Well done Stefan and Paul!

>> i forgot the message

Huh what a pity, so you won't be able to confirm Paul's decode... unless someone with a >10 GB machine manages to pull it out from the DL0AO wav ;-)

73, Markus


-----Ursprüngliche Mitteilung-----
Von: DK7FC <[email protected]>
An: rsgb_lf_group <[email protected]>
Verschickt: Mo, 12. Jun 2017 14:23
Betreff: Re: VLF: This weekend...

Hello Paul, VLF,

That's nice, a new record for that week, 100 characters on 6.47 kHz in
a single (non-stacked) transmission, even in June!! :-)
Many thanks for the decode. It looks like you are further improving your
software, so it become a even more powerful tool! I run some very basic
exercises in the last days and i can say i'm impressed and i will become
a friend of the tool(s).

Reporting the decoded message indirectly by using the md5 hash is a
cleaver idea, but i find it somehow spoils the fun a bit. In principal
the message content is not relevant because it could be any message. But
it looks like a bit far from amateur radio when the message is not told
directly.
Always think about potential newcomers: If someone reads
"9f7707ee7424c3c07c3524616fe45b9b" then he could think "What the hell
are they doing there??? but if he reads "CQ CQ on the 46 km band de
DK7FC in JN49IK, pse KK" then he may rather think "Wow they are really
transferring messages on such low frequencies, much longer messages than
in JT9, i want to try that too"...

Another thing is that i forgot the message, so i'm curious of what you
have decoded :-) Can you tell the message please?

94% channel capacity and just 31.1 % bit error rate, how can this be?

73, Stefan

Am 12.06.2017 11:25, schrieb Paul Nicholson:
>
> Stefan schrieb:
>
> > f = 6470.100000 Hz
> > Start time: 10.Jun.2017 04:00:00 UTC
> > Symbol period: 20 s
> > Characters: 100
> > CRC bits: 8
> > Coding 8K21A
>
> Copied a message with md5 hash (*): 9f7707ee7424c3c07c3524616fe45b9b
> Constant reference phase.
>
> BER 31.1%
>
> Eb/N0 = -1.1 dB, 94% of channel capacity.
>
> Carrier S/N: 26.42 dB in 10.0 uHz,
> -23.60 dB in 1Hz,
> -57.58 dB in 2.5kHz
>
> List rank 1159.
> Decoder RAM: 10Gbyte with list length 10k.
>
> I used ebnaut option -a 100 to normalise the amplitudes in the
> decoder. It didn't decode at all without that.
I will try that here too, thanks. It didn't see the -a 100 option in the
documentation.
>
> 100 chars is a record message length at 6470 Hz.
>
> Error map http://abelian.org/vlf/tmp/170612a.png
>
> I used a night blanker setting -a 12 (in version 8 onwards),
> or about -a1.2 before version 8.
>
> (*) echo -n 'XXX...XXX' | md5sum
>
> --
> Paul Nicholson
> --
>