From: Rik Strobbe <Rik.Strobbe@fys.kuleuven.be>
To: "rsgb_lf_group@blacksheep.org" <rsgb_lf_group@blacksheep.org>
Date: Wed, 1 Feb 2012 14:40:09 +0100
Thread-Topic: LF: RE: OPERA and QRS
Thread-Index: Aczg4req0N1cQt8XQaSVca3b1AnSVAAAX0Yo
Message-ID: <BF4A524700075746A6467658DFC7102CB0B68CCD71@ICTS-S-EXC2-CA.luna.kuleuven.be>
References: <4F2911FC.3020709@talktalk.net>
 <BF4A524700075746A6467658DFC7102CB0B68CCD6D@ICTS-S-EXC2-CA.luna.kuleuven.be>,<4F2938EB.8010504@talktalk.net>
In-Reply-To: <4F2938EB.8010504@talktalk.net>
Accept-Language: nl-NL, nl-BE
Content-Language: nl-BE
acceptlanguage: nl-NL, nl-BE
MIME-Version: 1.0
Subject: RE: LF: RE: OPERA and QRS
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
X-AOL-SCOLL-URL_COUNT: 0  
x-aol-sid: 3039ac1db4054f2940ea40e1
X-AOL-IP: 195.171.43.25
X-AOL-SPF: domain : blacksheep.org SPF : none


Hello Eddy,

Opera (and WSPR) are using the best way to improve SNR, in their case FEC.
For QRSS using a shorter dotlength and repeating the message (what is in fa=
ct and kind of FEC) is probably not the best way. Increasing dotlength is.
To give an extreme example: would repeating a callsign at 20WPM during 4 mi=
nutes ever be competitive to Opera4 or QRSS3. Certainly not, because this m=
ethod is no good in improving SNR.=20
So "fair" to me sounds: let each mode use it best capabilities to transmit =
a given message in a given timeframe.

About PA7RIS / YQ0YQY: the last is a worst case example for QRSS. Opera nee=
ds the same time to transmit a 6 character call, that is a choise the desig=
ner made. CW (and thus QRSS) took another option.

Comparing Opera4 with QRSS1 maybe an interesting test in a "men against mac=
hine" competition, but rather meaningless in real world. And without any do=
ubt it will be in favour of Opera, as this is in advantage of FEC (used by =
Opera, nut used by QRSS).
The other way around: could an Opera version without any FEC and 3 sec. bit=
length compete with QRSS3. I doubt it.

73, Rik  ON7YD / OR7T


________________________________________
Van: owner-rsgb_lf_group@blacksheep.org [owner-rsgb_lf_group@blacksheep.org=
] namens qrss [qrss@talktalk.net]
Verzonden: woensdag 1 februari 2012 14:06
Aan: rsgb_lf_group@blacksheep.org
Onderwerp: Re: LF: RE: OPERA and QRS

Hello Rik

I am glad you used the word fair, I have been trying to avoid it. my
first though when viewing my TX sending QRS3, then OPERA was, and is
every time I look,  this is not fair.

Yes digital modes are allowed FEC, as Andy says QRS is allowed the brain
which is far more powerful than the computer, surely it is unfair to ask
for longer bit time to compensate for repetition, let QRS be repeated  4
times at QRS1.024. No need in my test transmission though I
think.......unless someone would like it.

The PA7RIS idea sounds OK but again would not be fair as the full
capability of OPERA would not be used so favouring QRS. Only YQ0YQY will
do, oh no, now every one knows the code call.

73 Eddie


On 01/02/2012 10:52, Rik Strobbe wrote:
> Hello Eddy,
>
> comparing Opera and QRSS with identical bit lengths seems not correct to =
me as Opera uses "ticks" (such as Forward Error Correction) to improve SNR.
> Assuming a simple 8 bit character set it takes only 48 bits to transmit a=
 6 character call. With clever coding (as used in WSPR) it is even possible=
 to fit a 6 character call into 28 bits. But each Opera transmission contai=
ns 240 bits.
> So if Opera is allowed to use FEC in order to improve, QRSS should be all=
owed to increase dotlength for the same reason.
> Same transmission duration seems far more fair than same bit/dot length.
> But then the dot length would have to differ for every call, so maybe we =
should agree on an "average call" to determine the QRSS dot length. A tradi=
tional method to determine CW speed is the PARIS method (PARIS includes exa=
ctly 50 dots, including the word spacing), inserting an average length numb=
er (7) would give us the nice call PA7RIS, that contains 66 dots, including=
 the word spacing. As Opera transmits only a single word (call) the word sp=
acing is not used, so it can be ommted also in QRSS what leads to a (conven=
ient) number of 60 dots.
> Opera4 takes 246 seconds, so the QRSS speed to compare with should be 246=
/60 =3D 4(.096) seconds dot lenght.
>
> 73, Rik  ON7YD - OR7T
>
> ________________________________________
> Van: owner-rsgb_lf_group@blacksheep.org [owner-rsgb_lf_group@blacksheep.o=
rg] namens qrss [qrss@talktalk.net]
> Verzonden: woensdag 1 februari 2012 11:20
> Aan: rsgb_lf_group@blacksheep.org
> Onderwerp: LF: OPERA and QRS
>
> We all seem to have agreed to compare OPERA with a QRS speed which takes
> the same time to send a call sign, correct or not?
> A QRS beacon can send the call continuously all day and successive
> periods are often used to establish the information.
>
> How about comparing like for like in the timing of Data Bits, OPERA uses
> digital techniques and no doubt repeats the data in  the 4 minute period
> of OP4. In this case we would need to be comparing it with QRS1 or
> QRS1.025 if you like.
>
> I feed a coding session coming on.
>
> 73 Eddie G3ZJO
>=