going to be around for a very long time yet, but digital will creep in for certain purposes, as technology and access to technology improves. 73 de VK3JED / VK3IRL http://vkradio.com
