G2HDQ wrote:
Apart from the obvious i.e. the various decoders with text output, I would
put software such as Spectran (VY NICE program, by the way!), Spectrogram,
etc. in the "machine modes" category. This is because it takes a wanted
audio signal which is buried in QRM/QRN and therefore indistinguishable,
and translates it into a visual form, separating it from the unwanted
stuff as a discrete trace on a screen. The only part the brain is required
to play in this process is to interpret the trace as letters and numbers -
most of the hard work has been done already.
See my other message. If the only work was to interpret the trace
as letters and numbers, it would be trivial for a bit more coding to
be included to do just that - and of course 'strong' signals (ie those
that are louder than a mere 10db below noise) come into this
category. However, it is quite possible to dig much deeper than this
by using the brain to work out what is rubbish and what is signal.
Most machines are quite incapable of ignoring static and other
noise.
In the few reports I have read about the various PSK modes, they
appear to have similar error rate to CW, which is very impressive for
something read by a machine (conventional CW read by a machine
is terrible in the presence of QRM), but at sub-noise levels the error
rate increases. This is not the case with visual CW modes for the
reasons mentioned above. Of course, very low bit rate PSK may
well be a different matter, and that is really comparing like for like.
Mike, G3XDV (IO91VT)
http://www.dennison.demon.co.uk/activity.htm
|
