DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=Kc3d1CkzliMLHU5JsRvJZfDlKoqC8e5/noupKH1VjfFqQyj+3asgOv/lR8TASLgiKc
         LZpFnpPjpRfOsF9ZpBdXZ2ShBJhuHvkM0khtcEDWwv+vZV9rilrjpXc/gLfy09fZdM2d
         O/yga7pERkuU1SyrDOOa0SnFMJbqyDKFLuUQc=
MIME-Version: 1.0
In-Reply-To: <4D146CA3.7090708@online.fr>
References: <4D122CE9.26517.19CC01F@mike.dennison.ntlworld.com>
	<BF4A524700075746A6467658DFC7102C88715AE9BD@ICTS-S-EXC2-CA.luna.kuleuven.be>
	<E93ACBBD7EA842B0AB8E3A577C3F41CA@IBM7FFA209F07C>
	<776F8577325D4C4881D6DFA12096E0D3@AGB>
	<001601cba2cc$b7138910$0401a8c0@xphd97xgq27nyf>
	<1293134076.24424.16.camel@pat-compaq-evo>
	<000701cba312$63924420$0301a8c0@your91hoehfy9g>
	<4D145A2E.4080706@telus.net>
	<4D146CA3.7090708@online.fr>
Date: Fri, 24 Dec 2010 10:12:07 +0000
Message-ID: <AANLkTim2xTemSG+e2KrQkPiJfW6RRs9_XCKAtW+K04Pd@mail.gmail.com>
From: Andy Talbot <andy.g4jnt@gmail.com>
To: rsgb_lf_group@blacksheep.org
DomainKey-Status: good (testing) 
Subject: Re: LF: Re: RE: Full ID
Content-Type: multipart/alternative; boundary=90e6ba4fc514061d8504982538b8
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


--90e6ba4fc514061d8504982538b8
Content-Type: text/plain; charset=ISO-8859-1

Joe Taylor has come up with the tightest source coding of amateur callsigns
as built into the WSPR and JT65  protocols.
By examining the make up of callsigns from letters numbers and spaces, with
even those restricted in where they lie in the callsign, he compresses all
possible ones - bar a few special event types - into 28 bits.   Full coding
details are in http://www.g4jnt.com/Coding/WSPR_Coding_Process.pdf

What we need now, for LF, is a way to turn the 28 bits of data into a
frequency shift + on/off keyed scheme for reading directly off
screen.  Something better can probably be done than Morse / DFCW,  although
to read without the use of computer decoding will require a new operator
skill / language to be learned.   (Somehting for the already perfect CW OPs
to get their teeth into  perhaps :-)

It ought to be poossible to code something up that is appreciably shorter
than the full callsign, yet conveys that unambiguously.  But, note that as
the callsign has now had all its redundancy removed, ANY errors in decoding
will generate an equally valid-looking but erroneous callsign.    That's the
penalty of good source coding - we see it occasionally on WSPR reports when
noise/QRM manages to pass the error detection tests.

Andy
www.g4jnt.com

--90e6ba4fc514061d8504982538b8
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div>Joe Taylor has come up with the tightest source=A0coding of amateur=
 callsigns as=A0built into the WSPR=A0and JT65 =A0protocols.=A0=A0=A0 </di=
v>
<div>By examining the make up of callsigns=A0from=A0letters numbers and sp=
aces,=A0with even those restricted in where they lie in the callsign, he=
 compresses all possible=A0ones - bar a few special event types -=A0into=
 28 bits.=A0=A0 Full coding details are in <a href=3D"http://www.g4jnt.com=
/Coding/WSPR_Coding_Process.pdf">http://www.g4jnt.com/Coding/WSPR_Coding_P=
rocess.pdf</a>=A0 </div>

<div>=A0</div>
<div>What we need now, for LF, is a way to turn the=A028 bits of data into=
 a frequency=A0shift + on/off keyed scheme for reading directly off screen=
.=A0=A0Something better can probably be done than=A0Morse / DFCW,=A0 altho=
ugh to read without=A0the use of computer decoding will require a new oper=
ator skill / language to be learned.=A0=A0=A0(Somehting for the already pe=
rfect CW OPs to get their teeth into=A0 perhaps :-)</div>

<div><br>It ought to be poossible to code something up that is appreciably=
 shorter than the full callsign, yet conveys=A0that unambiguously.=A0 But,=
 note that as the callsign has now had all its redundancy removed, ANY err=
ors in decoding will generate an equally valid-looking but erroneous calls=
ign.=A0=A0=A0 That&#39;s the penalty of good source coding - we see it occ=
asionally on WSPR reports when noise/QRM manages to pass the error detecti=
on tests.</div>

<div>=A0</div>
<div>Andy</div>
<div><a href=3D"http://www.g4jnt.com/">www.g4jnt.com</a></div>
<div>=A0</div>

--90e6ba4fc514061d8504982538b8--