Software is better for machine-generated CW. There are still a lot of
bugs/sloppy fists out there that skimmer can't copy. And there are
still some calls generated by computer/keyer that skimmer miscopies a
lot (LW3LPL and (K)B4JR, for example).
Watson beat the best players on Jeopardy. Should the game show go away?
Bottom line, CW was created as a "language" for humans. So what if a
computer can copy better? Jukka, did you type your email by hand or use
voice to text?
Barry W2UP
On 4/5/2016 06:28, Jukka Klemola wrote:
Software CW copying capability is judged - once again.
To judge human vs SW capability to copy CW with a more realistic data set,
please take a look at Dayton 2009 Pile Up contest results.
http://www.kcdxclub.com/pileup2009.html
Skimmer won.
The year was 2009.
Already back then, Skimmer was copying 61 calls while the best human
participants copied 51.
Then, look at who the human participants are on the score listing.
I think we can be satisfied with the result and declare the winner of the
human vs computer CW copying competition.
Looking forward from 2009:
My understanding is the software has developed by some amount after that
2009 event.
The human evolution takes much longer time to get another call from the
recording.
If you are interested in accelerating the human development, click
http://www.no5w.com/PileUpNetPractice.php
The service from KCDXClub and NO5W is excellent.
If you wonder about Skimmer, outside a laboratory grade environment, in a
real shortwave use scenario:
https://www.youtube.com/watch?v=bDVdwWfiAZw
73,
Jukka OH6LI
2016-04-05 14:07 GMT+03:00 Gordon LaPoint <gordon.lapoint@gmail.com>:
Hank,
...
NO software is as good as a person, but both of the above can help a
newcomer get started. Both of the above use the rig in CW mode with the
audio out to the computer sound card.
Gordon - N1MGO
On 4/4/2016 11:01 PM, Hank Greeb wrote:
I'm trying to help a relative newcomer to set up his rig for the MiQP.
He's just learning CW, we note that FLdigi can send and decode CW. I've
used FLdigi for Hellschreiber and PSK-31, but never with trying to decode
CW under contest conditions. From what I've encountered in MiQP and
similar contests, there's seldom a condition where "good ears" couldn't
work with a recevier 2 KHz bandwidth or so, which these software packages
generally use with a rig in the SSB mode.
Has anyone had experience with using it, or another "digital" software
package, in a simple CW contest like MiQP? It would help his multiplier
count if he could make some CW contacts with an aid like FLdigi.
Please don't "put this question down" - I'd like to encourage the newbies
to contesting to give it their all, and use all the technology available.
72/73 de n8xx Hg
QRP >99.44% of the time
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
_______________________________________________
CQ-Contest mailing list
CQ-Contest@contesting.com
http://lists.contesting.com/mailman/listinfo/cq-contest
|