Some more info: I found a paper that describes some tests done with 2.4Kbps and 1.2kbps voice transmission over HF paths. It sounds like the 1.2 kbps gives useable voice quality.
I talked to a friend who had done some research for the military back in the '80s on digital voice transmission over HF. He said that the state of the art back then was about 2.4Kbps but there was a technique developed for decoding the speech into text with a very limited vocabulary for tactical operations - around 300 words - and sending the words with minimum coding required for the limited vocabulary. The transmitted data was then reconstructed into speech at the receiving end. The data rates achieved with this method were as low as 300 bps. There were a few problems with this method, though. First it introduced substantial delays due to the processing required for the speech recognition. Second, the Aussies got very upset when their speech came out the other end of the link with a midwestern accent. :) Finally, when attempts were made to apply this method to an application for the Marines for fire control communications, the Marines were unable to come up with a 300 word vocabularly that didn't contain profanity or obscenities so the project was killed. Seriously, long distance HF propagation imposes some very challenging problems with long fades that can only be overcome with long interleaving of the data which creates long delays in the transmission. I will continue to study the issue and report any interesting ideas that show promise for weak signal digitzed voice communications. Ed --- In digitalradio@yahoogroups.com, "Ed Hekman" <[EMAIL PROTECTED]> wrote: > > --- In digitalradio@yahoogroups.com, Tony <DXDX@> wrote: > > > > Ed wrote: > > > > > Are there any communications engineers in this group that can > > > give us some idea whether a useable quality digitized voice can > be > > > sent over a 2.5 KHz wide HF channel with SNR comparable to or > less > than > > > what is required for analog voice? > > > > I was thinking about this today Ed. I'd sacrifice a bit of voice > qaulity for > > better SNR performance. > > > > WinDRM defaults to the MELP codec and I was wondering if the SPEEX > or LP-10 > > offer an improvement in SNR performance? > > > > Tony KT2Q > > Tony, > > The key to better weak signal performance will be primarily in how > the data is sent over the air. Observing the signal for digital > SSTV on 20M has been very interesting. Often a hole in the spectrum > can be observed as it moves across the spectrum. It takes out about > 20% of the spectrum and has a time span at one frequency of around a > second. The packaging of the data must be done with enough > redundancy spread throughout the spectrum so it can be recovered in > spite of these spectrum holes. On 80M atmospheric noise tends to > have short impulses that take out the entire signal for much shorter > periods of time - much less than a second. To mitigate this, the > data redundancy must be spread over time so the data can be > recovered in spite of a complete loss of signal for a few > milliseconds. > > Of course adding redundancy means reducing the data rate for the > encoded voice data. This would probably require some adjustments to > the voice coding algorithm. With cellular signals, although the > maximum data rate may be 8K bps, the effective rate is usually less > than 1/2 that due to the fact that speech is not a constant signal - > there are holes in it for short periods of time during which no data > needs to be sent. Need a Digital mode QSO? Connect to Telnet://cluster.dynalias.org Other areas of interest: The MixW Reflector : http://groups.yahoo.com/group/themixwgroup/ DigiPol: http://groups.yahoo.com/group/Digipol (band plan policy discussion) Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/digitalradio/ <*> Your email settings: Individual Email | Traditional <*> To change settings online go to: http://groups.yahoo.com/group/digitalradio/join (Yahoo! ID required) <*> To change settings via email: mailto:[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/