Glenn --
If you do not have time or inclination to study the source code, you
should at least start with the User Guide.
Decoders in WSJT-X have advantageously used available "a priori" (AP)
information for many years. A whole section of the WSJT-X User Guide is
devoted to explaining what it does, and how to use it.
https://wsjt.sourceforge.io/wsjtx-doc/wsjtx-main-2.7.0-rc7.html#_ap_decoding
It's worth noting that aside from the "bare CQ" message, AP decoding is
not very useful for SuperFox messages because only a tiny portion of the
payload is predictable.
-- 73, Joe, K1JT
On 9/30/2024 1:20 PM, Glenn Williams via wsjt-devel wrote:
I have been out of the software engineering loop too long, so it would
be a chore to dig out the answers to the following questions from the
released code.
A suggestion for SF decoding?
I presume that each time the SuperFox transmits a CQ the message that is
encoded is exactly the same. I presume that portions of the signal
report and 73 message contain exactly the same encoding. I presume
therefore that the modulation for "exactly the same" data is also
exactly the same. Therefore once a SuperHound receives those sections
of messages, the software can "learn" what tones to expect and where in
the stream. So in the case of QSB the learned data could be compared to
(correlated with?) the most recently received audio and thereby an
improved decode presumption could be supplied.
--73, Glenn, AF8C
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
