[TenTec] Digital Speech Processing & Pegasus

George, W5YR w5yr@att.net
Mon, 18 Feb 2002 19:42:15 -0600


Mark, historically speaking, processing at audio baseband has not proved to
be all that effective in enhancing articulation and improving effective S/N
for SSB signals. True, it can be highly effective in increasing average
power, which is alright is one is willing to accept that most of the
increase comes from distortion products within the baseband.

It is for that reason primarily that audio "compression" or clipping has
fallen into disrepute and is seldom used in linear systems such as an SSB
transmitter. FM rigs can benefit from direct audio processing, as can AM
rigs to an extent, but not SSB.

Over the years a number of r-f clipping and filtering systems have been
developed and marketed, most of which I have had an opportunity to work
with. My pick of them all has been the Comdel processor which generated
from baseband audio input an SSB signal at a 500 KHz suppressed carrier
frequency. 

This signal was then heavily clipped and filtered at 500 KHz to remove all
the distortion products which theoretically should all lie outside the
original audio passband.
Next, the clipped and filtered signal is heterodyned back to baseband with
the same oscillator that was used to generate the 500 KHz SSB signal, thus
avoiding any frequency-shift distortion. 

The resulting audio signal is then amplified and low-pass filtered and
output to the mic input of the transmitter. The audio output when used with
speech is of notably higher average power with a minimum of distortion if
the increase is held to about 6 db to 9 db. It is possible to obtain 10-15
db increase but the distortion by then is noticeable to the ear. At 6 db, a
sine wave input is indistinguishable from the Comdel output, the distortion
is that low.

Several systems of that type have appeared over the years: DX Engineering,
Waters, and others no longer around. TenTec used to make a similar product
that I understand was quite effective - I do not know if it is still being
made; I doubt it.

The reason is that most modern rigs now incorporate some form of
"processing" if for no other reason that to be able to make the marketing
claim in the specs. Usually this takes the form of mild compression of the
signal somewhere along the IF chain.

The one completely audio processor that was both effective and
low-distortion was the Vomax made by Maximillian Associates. I mention this
especially since it represents an approach that you perhaps might want to
explore for implementing in the computer with .wav files as input.

The Vomax was a "split-band" processor that functions altogether at
baseband. The incoming mic signal is split into four adjacent audio
frequency channels, and the signal in each channel is clipped and filtered
as with r-f processing. Then, the magic comes when the clipped and filtered
multi-band signals are adroitly recombined with minimum distortion to
produce a baseband output with improved articulation and a modest increase
in average power.
The limit to its capability, as with the others, lay in how much distortion
you were willing to accept for how much power increase.

I have used all of these types on the air usually with outboard units. I
may be wrong here, but I have been told that the Icom PRO implements a form
of split-band audio processing within its DSP core. And I have heard that
the PRO II has improved upon this to generate a true r-f clipped and
filtered signal in DSP. If you have access to old issues of Ham Radio
Magazine, the prototype of the Vomax was a feature article sometime in the
early 70's as I recall. Later on, QST published an updated design using the
same approach but by a different author.

So, perhaps all this trivia may encourage you to pursue some form of speech
processing capablity in your software. I would be leery of simple "dynamic
limiting" however. Unless done with fairly long time constants, in which
case the effectiveness is impaired, then the effective clipping of the
audio signal generates distortion products that remain within the audio
baseband signal. 

I suspect that you will also find it advisable to afford the user with
means to tailor the audio chacteristics of the input and/or output signal
in your processing. Standing out in a pileup is not done readily with a
broadband, but beautiful sounding, SSB signal with tons of bass content.

Please keep us informed on your progress. Speech processing was a hobby
within a hobby for me back in the 70's when I designed and built an r-f
processor for my Heathkit SB-401 transmitter that was quite successful.

72/73/oo, George W5YR - the Yellow Rose of Texas         
Fairview, TX 30 mi NE of Dallas in Collin county EM13qe   
Amateur Radio W5YR, in the 56th year and it just keeps getting better!
QRP-L 1373 NETXQRP 6 SOC 262 COG 8 FPQRP 404 TEN-X 11771
Icom IC-756PRO #02121  Kachina #91900556  IC-765 #02437

All outgoing email virus-checked by Norton Anti-Virus 2002

Mark Erbaugh wrote:
> 
> As I continue my work on my own custom Pegasus control software, I have now
> incorporated the ability to play .WAV files through the aux input. This will
> allow me to use the computer as a digital voice keyer.
> 
> However, I think that I can take advantage of some sound processing software
> on the computer to customize the .WAV file to give it more punch. Other than
> doing some dynamic limiting to increase the average power, are there things
> that can be done to increase the chances of standing out or at least being
> heard in a pileup?  Is there a model of HF propagation that can be used to
> experiment with the effects of various enahancements?