Hello Wolf, all,
Yes, that's why i suggested to allow input frequency ranges of 0...24
kHz. There could be a text field where to type the desired center audio
frequency and the spectrogram window automatically centeres to that
frequency...
I also kept in mind that there may be a time delay in the SL filter so
i used "only" 4096 FFT points for the filter.
A few minutes ago i continued to attempt to decode DF6NM and actually
there have been several decodes at -12...-15 dB SNR. But as soon as i
opened the remote control software (Teamviewer) to attach the TX
notebook to start a call for him, there were no decodes from that time.
A restart didn't change the situation. My time server program is
updating each minute or each 5 minutes likewise. I cannot see a
systematic behaviour why there are decodes and in the next moment they
fail. As said, in WSPR i never observed any problem with non decodes,
time delay or CPU problems on the same PC...
Now i saw PA0A calling a few minutes ago. Albert, if that was you,
could you go on with a few sequences? I saw you clearly in the
waterfall but restarted the program and later you stopped...
73, Stefan/DK7FC
Am 02.11.2012 15:54, schrieb wolf_dl4yhf:
Just some guesswork:
We may need an option to compensate the delay between audio input (to
the soundcard), and the time read by WSJT-X from the PC's system clock.
Since Stefan uses Spectrum Lab as an 'input preprocessor', plus a
virtual audio cable, the total delay may be too much (in SL, very long
FFTs and inverse FFTs are sometimes used for audio processing, which
introduces delays over one second. The ultimate solution would be to
feed timestamped data into the WSJT decoder, but a simple input field
(in 'seconds added to the PC's system time) to compensate such a delay
manually would also be ok.
Here, with WSJT-X directly connected to the soundcard's input, I
receive some overs with almost perfect sync but no text decoded.
Unfortunately wave-file-logging was turned off at that time:
73,
Wolf DL4YHF .
|
|
