I went about this as I would a numbers TX, although like most people I think it is most probably a spoof. The first thing I did was look for patterns and habits, not worrying yet about a possible code.
It looks and sounds like the second half of the message is identical to the first half. If so then it would appear that all groups are 5 figure. This works with all of the recordings I have heard also, I think I have never heard a "for sure" 4 figure group in any of the recordings people have sent me. Unfortunately I have very poor reception of this TX, so my own recordings are mostly noise.
It would appear that 4 transmissions have been heard, from various people. WEAK heard the first on 11/28. Several people, including myself, heard the 12/04 and 12/05. Textim (and someone else in #pirateradio or #wunclub, but I forget who…DOH!) heard one on 12/06 that I really could not hear, the Plasma TV was on at the time and I could tell something was in there, but not exactly what it was.
The 11/28 recording from WEAK matches exactly the transmission heard 12/04. I played the 11/28 recording in the left channel while I had the 12/04 in the right (this is something I often do with numbers stations). I synchronized them at the first point both recordings had in common, one of the circus music starts. They (11/28 and 12/04) were absolutely identical in run time and content, of course they suffered slightly different fades and such. Both of these were Apple in the Tree transmissions. All of the message Ids I can hear all sound like 1913, not 1918.
My transcription differs a little from yours. Putting one recording in each ear really makes the signals more understandable (after you confirm they are identical) and I made an assumption that when a couple of groups in one recording got covered by OTHR or some other fade that the other recording was identical at that point, since ALL parts that could be heard clearly were indeed identical. Also it helps that I have multiple recordings to choose from for the 12/04 TX. So this is a “fixed” version, based on assuming the 4 bodies (2 each 11/28 and 12/04) are identical and using each to fill in the blanks.
MSG Body:
9 8 3 4 9 / 1 3 2 1 4 / 0 3 1 6 0 / 7 3 1 4 9 / 0 7 5 1 0 /
9 3 4 8 2 / 1 9 3 8 5 / 4 7 1 5 7 / 0 9 1 3 4 / 2 5 1 0 0 /
5 4 3 9 8 / 2 3 8 4 6 / 5 9 7 0 1 / 2 5 1 5 6 / 5 7 5 6 2 /
3 4 1 3 8 / 5 1 3 8 0 / 2 5 4 7 3 / 8 1 0 1 9 / 8 5 4 6 7 /
1 0 1 3 4
The 12/05 TX was the first reported Apple on the Ground message. I have not fully transcribed it but it was substantially different from the Apple in the Tree message when I tried to synchronize the sounds. I also tried to synch just the numbers groups, and that also showed they are different message bodies with different numbers of groups (from the “Tree” messages) but both halves of the “Ground” message are identical. However, the last four groups in each message (first and second half) appear the same for both the "Tree" message and the "Ground" message, “2 5 4 7 3 / 8 1 0 1 9 / 8 5 4 6 7 / 1 0 1 3 4”
I have not done a full transcription of the “ground” message body yet, but it looks like the message ID is different, it sounds like 3103 to me. Only having one recording may mean I have a few more holes in this one in the end.
The 12/06 TX was an Apple on the Ground message. I have not heard a recording of it, so I do not know if it matches the other “ground” message on 12/05.
So, four transmissions total that I know of, although I have only heard two at my location for myself.
11/28/2010 (Sunday) at about 0230 “Tree”, Message 1913
12/04/2010 (Saturday) at about 0100 “Tree”, Message 1913
12/05/2010 (Sunday) at about 0230 “Ground”, Message 3103
12/06/2010 (Monday) at about 0600 “Ground”, Unk Message no.