Dee Flint wrote:
"Chris" wrote in message
...
On Wed, 25 Oct 2006 20:14:03 -0400, Dee Flint
wrote:

... Only the finest operators can send code well enough with a hand key
that a computer can copy it anyway. Only exceptionally good operators
can
send well enough with a bug that a computer can copy it. Only very good
operators can send well enough with paddles that computers can copy it.
Basically a computer is good at copying computer generated code.

That may have been true in the 80's, back when people were just
getting started on the problem of copying CW with a personal computer,
but the algorithms have improved greatly since then, and they are now
quite good at copying manually generated Morse code. Even the area
where humans excelled - copying CW in the presence of QRM and QRN - is
now handled quite well by most modern algorithms. Currently, the most
popular program seems to be CwGet - a Windows program which Breakin
Magazine rates very highly. With gigahertz microprocessors and
built-in A/D converters, the modern PC is more than up to the task of
dealing with computations that were once only practical on mainframes.


I've tried CWGet and it doesn't copy the signals that I want to copy. It
still is subject to problems with QRN, QRM, QSB, and less than perfect
fists. It can't copy any of the signals distorted by aurora. So while it
is the best of the available programs, it still falls far short of a good
human operator. And I'm speaking from experience with the program. It's
not up to the task that I want it to do.


You can sit and struggle with trying to train yourself to receive 20
wpm Morse, or you can download and install CwGet and start copying the
high speed CW nets immediately. There's no longer any real need for a
human to be in the decoding loop, a sure sign of just how
anachronistic human-decoded CW really is. Samuel Morse originally
designed his code to be copied by machine, so in reality we're only
catching up with what he intended to do way back in the 1800's.

Already tried it. As I said while it is the best that is available, it is
still far below the capabilities of a human operator. I've tried it under a
wide range of conditions and CWGet still needs a pretty good signal to
function.

Dee, N8UZE

Morse Myth #119: All CW signals are good signals (Its the corollary of
Morse Myth #1: CW always gets through).