Hi Joe and All,
Many thanks to Mikhail R3LW, who pointed me to this issue.
In order to reproduce it one shall tune FTRSD decoder to maximum sensitivity
for second pass:
1st pass thresh0=2.4
2nd pass thresh0=1.0
_________________________
Test #1, original audio file brings these JT65 decodes:
2227 -15 0.8 1262 # CQ IZ0MIT JN61
2227 -21 1.0 1360 # CQ IK2HTY JN45
2227 -15 1.3 1626 # ZP9CTS HA3LI JN96
2227 -9 0.8 2030 # CQ I1RJP JN45
2227 -14 0.9 2128 # JF3SUL SQ5PFF 73
2227 -21 1.3 2320 # E9DX YT1DL KN04
__________________________
Test #2, 800...1000 Hz band pass filter is applied, now modified audio file
brings additional decode:
0000 -25 0.9 2169 # W4MMP WB9OTX -01
__________________________
Test #3, 600...800 Hz band pass filter is applied, now modified audio file
brings additional decodes:
0000 -20 1.6 1962 # KJ3N 9A6CC -20
0000 -25 0.9 2169 # W4MMP WB9OTX -01
__________________________
Test #4, 5700...6000 Hz band pass filter is applied, now modified audio file
brings additional decodes:
0000 -27 1.1 799 # ER1PB PD2HAB JO22
0000 -25 0.9 2169 # W4MMP WB9OTX -01
All these tests raise question on correctness of the FFT implementation in the
WSJT-X code.
For repro I used thresh0 modified r6373, JT65+JT9, Random Erasure Patterns=8.
Bandpass filter has 30dB rejection outside it's bandwidth.
I have put audio files there: https://cloud.mail.ru/public/2FYq/iTExXKrnp
73 Igor UA3DJY
------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
wsjt-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wsjt-devel
