Tenth edition of the N&O column / Spooks newsletter

(Date: Sat, 30 Jan 1999 13:37:02 +0100 CET)

Press release | Locations | Morse stations | Voice stations | Other modes
Polyalphabetic substitution ciphers | ENIGMA list update | Logs
Index | NS NL Home


Other modes

Polytone stations [XPH]

Valeriano Martin surprised us with a decoding theory for the Polytone stations, so I gladly hand over the mike to him:

"I've almost attained to decode XPH and I would like to confirm my results, but I've not seen any post in the matter. Have anyone been able to decode it?

I have identified 14 different tones within its alphabet, with a step of 16 Hz between two correlative tones. I have initially assigned an arbitrary letter to each tone, according to this list:

Letter Freq. (approx)
A 305Hz
B 321
C 337
D 353
E 369
F 385
G 401
H 417
I 433
J 449
K 465
L 481
M 497
N 513

For instance, the sample 39 in CONET CD#2 could be written in this way:
AGCHADAGCHAGCHAGCHADAGCHAGCHAGCHADAGCHAGCHAGCHADAGCHAGCHAGCH
ADAGCHAGCHAGCHADAGCHAGCHAGCHADA
mnmnmnmnmnmnmnmnmnmnaaaaaaaaaa
cnegiacncehaflhenalhcfnalckefadjfchacenifaikleda
cfihcadiknkacekchadcjneacjcneacfinjadgjijacdelca
fhcflaknhckaicjkgaienehainjnlagldhkacgighadkghja
kjikgagenfhaehnjn
bnbnbnbnbnbnbnbnbnbn

And the track 42 (null message),
AGDCAGDCAGDCACNCAGDCAGDCAGDCACNCAGDCAGDCAGDCACNCA
mnmnmnmnmnmnmnmnmnmnaaaaaaaaaa
bnbnbnbnbnbnbnbnbnbn

Where upper case letters represent 1 sec duration and lower case 133 msec. I have noted:

Therefore, we have the equivalences:

A Space
B ???
C 0
D 1
E 2
F 3
G 4
H 5
I 6
J 7
K 8
L 9
M ???
N Repetition

Making the apropriate substitutions, the callups turn into:
1st sample: 405 1 405 405 405 1 405 405 405 1 405 405 405 1 405 405 405 1 405 405 405 1 405 405 405 1
2nd sample: 410 410 410 000 410 410 410 000 410 410 410 000

It looks like S7, V7, M12... isn't it? Three digit group seems to be a frequency id and the single figure, the number of messages to be sent. I have checked it. For example, the call up on Friday, 8 Jan 199999 was:
GFNAGFNAGFNAEA or 433 433 433 2

And the sked:
11462kHz at 0700

12362 at 0720
13362 at 0740

It sent two messages, separated by something like cncncacncncaccccca (I'm not sure, the reception was not good). First and second groups in each message were ID and GC, as we can expect:
cdjihacndkh...(1110 tones)...cncncacncncacccccc
cjcheacncjl...(474 tones)...bn(x10)
that is
01765 00185...(185 x 5fg )...00000 00000 000000
07052 00079...( 79 x 5fg )...bn(x10)

Comments appreciated."

Wow, great job Valeriano! Chris Smolinski reacted as follows:

"Very interesting results! A few thoughts...

Using a tone to indicate "duplicate" makes sense. Forcing sequential tones to be different probably helps improve the tone detection and decoding. That way, it is known when one character ends and the next begins, without having to rely just on timing.

The recording of XPH I have has some 'pulses' between each of the 'a' characters in the 'aaaaaaaaaa' sequence just prior to the start of the message text.

I wonder if these ten 'a' (305 Hz) tones sent just before the message are a sort of timing series, to get the decoder synchronized and ready to copy the message text? Likewise, the repeating "mn" series could serve to indicate that the synchronization is about to begin.

Now what we really need - a program to decode XPH messages! (Well, convert to numbers is a better term, since we're really not decoding the message text). I'll gladly do a Mac version. Anyone want to team up to work on a combined Mac/Windows program?"

This certainly is a challenge guys! Anyone for a Windows version of the program? Keep us informed please?! -Ary-

See also 'Locations' in this Newsletter .
See also Newsletter 12.

Press release | Locations | Morse stations | Voice stations | Other modes
Polyalphabetic substitution ciphers | ENIGMA list update | Logs
Index | NS NL Home

---