I played with watching the difference and I don't see any difference if I
have the Xonar at -58dB or 0. The waterfall looks exactly the same. I
don't see anything that doesn't decode either way. Have to wait and get
some low-level signals to test some more. Too band I can't run it
side-by-side as doing it this way feels subjective.
Part of what you said here echos what I was thinking.
Earlier you said it would nice to have the WSJT-X slider in the middle.
Isn't that all digital gain? Wouldn't you want that closer to zero?
I see the ingain getting passed to symspec and hspec_ so what is the
purpose of changing the scale inside symspec?
Does this linear rescaling in symspec really matter?
RRR
Mike W9MDB
On Tue, Jan 19, 2016 at 7:36 PM, Bill Somerville <[email protected]>
wrote:
> On 20/01/2016 00:53, Joe Taylor wrote:
> > Sure, we could set the level higher. WSJT-X works OK with the noise
> > level set anywhere between about 20 dB (rms=10) and 60 dB (rms=1000).
> Hi Joe & all,
>
> so it will be interesting to for Mike to report back what difference he
> sees now he has removed what appeared to be ~60dB digital attenuation
> from his system and probably a downstream ~60dB digital gain to compensate.
>
> I will try and summarise a couple of points:
>
> a) having the ADC clipping due to excessive audio level is bad, once the
> sampling circuit is driven to the maximum count it is no longer linear
> and cannot express the cumulation of the many cosine waves that are the
> signals and noise. Jim points out that the peak levels from radio
> receivers are often impulses so the slow modes like JT9 and JT65 are
> very tolerant of short duration sampling non-linearities but we are
> still better off without them. We should strive to feed just enough
> audio level to the ADC but not too much.
>
> b) Joe points out that the one LSB quantization error of the ADC is
> swamped by the receiver and band noise but if we have our input audio
> level to low and have to apply digital gain after sampling to get a
> usable signal then for every dB of digital gain applied we are
> increasing that quantization error by the same amount. That's is OK up
> to the point where the quantization error is no longer buried in the
> noise, then a threshold is crossed and the wanted signals have their
> signal to noise ratio reduced.
>
> 73
> Bill
> G4WJS.
>
>
> ------------------------------------------------------------------------------
> 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
>
------------------------------------------------------------------------------
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
