G2 G2a E23
E23 "Cynthia Voice"  wma
8188 kHz 23 SEP 04
55555 66666 77777 88888 99999 count in.
E23 "Cynthia Voice"  wma
8188 kHz 21 OCT 04
Swedish Rhapsody from 1989 wma 5 minutes long
Swedish Rhapsody  pre-1998  
Swedish Rhapsody  from Leen, NL. via spook007
Swedish Rhapsody1 from Leen, NL. via spook007
A clip sent in by Andrew T. recorded on 15.1MHz, 1 AUG 1977 at 0905 UTC.
Swedish Rhapsody 1977
A clip from Jochen Schafer 17 JAN 82 0800 8188 kHz    
Swedish Rhapsody 1982
A clip from Jochen Schafer 17 FEB 82 0800 8188 kHz    
Swedish Rhapsody 1982 Different voice
A clip from Jochen Schafer  6507 kHz    
Swedish Rhapsody with comments from a young Jochen


Comments by Peter H in Germany
"I've never had a problem with identifying this
voice as that of a grown-up woman. I always thought
that she was of East German origin, because the "o" in
"zwo" sounds very much like that pronounced by people
from Saxony."

Description by ENIGMA staff

This station has recently undergone a massive change, of which more later. However let's begin with the original station with the child's voice and the rather tacky music box intro. Langley Pierce has identified this station as being operated by Austrian Intelligence. However, it has recently (late 1998) been traced to Poland. This is his description of the station: The preamble consists of a piece of music played by a music box repeated for five minutes beginning on the hour. The piece is called 'Swedish Rhapsody.' The piece is repeated 23 times.

Message. 73242 73242 95222 95222 04528 04528
Preamble: Swedish Rhapsody
This is the traffic list. Numbers refer to traffic identifiers. Note the absence of any group count..
Achtung! 73242 73242
Achtung! 40023 40023 67152 67152 76997 76997 Ende
Achtung! 95222 95222
Achtung! 83633 83633 84878 84878 21737 21737 Ende
Achtung! 04528 04528
Achtung! 57625 57625 92622 92622 71419 71419 Ende.

If a further transmission is scheduled for the next hour, instead of the transmitter being shut down, it instead transmits a 0.5 second 1100 Hz tone every 3 seconds up until the next transmission. This station also operates a secondary format. the preamble as which is made up of two distinct parts.

Firstly ten minutes before the hour or half hour depending on transmission time. a 0.5 second 1100 Hz tone is transmitted every 3 seconds. This continues up until the beginning of the musical preamble. And secondly, the music box piece, Swedish Rhapsody is played three times, followed by a count from 1 to 0 in two parts. repeated twice:
12345 67890
12345 67890
This sequence is repeated a total of six times. The message is proceeded by a further three playings of Swedish Rhapsody. The message follows as before:

2. Traffic. Traffic is presented in standard non-standard synthesised female voice at 17 groups per minute. Each digit Is presented smoothly, although final digits are often subjected to being prematurely cut off.

This effect is particularly noticeable with 'Seiben'. The station is characterised by the standard variants of funf, neun and noll being rendered aa "funnif", "noiyin" and "nal"
1.eins   6. sechs
2.Tsvo   7. seiben  
3.dri      8.acht
4. fier     9. noi-yin
5. funnef 0. Nall

3. Broadcast Schedule. Traffic is usually broadcast in the evening on pseudo-regular schedule. although transmissions have been heard in the early morning. Due to the length of transmission about 35 minutes sign-on times are restricted to begin on the hour. One interesting feature of this traffic is that every fifth individual group is followed by an extended pause.

Day Time (UTC) Frequency  Alternate Frequency
Mon 1700 5748 kHz
Mon 2200 5748 kHz 6200 kHz
Mon 2300 5748 kHz 6200 kHz
Tue 1900  5748 kHz 6507 kHz
Tue 2100  4779 kHz
Tue 2200  4832 kHz
Tue 2300  5748 kHz
Wed 0000 5340 kHz
Wed 1600 5748 kHz
Wed 1700 5748 kHz
Wed 2100 4751 kHz
Wed 2300 5340 kHz  5748 or 6200 kHz
Wed 2300 5748 kHz  6200kHz
Thu 2000  5340 kHz
Thu 2100  5340 kHz
Thu 2200  5340 kHz
Sat 1000 9457 kHz
Sat 1900 5340 kHz  6507 kHz
Sat 2000 5340 kHz  6507 kHz
Sat 2100 5340 kHz  6507kHz
Sat 2200 5340 kHz 6507 kHz
Sat 2300 5340 kHz 6507 kHz
Sun 1000  9457 kHz
Sun 1800  7584 kHz
Sun 1900  5535 kHz
Sun 2100  5340 kHz 6507 kHz
Sun 2200  5340 kHz 6507 kHz
Sun 2300  4779 kHz


Week Beginning


0752 0952 1152 1252
1st Sat u.s.b - 6.507 8.188 5.34
2nd Sat u.s.b - 7.250 8.188 5.748
3rd Sat u.s.b 4.832 6.200 8.188 6.507
4th Sat u.s.b 5.340 8.188 7.25 -

This type is one of only a few stations that does not broadcast a group count. The reason is quite simple Each item of traffic contains 100 groups. This is a fixed value and never changes The reasons why such traffic should always contain 100 groups opens up an entirely new avenue of reasoning. Most popular possibility for consistent group counts is that the traffic contains a set form in which only certain groups have to be altered to convey the necessary information.

During April 1998, the station underwent a radical transformation as described by the ENIGMA group:

This station, noted for its stability, has recently undergone two radical changes. On 1st April 1998 the old schedules were entirely replaced, turning the unruffled world of Swedish Rhapsody upside-down. Perhaps the biggest change was the total extinction of G2A, whose schedule hadn't altered for very many years. In the third week of April, all this changed again - this time, although the new schedules remained, the familiar 'little girl voice' along with her musical toy disappeared and has not been heard since. Yet another blow to eccentricity which will be sadly lost. The voice in German has been replaced by that of the brash American English - Cynthia. What conclusions can we draw from this?

Format has also changed - after a short settling down period. Gone are the single message voice transmissions; all are now of the 100/100/150 group triple type. Transmissions commence at 3 minutes to the hour with a call reminiscent of E5, however, no ID is included, and the counting is in the form of 5 Fig stutter groups. E.g. 11111 22222 33333 44444 55555 66666 77777 88888 99999 00000 repeated until the hour. On the hour the three message headers (still encrypted 5F are read out three times, followed by the first header twice which is immediately followed by the first message (100 pairs, as usual).

 Without pause the second header is read out twice, and the second message follows immediately. The same applies to the third (50 group) message, which ends abruptly, although the word "END" has also been noted. The main difference between this new format and the old one is the complete lack of pauses and announcements (no English equivalent of "Achtung!") - from beginning of traffic list to end of last message. Headers are read no differently as the message groups themselves, as if they were merely message groups. This gives the impression of one long 162 group message, which of course, it is not.

What does this imply? It now implies more so that all three message blocks are intended for the same recipients, and not separate addressees, as would be the case in a station like M10, S10, G3 or G8 for example. It also implies that blocks 2 and 3 are probably continuations (when needed) of the message encrypted in the first block. This means that the third block would always be the least likely one to carry valid traffic. The recipient would first of all transcribe the three headers as headings on a pre-printed (?) form. A grid of 100 spaces (10 x 10) below the first heading would then be filled in as the first block groups are read out. A prearranged

Message End Marker, say a particular run of 3 figures (in clear), would be looked for by the recipient and at that point all further groups read out would be ignored, the receiver being switched off. The headers probably provide decryption information, and possibly also a (simply?) encrypted group count - in which case a Message End Marker would not be needed. The message ending could be marked at the appropriate grid square - these squares may even be numbered 1-150. Perhaps the need for three blocks is due to a need (added security? An inherent weakness of the system?) of the encryption system used, a system designed for a maximum of 100 groups.

Why is the final block only half the length of the other two? Maybe it's an historical anomaly dating back to a period when time was needed to retune transmitters in time for the commencement of the 10min tone period prior to the next transmission. A 100 group final block would leave very little time to arrange this. 50 groups ending at h + 37.5, would make all the difference; it could be as simple as that. Maybe, for reasons unknown to the uninitiated, there is never any need for more than 150 groups, and this figure depends on a very specific kind of message -perhaps even so specific that precisely 100, 100 & 50 groups are always required, but this seems the least probable theory. We are open to further suggestions and wild speculations!