[Lowfer] WSJT-X 1.9 v 1.8 second night

[Garry]

Paul,

I ran a parallel WSJT-X1.9 vs 1.8 test last night too, listening mainly 
on 474.2 kHz (1333 decodes) but also copying some signals on 136.0 kHz 
(28 decodes) and 74.1 kHz (16 decodes).

I found the difference to be NIL for SNR on the single station active on 
74.1 kHz and only 3 of 16 decodes differed from zero DT.

On 136 kHz I copied 4 different stations whose SNRs agreed exactly on 27 
of 28 cases (1.9 decoded one case at SNR=-31 dB that 1.8 missed).
There was a small DT difference in just one of the 27 shared cases.

On 474.2 kHz 1.8 decoded 2 signals at SNR=-32 dB while 1.9 decoded them 
and 3 more. At SNR=-31 dB 1.8 decoded 5 cases while 1.9 decoded them and 
3 more. Version 1.9 decoded 1333 cases in common with 1.8 but also 
included 18 more. Both programs erroneously decoded the call VG4FSE once 
and each program had one more bogus decode (ED5/086JQX for 1.8 and 
RSB/BP1LSV 1.9).

Overall the decoding ability of 1.9 was only slightly better than 1.8, 
but it can pull through weak signals that otherwise would be missed. 
Earlier in the afternoon one station not a part of the current data set 
regularly decoded via ground wave at -32 dB SNR with 1.9 while 1.8 gave 
nothing. And of course there are other reasons like features, ease of 
use, more information displayed, etc. to make 1.9 a desirable upgrade.
-- 
Garry, K3SIW, EN52ta, Elgin, IL