From: DL4YHF@aol.com
Message-ID: <117.14e608a5.2a707199@aol.com>
Date: Wed, 24 Jul 2002 17:09:45 EDT
Subject: Re: LF: DL2NDO/P (-> noise level estimation)
To: rsgb_lf_group@blacksheep.org
MIME-Version: 1.0
Content-Type: text/html; charset=windows-1252
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Sender: <majordom@post.thorcom.com>
Content-transfer-encoding: 8bit

<HTML><FONT FACE=arial,helvetica><FONT  SIZE=2>Hello Jim and group,
<BR>
<BR>&gt;&gt; You could also do a similar calculation using SpecLab to calibrate the 
<BR>signal level - here you have to take care not to adjust the RX gain (and 
<BR>keep the AGC turned off) between calibrating and measuring. To interpret 
<BR>the noise level you would have to take the bandwidth/resolution settings 
<BR>into account.
<BR>&lt;&lt;
<BR>
<BR>Agreed, if you are using the function "noise(f1,f2)" to do this. It returns the noise level as seen on the waterfall/spectrum, so it depends on the frequency resolution (or, in other words, the decimation- and FFT-gain). But there is another function noise_n(..) which does basically the same, but normalizes the noise level for a 1-Hz-RX bandwidth, taking the FFT bin width into account. 
<BR>Be aware to set the frequency range in which the noise is measured away from the edges of the IF filter's edge, and don't make the observed frequency range too narrow. You may find yourself observing the noisy slopes of hard-keyed (clicking) CW signals, etc....
<BR>By the way - the noise( ) - function is is based on an algorithm described by Andy 'JNT. 
<BR>
<BR>Regards,
<BR> Wolf DL4YHF
<BR></FONT></HTML>