[Elecraft] CW Decoding - Your Brain is best

[Dyarnes]

Hi All,

I'm not going to totally throw cold water on those trying to rely heavily on 
the CW decoding capability of the K3 (and KX3), but I think you will be 
mostly frustrated.  First of all, the decoding capability of these rigs is 
heavily dependent on several factors, including a pretty good signal and 
properly sent characters (including spacing).  Even if the characters are 
well formed, you can still think you are getting some gibberish if the 
spacing isn't correct--you have to focus on properly separating the 
characters, and some ops really make that a chore!

I do appreciate the capability these rigs have, and even use it once in a 
while--very seldom though.  One example is when I run across a handful of 
ops on 40 meters who like to gather nearly every day and play "burn out"!  A 
couple of these guys are going 40 wpm plus, and that is a bit too fast for 
me--I get a lot of it, but not all.  Fortunately, they tend to send pretty 
good code (I am suspicious that some of them may be using keyboards), so I 
can "fill" in what I miss from the code reading feature on my K3 and KX3. 
I'm not participating in the conversation, but use it for code practice.

Probably the biggest benefit of the code reader is when I am showing off the 
radio to someone who doesn't know code.  Not only can they see what I am 
hearing (at least most of it), they can see what I am sending as well.  The 
abbreviations we use on CW will throw them a bit, but they tend to get the 
gist of the conversation--I can explain the "shorthand" later.  This really 
is helpful!  An observer's eyes can glaze over pretty quickly when all they 
hear is a bunch of dits and dahs that have no meaning.  However, when they 
can see that real information is being passed back and forth, and they can 
interpret it, their attention span, and interest, is much greater.

I applaud Wayne and Lyle's efforts to try and make the code reading 
capability better, but I am somewhat pessimistic that they will be able to 
make substantial improvements.  I say that primarily because I think the 
deficiencies of the sender are apt to be too great, and too variant, to 
really overcome.  Too many ops out there just don't seem to be motivated to 
try and emulate machine quality code.  I don't know, from a technical 
standpoint, what adjustments Wayne and Lyle are trying to make, so I can't 
really say with any certainty how successful they will be.  Perhaps they can 
get the reader to accept variations in the "1 to 3" ratio more readily. 
That could help I suppose.  I think they inferred that was one objective. 
However, I don't know how you can ever overcome most of the spacing issues. 
That's my biggest problem in just trying to copy CW, and it seems to be 
where most code reading software tends to go sideways.  If someone is going 
to insist on running characters together, only the human brain can probably 
figure that out--eventually!  If you are dealing with weak signal problems 
I'm not sure what improvements can be made.

I like the code reading capability on these rigs--it's clever and sometimes 
beneficial.  However, I see very little promise of it becoming something an 
operator can rely on primarily.  Your best chance of doing so is if 
keyboards are being used.  However, if you do that, without being able to 
translate things yourself, it really does become "just another digital 
mode", and not much different than using something like PSK31.

Dave W7AQK