CW Decoding

  • 16
  • Idea
  • Updated 1 year ago
  • Under Consideration
Yep, I know we can get this through a 3rd party product - CW Skimmer -- but with the ICOM 7300 and Elecrafts providing this natively, why not add it to the feature list of the 6000 series as well?  It would be so cool to just click a tab and have a CW decoding window pop up at the bottom of the screen.
Photo of VE7ATJ

VE7ATJ

  • 106 Posts
  • 20 Reply Likes

Posted 2 years ago

  • 16
Photo of Al / NN4ZZ

Al / NN4ZZ

  • 1678 Posts
  • 573 Reply Likes
Good idea....there is an idea to add spots from CW Skimmer to the SSDR display....

Adding a decoding window is a good addition and I added your comments there too.  (hope it's okay)

Here is the link: 

https://community.flexradio.com/flexradio/topics/pfskimer_what_is_the_possibility_for_a_smartsdr_dis...

Regards, Al / NN4ZZ  
al (at) nn4zz (dot) com
6700 - HW.................. V 1.8.4.84
SSDR / DAX / CAT...... V 1.8.4.168
Win10


 
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 642 Reply Likes
That was actually part and parcel of a question I asked the other day that now I had the DAX audio now what? That was actually a multipart question as it isn't clear what one doublet of information represents, is it temporal or spatial? In other words, is it a point in time and/or is it a point in space?  I am thinking it is a point in time (temporal).

We tend to think of a spectrum display having real estate or an oscilloscope showing a wave over time but, as I pondered this question I concluded it is a sliver of time where the trace would be over time had it had a X axis.

I am thinking therefore that two channels is likely superfluous as are, for the sake of CW, the negative values. Collecting what's left over, and time stamp them, we'd have a series of amplitudes, some short, some long, some really long. The really long spacing of low amplitudes is the inter-word spacing, the really short lived high amplitude values are the dots and the intermediate lived high amplitudes are the dashes. Then just look up the values! Where one could get creative is in filtering out noise and adjusting for legacy keyers. I did try to google to see if I could find published articles on signal processing. As I believe I mentioned to Howard, I really enjoyed physics classes but never took the advanced ones where I am sure this sort of subject came up.                                                                                                                 
Photo of W5XZ - dan

W5XZ - dan

  • 562 Posts
  • 83 Reply Likes
like, some software than can pull cw out of the noise better than the human ear?

73, w5xz, dan
Photo of KK9W - Steve

KK9W - Steve

  • 552 Posts
  • 173 Reply Likes
I would like it as I have some hearing loss and have a hard time with it.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 642 Reply Likes
Ag1le blog - OK, I feel a decoder in my future.

Darn...no, I don't believe my previous is entirely correct. The DAX signal represents everything in the passband. In the case of a 100Hz CW filter, yes, one can conclude there is but one discrete signal there but, let's say you opened it up to 3KHz on 14.070 (or .100) for psk. In that case there are more than one active signals such that, over time, multiple concurrent signals will be oscillating around 0. My head is about to explode.
(Edited)
Photo of Peter

Peter

  • 79 Posts
  • 5 Reply Likes
I have wondered the same thing seeing this to be basic needed feature that can be easily implemented instead of depending on third party tools. Guess FRS don't see this as necessary at the moment, or maybe implemented in the paid ver. of SmartSDR!

Peter 
Photo of ka7gzr

ka7gzr

  • 212 Posts
  • 36 Reply Likes
The IC-7300 has a native CW decoder? It certainly has RTTY but mine doesn't have a CW decoder.
Photo of KM6CQ - Dan

KM6CQ - Dan

  • 156 Posts
  • 33 Reply Likes
This would be an incentive to purchase 2.0 

Dan
Photo of James Whiteway

James Whiteway

  • 856 Posts
  • 186 Reply Likes
I thought real hams knew code ! No need for a decoder!
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 3875 Posts
  • 1179 Reply Likes
Many hams know code, James.  
I even passed 20 wpm the hard way, after years of frustration and fighting, no less...

But just last night I was trying to bust a pileup on 40 Meters, and the guy was sending at 45 wpm!  In my book, there is not reason in the world for a DX pileup to run that fast.  I couldn't even have copied my OWN call at that speed.
That is why I would like to find a top-notch decoder to assist my CW skills.....
Photo of Rick Hadley - W0FG

Rick Hadley - W0FG

  • 556 Posts
  • 118 Reply Likes
You must mean TU5MH.  He was actually running 38-40 wpm.  I speed matched him with CWX and got him at 02:14. I can't carry on a full QSO at those speeds, but can copy calls.  Very unusual to find a DXpedition running at those speeds, but it may have been a way for him to reduce the pileup.
Photo of Kevin K4VD, Elroy

Kevin K4VD, Elroy

  • 775 Posts
  • 171 Reply Likes
There is no CW decoder that I found that is accurate enough to be embedded into the Flex radio.

To me, a skimmer show's who's out there and, only secondarily, what they might be saying. I think skimmers take advantage of one of the same features the human brain uses to decode a fast contest or DX exchange... anticipation and format. Adding skimmer capabilities to SmartSDR seems to fit very nicely. Relying on the radio or SmartSDR to decode code, less than useful fit.

In a DX exchange I expect to hear my callsign and 5NN. Send it at 50+ WPM and I can still pick out my callsign and the RST because I expect to hear it and I hear the format the DX station is using. I couldn't have a conversation with the DX station at those speeds but then again, everyone else in the pileup probably appreciates that.

I have not found a single decoder (cw-get, skimmer, KX3 built-in, and various commercial and homebrew hardware kits) that have not actually impaired my ability to copy by ear because of their inability to copy anything but a perfect signal.

Congrats on TU5MH Rick.

It's just me.

Kev
Photo of Ken - NM9P

Ken - NM9P, Elmer

  • 3875 Posts
  • 1179 Reply Likes
Rick, perhaps it reduced some of the pileup because so many may have given up.  But it probably also added some to the congestion because so many people could not be sure what call he was sending and continued to throw theirs in just in case...

According to Reverse Beacon Network and he was anywhere between 40-45 when I was trying to copy him.  I don't know how accurate RBS reports are.  In any case, it was faster than I ever hear on 99% of the DX pileups, especially on the noisier lower bands.

The combined effect of the high speed, weighting/ratio he was using, rapid fading, noise, and water/auroral effect on his 40 meter signals, the high speed code was extremely difficult to distinguish at my QTH in Southern Indiana.  

I have fought many contests and DX pileups in the 28-35 WPM range successfully, but this was a particularly difficult copy.  I eventually gave up with plans to try again when the lightning is further away.....perhaps after working the big guns he will slow down to about 32 wpm, which I have read many places is about the optimum compromise between speed efficiency and needed repeats due to excessive speed.  Others may differ on that point.

Ken - NM9P
Photo of Rick Hadley - W0FG

Rick Hadley - W0FG

  • 556 Posts
  • 118 Reply Likes
I have to agree with Elroy.  I've never found a CW decoder that worked well on anything except computer generated code.  As noted, CWS does a pretty fair job on callsigns, but I consider it unusable for anything else, which is OK as CW was always designed to work with wetware...lol.  As far as TU5MH goes, I had pretty good band conditions and found him remarkably easy to copy, given his speed.  I had plenty of time to listen to him, as I'm just running barefoot to a vertical on 40, so it took me a while to snag him.

Rick, W0FG
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 642 Reply Likes
In fairness, you also purchased 1.x. It's cost was embedded in the price of the radio. Should someone buy a 6000 next year, per my understanding, that $200 is part of the cost of the 6000 next year. That purchaser will be buying 2.x rather than 1.x everyone here purchased already.
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2354 Posts
  • 535 Reply Likes
I think from the beginning of SSDR the idea was to build up a library of API software to plug into the SSDR. This makes things we can do with our radios almost endless. The idea of 3rd party software working with the Flex is the main idea of owning a Flex radio.

The SSDR has been made, taking thousands of hours to get to a fully working system. To mostly start over to add a new software into it will not happen I'm sure.

And this is were features we could purchase comes in. Ater V2.0
(Edited)
Photo of KK9W - Steve

KK9W - Steve

  • 552 Posts
  • 173 Reply Likes
No biggie as I use CW skimmer to assist me.
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2354 Posts
  • 535 Reply Likes
Me too Steve, guess I'm not a real ham,,oh well...
Photo of Bob G   W1GLV

Bob G W1GLV

  • 639 Posts
  • 107 Reply Likes
I could never understand why you had to decode morse code to become a ham. That requirement caused a lot of potential hams to walk away from the hobby. I guess it came from the railroad days.
Photo of KY6LA - Howard

KY6LA - Howard, Elmer

  • 3471 Posts
  • 1335 Reply Likes
It actually was a vestige of WW2 where we needed trained CW operators and later fears over Sputnick. I agree that the requirement was a huge mistake that almost killed the hobby.
Photo of KK9W - Steve

KK9W - Steve

  • 552 Posts
  • 173 Reply Likes
I used to copy 40 wpm still can on a real strong signal but with age things start to go like hearing and eyesight. Thank god that's it. That narrow bandwidth just does not play well with me.
Photo of Peter

Peter

  • 79 Posts
  • 5 Reply Likes
What is needed is a QRM filter to filter out the arrogance on this thread. The hobby should be fun, not work, I am lazy and too old to copy CW....I would rather read it that try to guess what some VERY CW operators are sending.

Peter
Photo of Bill -VA3WTB

Bill -VA3WTB

  • 2121 Posts
  • 503 Reply Likes
Peter, the best filter is between our ears. For a CW nut, this thread should have little interest to them to begin with.
Photo of Rick Hadley - W0FG

Rick Hadley - W0FG

  • 556 Posts
  • 118 Reply Likes
I'd like to see a CW decoder that really works.  The one in CW Skimmer is only passable, and it's far better than the one in DM780.  I don't know how many DX Cluster spots I've sent out correcting callsigns that somebody's decoder misread.  As far as hearing goes, my failing 70-yr old ears still hear CW much better than SSB, or my wife across the room...lol.  All one needs for CW is sidetone set to a pitch that's still within ones auditory comfort zone and a good set of NR cans that block out extraneous sounds.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 642 Reply Likes
Totally agree with that Rick. I think DM780 is darn near worthless as it appears wrong more than it's right. CW Skimmer has gotten better over time. About a yr or so ago, when Al was espousing it's virtue, I found it to be nonsensical in decoding. I tried it again (different version) maybe six months ago and it was markedly better than I remembered it. But yes, I think the science is still in its infancy.
Photo of K3DCW

K3DCW

  • 55 Posts
  • 17 Reply Likes
MRP40, from Polar Electric, has probably the best software-based CW decoder I've used.  It does a surprisingly good job copying hand-sent CW, but it is not a skimmer; it works best with single signals. 

That being said, there's always some really poor CW that nothing other than the human ear and experience could copy. 
(Edited)
Photo of Stan - VA7NF

Stan - VA7NF

  • 401 Posts
  • 86 Reply Likes

In fairness there are many groups and subgroups of hams. Full disclosure: I'm in the contesting/ARES main group and CW lover subgroup with a close tie to administrative as President of the Surrey Amateur Radio Club.

When I contest RTTY there are two decoders running on screen and I (N1MM) click on the most consistent callsign that appears.  This may be applicable to CW. 

(See bias above) CW is just a more efficient digital mode; having two or more (clickable) on screen decoders running may, optionally, be supplemented by the between the ears decoder.  The choice becomes the best of three CW decodes (no minority reports planned).

Walt, Please avoid brain explosions BUT to be most effective the latency must be very low, as a contesting run reply has ONE repetition, unknown speed, unknown callsign rules, on or off frequency, (with diversity) locational stereo phase relationship, in passband or adjacent, then tonal separations of only a few hertz, and don't forget propagation phasing and the infamous chirp.

The best place to do all this is in radio - But no way I would suggest FRS move their radio server development into CW decode (Editorial: historically CW is not a FRS strong point).  That leaves 3rd party DAX IQ development with possible supplement of video card (look up CUDA) array processing, coupled with a FRS option to add rules for 3rd party panadapter "leased screen space".  That is brain exploding!

My in-brain CW decoder is multi-threaded: 1) Variable speed up to 60WPM conversion to a virtual/visual "ticker tape", correctable as a source letter is formed. and 2)  A mental "ticker tape" reader that can correct for errors in 1) and mis-sends.  This system fails below 20WPM and must use paper intermediate.

Stan

Photo of Ned K1NJ

Ned K1NJ

  • 301 Posts
  • 75 Reply Likes
  
    Let Flex concentrate on the radio technology.  That's what they do best

Ned,  K1NJ
Photo of AA0KM

AA0KM

  • 287 Posts
  • 50 Reply Likes

Computer generated CW and received generated CW would be another digital mode.

:)-  :(