WSPR Decoding Question

[cmh716]

I am trying to use KA9Q's built-in FT4/FT8/WSPR decoding in conjunction with ftlib-pskreporter to report to pskreporter. FT4 and FT8 are working well. For WSPR, I am building wsprd from source and jt-decoded is decoding. But, wspr.log is not revealing the actual frequency and is instead only reporting the offset, leading to incorrect spots being reported on pskreporter.

chris@hf:~$ /usr/local/bin/jt-decoded -v -w -d /home/chris/wspr wspr.local
 [0]/usr/local/bin/jt-decoded [1]-v [2]-w [3]-d [4]/home/chris/wspr [5]wspr.local


wsprd -f 0.021095 -w /home/chris/wspr/21095/251221_1400.wav
wsprd -f 0.013554 -w /home/chris/wspr/13554/251221_1400.wav
wsprd -f 0.010139 -w /home/chris/wspr/10139/251221_1400.wav
wsprd -f 0.003569 -w /home/chris/wspr/3569/251221_1400.wav
<DecodeFinished>
unlink(/home/chris/wspr/3569/251221_1400.wav)
<DecodeFinished>
unlink(/home/chris/wspr/24925/251221_1400.wav)
<DecodeFinished>
<DecodeFinished>
unlink(/home/chris/wspr/5365/251221_1400.wav)
unlink(/home/chris/wspr/3593/251221_1400.wav)
<DecodeFinished>
unlink(/home/chris/wspr/1837/251221_1400.wav)
<DecodeFinished>
<DecodeFinished>
unlink(/home/chris/wspr/13554/251221_1400.wav)
unlink(/home/chris/wspr/5287/251221_1400.wav)
<DecodeFinished>
unlink(/home/chris/wspr/28125/251221_1400.wav)
1400 -26 -0.0   0.022490  1  KK4INZ EL89 23 
<DecodeFinished>
unlink(/home/chris/wspr/21095/251221_1400.wav)
1400 -23 -0.0   0.011691  1  WR4I FM08 23 
<DecodeFinished>
unlink(/home/chris/wspr/10139/251221_1400.wav)
1400 -27  0.2   0.019566  0  KK5PY EM26 23 
<DecodeFinished>
unlink(/home/chris/wspr/18105/251221_1400.wav)
<DecodeFinished>
unlink(/home/chris/wspr/14096/251221_1400.wav)
1400 -17  0.6   0.008440  0  VE3KU FN03 20 
1400 -19 -0.0   0.008517  0  K4LXY FM08 23 
1400 -24  0.1   0.008542  0  AC9YY EN51 30 
1400 -23  0.2   0.008582  0  N9VP FM06 23 
1400 -21  0.0   0.008606  0  K6AUS FN42 30 
1400 -24 -0.0   0.008628  0  K8EB EN73 23 
1400 -12  0.3   0.008641  0  <...> EN81AP 23 

Am I incorrect in thinking that the -f parameter should be passing the actual frequency and not just the frequency offset? Spots are being uploaded as 0.008 MHz

[cmh716]

Should https://github.com/ka9q/ka9q-radio/blob/main/src/jt-decoded.c#L476 be

snprintf(freq,sizeof(freq),"%lf",(double)sp->ssrc * 1e-3);

-3 instead of -6?

https://manpages.ubuntu.com/manpages/jammy/man1/wsprd.1.html
-f x
x is transceiver dial frequency in MHz

[cmh716]

After changing to 1e-3, I get this. Seems wsprd adds in the offsets, which explains the odd numbers I had before. Am I doing something wrong or should we change the code to -3? Thanks

1502 -25  0.1  18.106602  0  W1VR EL98 20 
<DecodeFinished>
1502 -29  0.0  10.140408  0  WA0RXR EN42 23 
1502 -25 -0.1  10.140457  0  KC2STA FN22 23 
1502 -17  0.1  10.140468  0  K3XR FN10 23 
1502 -23 -0.1  10.140501  0  N9WKW EM69 23 
1502 -23 -0.1  10.140512  0  K4LXY FM08 23 
<DecodeFinished>
1502 -21 -0.2  14.097539  0  K4RCC EM85 23 
<DecodeFinished>
1502 -19  0.6   7.040400  0  VE3KU FN03 20 
1502 -26  0.3   7.040470  0  WB2TKP FN20 23 
1502 -23 -0.0   7.040481  0  WA8DZP EN82 23 
1502  -1  0.8   7.040501  0  KB8EZX EN91 37 
1502 -23  0.1   7.040501  0  W2RWM FN30 23 
1502 -19  0.4   7.040511  0  K1TTL FM05 43 
1502 -16 -0.1   7.040520  0  W9RAG EN54 23 
1502 -24 -0.1   7.040526  0  KC3SDJ EN90 23 

[cmh716]

#193

Please reject my PR if my analysis is wrong. Thanks, Phil

[ka9q]

I have been maintaining the FT4 and FT8 decoders in ka9q-radio but I've deferred to Rob, AI6VN, for WSPR as he runs the WSPRdaemon project. I have thought about putting a stripped-down wspr decoder (similar to those for ft4 and ft8) into ka9q-radio for those who don't want to run all of wsprdaemon. But that will take some work.

[kholia]

Hi Folks -

I would like to share a stripped-down WSPR decoder - https://github.com/kholia/wsprd/tree/main. It is derived from the excellent work done by Pavel Demin.

While it is a stripped-down and standalone WSPR decoder, the decoding performance remains the same (README.md in that repository has an comparison example).

PS: Thanks for this awesome project - I just started testing it yesterday.

[ka9q]

Thanks, I'll check it out.