From: "Clemens Paul" <cpaul@gmx.net>
To: <rsgb_lf_group@blacksheep.org>
References: <095C9D4E586E45CAA3FE071EA4EA9818@HPG5401fr> <53E0F872.3030500@gmail.com> <64A3B1F0647248229C5C374887AA2FCB@F6CNI> <53E1EC0B.2060309@gmail.com> <1686C40AF97B4ABE99024972373C232A@HPG5401fr> <96662FC1BCA8402098A702D935EEC66F@F6CNI> <9556CA31D58145AFA57960399F67FE98@Clemens0811> <op.xkab16k7yzqh0k@pc-roelof>
In-Reply-To: <op.xkab16k7yzqh0k@pc-roelof>
Date: Fri, 8 Aug 2014 23:48:58 +0200
Message-ID: <EBC2CBB7132449978430A2D6E0A5FDED@Clemens0811>
MIME-Version: 1.0
Thread-Index: Ac+zMhsDGT5vp9E3SzW0TuXzFS/0UgAHrJhA
Subject: RE: LF: RE: Re: Testing now a new RX antenna on MF
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: O

Hi Roeloff,

while 0dB SNR referenced to the RX BW is commonly defined as the
"minimum discernible signal" 
I agree with you that with the listening skills of an experienced operator
and low cw speed 10dB below "MDS" are possible under certain
circumstances which you have described.
Just played again with G4JNT's genious utility SNDemo
http://www.g4jnt.com/DownLoad/SNDemo.exe
and it confirms what you are saying including
that 300Hz pitch is easier to decode than 500Hz or even 800Hz.

73
Clemens
DL4RAJ 


>-----Original Message-----
>From: owner-rsgb_lf_group@blacksheep.org 
>[mailto:owner-rsgb_lf_group@blacksheep.org] On Behalf Of Roelof Bakker
>Sent: Friday, August 08, 2014 7:55 PM
>To: rsgb_lf_group@blacksheep.org
>Subject: Re: LF: RE: Re: Testing now a new RX antenna on MF
>
>Hello Clemens,
>
>> E.g., if your FFT is set to a bin width of say 5Hz and a 
>carrier lies  
>> 20dB over
>> the DANL (displayed average noise level) it will have a SNR 
>of zero dB  
>> in a
>> 500Hz "IF"-bandwidth and therefore will be barely audible.
>
>The brain of a skilled listener can simulate a bandpass filter with a  
>width of 50 Hz.
>Hence your FFT bin width of say 5Hz and a carrier of 20dB over 
>DANL will  
>be clearly audible.
>Unfortunately the human brain needs white noise to work 
>against, so a band  
>full of interfering signals will exhaust the listener before 
>long. On a  
>quiet band it works great.
>This problem is solved by using a filter bandwidth of 10 - 20 
>Hz for aural  
>copy of Morse Code.
>
>I believe that the "human brain filter bandwidth" is frequency 
>dependent.
>When I started chasing NDB's, a beat note of 500 Hz was used.
>Over time this has changed to 300 Hz!
>Such a low beat note can be a problem with an analogue 
>receiver due to  
>less than perfect filter skirts. With SDR's the problem does not exist.
>
>73,
>Roelof, PAoRDT
>
> 
>