Audio resampling in Digiu mode

[neiljw...@gmail.com]

              

Hi all,

I was recently trying to receive some QRSS30 signals on 136kHz using a high resolution waterfall (in Spectrum Lab) fed from an HL2 in digiu mode.

At times the transmissions would appear as if the frequency  was

varying up from the nominal by up to 2-3Hz. At other times the frequency was very stable.

At first I though the transmission was the culprit, as I had never seen this effect before.

But I suspect I had never looked at the frequency with this level of resolution before!

After changing to a different HL2 (which did the same) and a lot of head scratching I wondered if the audio resampling might be on and making a difference.

Sure enough, turning it off removed the effect completely.

Screenshot attached, showing a good section followed by a problem section.

As you can see it makes copying the QRSS30 rather difficult.

This is a rather extreme situation - I think few people will have needed this resolution

in normal operation. Just thought I would mention it in case anyone else was as puzzled

as I was. It might also cause problems with the slower digital modes like FST4W as the

audio frequency is wandering more than the tone spacing. It shouldn't be a problem with

the built-in modes but may be if feeding a digiu output to WSJT-X.

In summary, if you are using digiu to feed audio to a high resolution or small tone spacing application, turn off audio resampling.

73,  Neil  G4BRK

[ahop...@googlemail.com]

Hi Neil,

that is interesting, what version of spark was it, I made a small tweak to the resampler in the very recent test versions .  The real solution to this is to convince the digi mode programs to adopt some protocol other than vac audio to transfer audio or iq.  Where there any ep6 errors when this was happening?

73 Alan M0NNB

[neiljw...@gmail.com]

Hi Alan,

I'm using 2.0.33.0 on Windows 10.

I get a few ep6 errors, maybe 1500 in a day on this instance as it is quite busy.

73, Neil  G4BRK