Paul Keinanen wrote:
spot frequencies are undetectable due to selective hearing
loss.

One approach would be to run the CW signal through some kind of FM
receive, limiting the carrier, but producing white noise when no
signal is available (and you would have to learn to copy negative-CW)
or alternatively use some amplitude detector to control a noise gate,
i.e. when there is a carrier present, the white noise would get
through, with no signal, the headphones would be silent.

Just an idea.

Paul OH3LWR

Andy writes:

A hard limiter decreases the signal to noise by about 5.6 db, and
that's
a mathematical fact...

If we know the characteristics of the noise, and the
characteristics of
when it occurs and the distribution of the energy, then we have
"a priori" experience , and it really isn't "noise" anymore, and it can
be
dealt with --- in some cases rejected - and it's effects on the
intelligibiliy
of a desired signal made less...

There are two characteristics familiar to a radar engineer, which
deals with detection of a signal in the presence of noise::

Probability of detection -- Which is the probablility that a
signal
will be detected in the presence of noise without
an error.

Probability of false alarm -- This is the probability that
detection of
a vailid signal will occur when there is, in
fact, no signal
present.

Entire volumes have been written on Pd versus Pfa, since this
means
life or death to an aircraft (for instance) when a missle lock may
happen.....

If the other guy detects you and sends his missle before you
do, you will probly be dead.
If you fire your missles off at a ghost and have none left, you will
probly
be dead..

In the final analysis, CW is just On-Off signals, much like pulses.

The bottom line in all this is that if you KNOW what the signal is
going to
do you can increase the chances of detecting it properly.

For instance, if the signal is repetitive, it can be stored,
integrated,
differentiated, or accumulated with weighing functions to recover the
intelligence ---- "a priori" knowledge is necessary..

If you know what the noise is going to do -- impulse, popcorn, static,
broadband, random, etc --- you can use techniques to reduce it
hopefully without reducing the signal..

So fancy noise limiters, signal enhancers, and innovative detectors
will
work on some types of interference, and not on others. To be a
universal
S/N improvement, it has to work on "unknown" interference...

That's what the ham bands are like. It could be AM splatter, white
noise, a welder machine, the "woodpecker", car ignition.....
whatever....

That's what restricting the bandpass does, usually. Sometimes it makes
the S/N worse, but only with "special" types of interference.

There ain't no "magic bullet".....

The subject is a LOT more complicated than just any single simple
technique for recovering a signal...

However ,our ear/brain, with PRACTICE is an adaptive filter. It's
amazing
how well it works, after someone has been copying CW for a while.
Perhaps a microprocessor controlled adaptive filter can be made to
approach it, but ADAPTIVE filtering is the only hope that I can see,
given the different types of QRM and QRN that I have encountered.
My best bet is that someday an adaptive CW filter would be to do
as good as my own ear could do today.....

It's like looking at a noisy signal on a scope. Someone with a lot of
practice can see a valid signal several db lower in S/N than a novice
can do.... Sonar operators can do the same.....

"Waterfall" displays simply integrate the signal and noise over time,
which is similar mathematically to restricting the bandwidth as far as
the
S/N "enhancement" properties.... The characteristics of the signal
and the characteristics of the noise are known beforehand, and
that is used, via the display, to increase the Pd and decrease the
Pfa....

End of rant..... I need a beer.


Andy W4OAH

(retired communications and RADAR systems engineer and
ham for about 45 years, or so.... hell, I don't remember any more )