G
4X1RF and GW0EZY took a recording of my WSPR transmission and sent it to
me via dropbox, a wav file recorded in SpecLab.
I run the recordings in an endless loop and produced a set of decodes, 5
decodes for each setting, just to make sure that there is no random
difference in the S/N shown for the same settings. Mostly the S/N shown
was the same for a certain setting, maybe 1 of 5 differed by 1 dB. The
results were quite clear. I've done the test with strong signals
(GW0EZY) and weak signals (4X1RF). There average improvement was 2 dB,
maybe a bit more.
73, Stefan
Am 03.10.2013 20:45, schrieb Graham:
How did you measure the 2 dB Stefan ?
--------------------------------------------------
From: "Stefan Schäfer" <[email protected]>
Sent: Thursday, October 03, 2013 6:47 PM
To: <[email protected]>
Subject: Re: LF: WSPRX_08r3575
Hi Tobias,
Am 02.10.2013 16:19, schrieb Tobias DG3LV:
Hi Stefan/LF !
Information on the SVN-server says, that there was a noise blanker
added.
Thanks for all the infos.
Finally :-) During the time he developed WSPR-15 we were in regular
email contact. To that time i made some tests with DF6NM's slow WSPR
version, tests with 4X1RF and GW0EZY and others. We found that the
SpecLab noise blanker in front of the WSPR input (using VAC) made a
S/N improvement of at least 2 dB. Obviously he anyway didn't include
such a tool (which is much more useful on LF/MF as on the HF bands i
think) in the previous versions.
Now the question is if the new intenal NB is more efficient than an
external SpecLab NB. One could do a test running 4 WSPR-x instances,
the old one, with and without a SpecLab NB and then the new one, with
and without a SpecLab NB... If the new one without a SpecLab NB
performs best, then this would be really an improvement over the
older version! If i can find the time i will do such a test at night.
73, Stefan/DK7FC
|
