DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=googlemail.com; s=gamma;
        h=mime-version:date:message-id:subject:from:to:content-type;
        b=BjlblmqyCOWCyTVbZbD1DLWyk8Y0rEhmtpBnSQ2h4Dzp84oCzxPaF4gYcwC4LDJ40Y
         a/zrdG1Thnsr/Ixgaedk+k2UA5rWA6UIpICMYccWtV+/zypxL/6/cJD9mAxT6qB035zX
         00p0D3Ht0WByJ7bLBYudgQNmwtuZOHsuuxX0M=
MIME-Version: 1.0
Date: Sun, 10 Jan 2010 23:18:43 +0000
Message-ID: <fc7eccec1001101518k466006b8gfa63e4d2f30bf4fa@mail.gmail.com>
From: Andy Talbot <andy.g4jnt@googlemail.com>
To: rsgb_lf_group@blacksheep.org
DomainKey-Status: good (testing) 
Subject: LF: CW S/N abilities
Content-Type: multipart/alternative; boundary=001485f795f44ef961047cd7a40e
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


--001485f795f44ef961047cd7a40e
Content-Type: text/plain; charset=ISO-8859-1

A serious appeal to all the CW and QRS(S) ops on this reflector.

I'm interested in what can *really* be achieved in terms of readability of
weak CW .  For the calculations mentioned earlier I made the assumption the
best listeners can *read* 20 - 30WPM in 10dB S/N in a 30Hz bandwidth,
without repeats, which equates to -9dB S/N in the standardised 2500Hz. It is
possible to detect a tone significantly lower than this, (I can positively
detect a 600Hz tone at -25dB S/N in 2500Hz, but wouldn't know immediately if
it was switched on/off)  but first-time readability is essential.   If
anyone can measure S/N accurately and have measured your CW receiving
capability, it would be interesting to get confirmation or otherwise that
this assumption is not too-far-out.
Same applies to QRS(S) although this I suspect is less amenable to any
additional processing gain from learning and experience.

S/N can be determined from any spectrogram software offering  a spectrum
rather than waterfall display.    Note the height of the tone spike above
the average noise level;  this is the S/N in the bandwidth of the FFT Bin
Determine this value if it is not stated as a 'resolution' or any similar
name.
If not given specifically, resolution BW can be calculated from Sampling
rate divided by FFT size.  Eg 11025Hz sampling with 4096 point FFT gives a
resolution of 2.7Hz.
The normalised 2500Hz value is then obtained from S/N [2.5k] =  S/N
[measured]   -  10.LOG (Resolution BW / 2500).
If you can't measure S/N, record the marginal you've 100% copied and I will
measure it.

Please be honest about this.  If some CW experts clain they can out-perform
any appliance operator, lets see the figures that prove it.

Andy
www.g4jnt.com

--001485f795f44ef961047cd7a40e
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div>A serious appeal to all the CW and QRS(S) ops on this reflector.=A0=
=A0</div>
<div>=A0</div>
<div>I&#39;m interested in what can <em>really</em> be achieved in terms=
 of readability of weak CW .=A0 For the calculations mentioned earlier=A0I=
 made the assumption the best listeners can <strong>read</strong> 20 - 30W=
PM in 10dB S/N in a 30Hz bandwidth, without repeats,=A0which equates to -9=
dB S/N in the standardised 2500Hz. It is possible to detect a=A0tone signi=
ficantly lower than this, (I can positively detect a 600Hz=A0tone at -25dB=
 S/N=A0in 2500Hz, but wouldn&#39;t know immediately if it was switched on/=
off) =A0but first-time readability is essential.=A0=A0 If anyone can measu=
re S/N accurately and have measured your CW receiving capability, it would=
 be interesting to get confirmation or otherwise that this assumption is=
 not too-far-out.=A0 </div>

<div>Same applies to QRS(S) although this I suspect is less amenable to an=
y additional processing gain from learning and experience.</div>
<div>=A0</div>
<div>S/N can be determined from any spectrogram software offering=A0 a=A0s=
pectrum rather than waterfall display.=A0=A0=A0 Note the=A0height of the=
 tone spike above the average noise level;=A0 this is the S/N in the bandw=
idth of the FFT Bin=A0=A0 Determine this=A0value if it is not stated as=A0=
a &#39;resolution&#39; or=A0any similar name.=A0 </div>

<div>If not given specifically, resolution BW can be=A0calculated from Sam=
pling rate=A0divided by=A0FFT size.=A0 Eg 11025Hz sampling with 4096 point=
 FFT gives a resolution of 2.7Hz.</div>
<div>The normalised 2500Hz value is then obtained from S/N [2.5k] =3D=A0=
 S/N [measured]=A0=A0 -=A0 10.LOG (Resolution BW / 2500).=A0=A0 </div>
<div>If you can&#39;t measure S/N, record the marginal you&#39;ve 100% cop=
ied=A0and I will measure it.</div>
<div>=A0</div>
<div>Please be honest about this.=A0 If=A0some CW experts clain they can=
 out-perform any appliance operator, lets see the figures=A0that prove it.=
</div>
<div><br clear=3D"all">Andy<br><a href=3D"http://www.g4jnt.com">www.g4jnt.=
com</a><br><br></div>

--001485f795f44ef961047cd7a40e--