Just on an airport run but will get on it when I get back. I'm convinced their is at least one or more life forms in the UAs, especially after seeing an image just now, part of the UA looks suspiciously like an eggsack
UNKNOWN ARTEFACT: Sound Engineers, apply here!
- Thread starter sohara
- Start date
Just on an airport run but will get on it when I get back. I'm convinced their is at least one or more life forms in the UAs, especially after seeing an image just now, part of the UA looks suspiciously like an eggsack
Don't know if this is helpful, but there's a minor chord occuring sometimes, in sequence 4, 5, 6. It's a bit of an oddity, because most tones go up.
It's the straight horizontal lines between 300 and 600 hz, easily visible here:
Especially 4 and 5 look like they have almost the same effects and distortions on them.
Hope it's not ingame music or anything.
Oh and I feel there's also more slow "zombie" chatter in those 3 than usual.
I will fiddle around some more tomorrow ;-)
It's the straight horizontal lines between 300 and 600 hz, easily visible here:
Especially 4 and 5 look like they have almost the same effects and distortions on them.
Hope it's not ingame music or anything.
Oh and I feel there's also more slow "zombie" chatter in those 3 than usual.
I will fiddle around some more tomorrow ;-)
Last edited:
How about an Egg that is crying for its parents? It looks a hell of a lot like an egg. The chitter and purrs could be the creatures inside while the howl would be the "mom where are you" cry.
Has anyone confirmed whether the sound is dependent on the health %? Not sure how long it takes for the sound to change pitch but if it's at a salvageable percentage it may be a good idea to dump it, record the sound until the pitch changes, pick it up, then dump it again. If the sound depends on the health% it will remain in higher pitch. If the sound is the same we'll know if it's the health that changes it or if it is unrelated to the health.
Edit: I want to find one of these. Are there people out there actively looking for them or are all the efforts aimed at analyzing the existing audio material?
I know little about sound engineering but I have very good ears (and speakers/headphones) and am very good at rational thinking.
Last edited:
The (apparent) structure is more convincing than the variety. In fact, it could have more variety and have less evidence for something hidden in there.
The fact that no three consecutive purrs share tone is kind of a big deal. If there were no such structure. If the tones were convincingly random, I would have much less suspicion that something is hiding in the message, but there would be far more variety.
Took me a while to find it but the background sound "the buzzing" how I like to call it reminds me of the TOS Episode "wink of an Eye" where the aliens disappeared because they sped way up and were outside the visible spectrum. Someone said in the episode there was even a hidden message...
[video=youtube;JN4v-tXgxTs]https://www.youtube.com/watch?v=JN4v-tXgxTs[/video]
Hey guys, quick update on my weekend work. So I spent much of time doing a fairly lengthy transcription of one of the long audio recordings. It's preliminary, but I think I may have a significantly different structure for the high/low bit notes; once I free up later today I'll re-analyze it and report back.
Last edited:
With regards trying to extract some form of message from these sounds, there are a few fairly basic alternatives.
For option 1 we would expect to see large sections of message repeated, even if it is in some way encrypted.
For option 2 we would probably still expect to see large chunks of message repeated.
For option 3 we would expect to find much less in the way in repeated sections.
So the first step should be to see if there is a repeating sequence. Most importantly, we need to be able to see this sequence in more than two samples.
I've been trying to parse the trasnscripts as posted using some simple scripts that treat the data as a bitstream and look for common sequences between two transcriptions, and there really isn't a repeating sequence there.
I used 23 transcribed sequences of typically 70-80 'bits' long, the longest matches I found including uknown/missing data were these (first two numbers reference my dataset so that I can hand check things)
The longest matches without unkown/missing 'bits' are
Notice that in both these cases I'm not seeing the same sequence match in more than two transcriptions. I could do some further work to quantify how many of the shorter matches appear in multiple transcriptions, but, to be honest, given the results from checking for matches between two transcriptions I don't think it's worth the effort...
So, based on the data seen so far I would say there isn't any message hidden in the way that people are currently looking at things.
Match length summary
We do know a couple of things
I think that a lot of what is being dicussed is a case of of coming up with a theory and then hunting through the data in an attempt to find some segment that fits that theory. Whilst this is a very normal thing for humans to do when solving problems and it works very well for the types of problem where we understand the parameters (it's my appraoch to basic problem solving in my day job) it's prone to leading you down blind alleys when the problem lies outside our normal experience, and/or there are a number of unknown parameters. It can lead to some fun and entertaining theories so I'm still enjoying catching up with this =)
“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”
― Arthur Conan Doyle, Sherlock Holmes
- There is a static message - possibly encrypted.
- There is a changing message - a countdown, or a variation based on location.
- There is no message - and it is effectively random noise.
For option 1 we would expect to see large sections of message repeated, even if it is in some way encrypted.
For option 2 we would probably still expect to see large chunks of message repeated.
For option 3 we would expect to find much less in the way in repeated sections.
So the first step should be to see if there is a repeating sequence. Most importantly, we need to be able to see this sequence in more than two samples.
I've been trying to parse the trasnscripts as posted using some simple scripts that treat the data as a bitstream and look for common sequences between two transcriptions, and there really isn't a repeating sequence there.
I used 23 transcribed sequences of typically 70-80 'bits' long, the longest matches I found including uknown/missing data were these (first two numbers reference my dataset so that I can hand check things)
Code:
25 'bits
20 38 0110?11011010101101101101
17 37 0110011011010101101101101
28 'bits'
20 11 110110010010010010100101?010
1 4 ????100100?00100101001011010The longest matches without unkown/missing 'bits' are
Code:
20 'bits'
11 57 11001001001001010010
20 14 11001001001001010010
21 'bits'
6 56 110011001001011101001
11 14 110011001001011101001
24 'bits'
4 60 010101011000110010110011
2 47 010101011000110010110011
24 'bits'
16 44 100110100110010110101001
8 22 100110100110010110101001Notice that in both these cases I'm not seeing the same sequence match in more than two transcriptions. I could do some further work to quantify how many of the shorter matches appear in multiple transcriptions, but, to be honest, given the results from checking for matches between two transcriptions I don't think it's worth the effort...
So, based on the data seen so far I would say there isn't any message hidden in the way that people are currently looking at things.
Match length summary
Code:
Length Number of matches
1 143798
2 83692
3 40476
4 25210
5 16213
6 9070
7 5601
8 3311
9 2008
10 1214
11 707
12 432
13 257
14 149
15 99
16 72
17 27
18 20
19 15
20 8
21 5
22 1
23 2
24 2
25 1
26 0
27 0
28 1We do know a couple of things
- We appear to be seeing some form of generated sequence
- It follows rules in timing and sequence of sounds (e.g. not repeating the same sound three times in a row)
I think that a lot of what is being dicussed is a case of of coming up with a theory and then hunting through the data in an attempt to find some segment that fits that theory. Whilst this is a very normal thing for humans to do when solving problems and it works very well for the types of problem where we understand the parameters (it's my appraoch to basic problem solving in my day job) it's prone to leading you down blind alleys when the problem lies outside our normal experience, and/or there are a number of unknown parameters. It can lead to some fun and entertaining theories so I'm still enjoying catching up with this =)
“It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.”
― Arthur Conan Doyle, Sherlock Holmes
Last edited:
As most of the tones come in groups of 7, I did wonder whether it might be a Hamming(7,4) code, giving 4 bit BCD numbers.
So I took jspoto's list from here https://forums.frontier.co.uk/showthread.php?t=142912&page=18&p=2205787&viewfull=1#post2205787
and ran it through a decoder, padding the 6 bit ones with an extra bit, but it doesn't appear to yield any useful pattern.
I've also tried various options like reversing the bit order, or inverting them, but that doesn't seem to make much difference.
But here are the results, in case they give anyone else some ideas.
Missing bits padded with 0
Missing bits padded with 1
So I took jspoto's list from here https://forums.frontier.co.uk/showthread.php?t=142912&page=18&p=2205787&viewfull=1#post2205787
and ran it through a decoder, padding the 6 bit ones with an extra bit, but it doesn't appear to yield any useful pattern.
I've also tried various options like reversing the bit order, or inverting them, but that doesn't seem to make much difference.
But here are the results, in case they give anyone else some ideas.
Missing bits padded with 0
Code:
Input Syn E Output Data BE LE
1001000 101 5 1001100 0100 4 2
0010010 101 5 0010110 1110 14 7
1001011 001 4 1000011 0011 3 12
0100101 000 0 0100101 0101 5 10
0110011 000 0 0110011 1011 11 13
1101010 100 1 0101010 0010 2 4
0011010 100 1 1011010 1010 10 5
1001010 110 3 1011010 1010 10 5
0110101 110 3 0100101 0101 5 10
0110110 010 2 0010110 1110 14 7
1001000 101 5 1001100 0100 4 2
0110101 110 3 0100101 0101 5 10
0100100 111 7 0100101 0101 5 10
1001011 001 4 1000011 0011 3 12
1100110 000 0 1100110 0110 6 6
1010010 001 4 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
0011001 000 0 0011001 1001 9 9
0110011 000 0 0110011 1011 11 13
0110110 010 2 0010110 1110 14 7
0100110 100 1 1100110 0110 6 6
0101011 111 7 0101010 0010 2 4
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
0100101 000 0 0100101 0101 5 10
1001010 110 3 1011010 1010 10 5
1001001 010 2 1101001 0001 1 8
0101011 111 7 0101010 0010 2 4
0010010 101 5 0010110 1110 14 7
1100100 011 6 1100110 0110 6 6
1010110 100 1 0010110 1110 14 7
1001000 101 5 1001100 0100 4 2
0110110 010 2 0010110 1110 14 7
1100100 011 6 1100110 0110 6 6
0101100 110 3 0111100 1100 12 3
0100100 111 7 0100101 0101 5 10
1001100 000 0 1001100 0100 4 2
0110101 110 3 0100101 0101 5 10
0101100 110 3 0111100 1100 12 3
0110010 111 7 0110011 1011 11 13
1100110 000 0 1100110 0110 6 6
1010010 001 4 1011010 1010 10 5
0101100 110 3 0111100 1100 12 3
0011011 011 6 0011001 1001 9 9
1011001 100 1 0011001 1001 9 9
0010010 101 5 0010110 1110 14 7
0010100 011 6 0010110 1110 14 7
0100110 100 1 1100110 0110 6 6
0110010 111 7 0110011 1011 11 13
1001011 001 4 1000011 0011 3 12
1010110 100 1 0010110 1110 14 7
0011001 000 0 0011001 1001 9 9
0011000 111 7 0011001 1001 9 9
0011010 100 1 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
0010011 010 2 0110011 1011 11 13
0110110 010 2 0010110 1110 14 7
1001101 111 7 1001100 0100 4 2
1001100 000 0 1001100 0100 4 2
1001011 001 4 1000011 0011 3 12
1010010 001 4 1011010 1010 10 5
0110100 001 4 0111100 1100 12 3
0101010 000 0 0101010 0010 2 4
0101011 111 7 0101010 0010 2 4
1001001 010 2 1101001 0001 1 8
0110010 111 7 0110011 1011 11 13
1010110 100 1 0010110 1110 14 7
1010100 111 7 1010101 1101 13 11
0110011 000 0 0110011 1011 11 13
0101001 100 1 1101001 0001 1 8
1101001 000 0 1101001 0001 1 8
0101101 001 4 0100101 0101 5 10
0011011 011 6 0011001 1001 9 9
1010110 100 1 0010110 1110 14 7
1001101 111 7 1001100 0100 4 2
0011001 000 0 0011001 1001 9 9
0110101 110 3 0100101 0101 5 10
0011001 000 0 0011001 1001 9 9
0110100 001 4 0111100 1100 12 3
1010011 110 3 1000011 0011 3 12
0100110 100 1 1100110 0110 6 6
0110110 010 2 0010110 1110 14 7
0110010 111 7 0110011 1011 11 13
1001101 111 7 1001100 0100 4 2
1001010 110 3 1011010 1010 10 5
0100100 111 7 0100101 0101 5 10
0010110 000 0 0010110 1110 14 7
1101001 000 0 1101001 0001 1 8
0010110 000 0 0010110 1110 14 7
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
1010100 111 7 1010101 1101 13 11
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
1001100 000 0 1001100 0100 4 2
1010010 001 4 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
1101100 010 2 1001100 0100 4 2
0110011 000 0 0110011 1011 11 13
1100110 000 0 1100110 0110 6 6
1001101 111 7 1001100 0100 4 2
0101101 001 4 0100101 0101 5 10
0101011 111 7 0101010 0010 2 4
1010011 110 3 1000011 0011 3 12
0110011 000 0 0110011 1011 11 13
1001100 000 0 1001100 0100 4 2
1001011 001 4 1000011 0011 3 12
1010011 110 3 1000011 0011 3 12
0100100 111 7 0100101 0101 5 10
1011011 111 7 1011010 1010 10 5
0110011 000 0 0110011 1011 11 13
1100100 011 6 1100110 0110 6 6
1001001 010 2 1101001 0001 1 8
0100101 000 0 0100101 0101 5 10
1100110 000 0 1100110 0110 6 6
1011000 011 6 1011010 1010 10 5
1001011 001 4 1000011 0011 3 12
0110010 111 7 0110011 1011 11 13
0010010 101 5 0010110 1110 14 7
0110011 000 0 0110011 1011 11 13
0101100 110 3 0111100 1100 12 3
0110110 010 2 0010110 1110 14 7
1001010 110 3 1011010 1010 10 5
1100100 011 6 1100110 0110 6 6
0100110 100 1 1100110 0110 6 6
0101001 100 1 1101001 0001 1 8
0101000 011 6 0101010 0010 2 4
1010100 111 7 1010101 1101 13 11
1011011 111 7 1011010 1010 10 5
1101100 010 2 1001100 0100 4 2
1001101 111 7 1001100 0100 4 2
0010100 011 6 0010110 1110 14 7
1010010 001 4 1011010 1010 10 5
0110110 010 2 0010110 1110 14 7
1010100 111 7 1010101 1101 13 11
1011001 100 1 0011001 1001 9 9
0110110 010 2 0010110 1110 14 7
0110010 111 7 0110011 1011 11 13
0110011 000 0 0110011 1011 11 13
1100110 000 0 1100110 0110 6 6
1100101 100 1 0100101 0101 5 10
0010010 101 5 0010110 1110 14 7
1101100 010 2 1001100 0100 4 2
0110100 001 4 0111100 1100 12 3
1100110 000 0 1100110 0110 6 6
0101010 000 0 0101010 0010 2 4
0010100 011 6 0010110 1110 14 7
0110110 010 2 0010110 1110 14 7
0110010 111 7 0110011 1011 11 13
1001100 000 0 1001100 0100 4 2
1001100 000 0 1001100 0100 4 2
0101100 110 3 0111100 1100 12 3
1010010 001 4 1011010 1010 10 5
1001000 101 5 1001100 0100 4 2
1001010 110 3 1011010 1010 10 5
0011001 000 0 0011001 1001 9 9
0110010 111 7 0110011 1011 11 13
1001011 001 4 1000011 0011 3 12
0110100 001 4 0111100 1100 12 3
0110010 111 7 0110011 1011 11 13
0010100 011 6 0010110 1110 14 7
1001000 101 5 1001100 0100 4 2
1011010 000 0 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
1100110 000 0 1100110 0110 6 6
1001100 000 0 1001100 0100 4 2
1011010 000 0 1011010 1010 10 5
1001000 101 5 1001100 0100 4 2
1011010 000 0 1011010 1010 10 5
0110010 111 7 0110011 1011 11 13
1101101 101 5 1101001 0001 1 8
0100100 111 7 0100101 0101 5 10
1001011 001 4 1000011 0011 3 12
0011001 000 0 0011001 1001 9 9
1011010 000 0 1011010 1010 10 5
1011011 111 7 1011010 1010 10 5
0110101 110 3 0100101 0101 5 10
0011001 000 0 0011001 1001 9 9
0100100 111 7 0100101 0101 5 10
1001101 111 7 1001100 0100 4 2
1011001 100 1 0011001 1001 9 9
0110010 111 7 0110011 1011 11 13
0010010 101 5 0010110 1110 14 7
1001101 111 7 1001100 0100 4 2
0100100 111 7 0100101 0101 5 10
1011001 100 1 0011001 1001 9 9
1001101 111 7 1001100 0100 4 2
0110101 110 3 0100101 0101 5 10
0110110 010 2 0010110 1110 14 7
1100101 100 1 0100101 0101 5 10
0100110 100 1 1100110 0110 6 6
0100110 100 1 1100110 0110 6 6
1100101 100 1 0100101 0101 5 10
1001001 010 2 1101001 0001 1 8
0011010 100 1 1011010 1010 10 5
1001101 111 7 1001100 0100 4 2
0011011 011 6 0011001 1001 9 9
0110110 010 2 0010110 1110 14 7
0100110 100 1 1100110 0110 6 6
1001100 000 0 1001100 0100 4 2
1010110 100 1 0010110 1110 14 7
0101101 001 4 0100101 0101 5 10
1001001 010 2 1101001 0001 1 8
0010010 101 5 0010110 1110 14 7
1001010 110 3 1011010 1010 10 5
0101100 110 3 0111100 1100 12 3
1101101 101 5 1101001 0001 1 8
0101101 001 4 0100101 0101 5 10
1011011 111 7 1011010 1010 10 5
1011010 000 0 1011010 1010 10 5
1010101 000 0 1010101 1101 13 11
1001101 111 7 1001100 0100 4 2
0100110 100 1 1100110 0110 6 6
1001010 110 3 1011010 1010 10 5
0011000 111 7 0011001 1001 9 9
0010110 000 0 0010110 1110 14 7
1100100 011 6 1100110 0110 6 6
1101011 011 6 1101001 0001 1 8
1101001 000 0 1101001 0001 1 8
1001100 000 0 1001100 0100 4 2
1010010 001 4 1011010 1010 10 5
0110110 010 2 0010110 1110 14 7
1011001 100 1 0011001 1001 9 9
0010011 010 2 0110011 1011 11 13
0011001 000 0 0011001 1001 9 9
0110100 001 4 0111100 1100 12 3
1101010 100 1 0101010 0010 2 4
1010010 001 4 1011010 1010 10 5
0100110 100 1 1100110 0110 6 6
1010110 100 1 0010110 1110 14 7
1001001 010 2 1101001 0001 1 8
0100101 000 0 0100101 0101 5 10
0011011 011 6 0011001 1001 9 9
0110011 000 0 0110011 1011 11 13
1010100 111 7 1010101 1101 13 11
1010110 100 1 0010110 1110 14 7
0101011 111 7 0101010 0010 2 4
0011010 100 1 1011010 1010 10 5
0100110 100 1 1100110 0110 6 6
1010100 111 7 1010101 1101 13 11
0110010 111 7 0110011 1011 11 13
0110100 001 4 0111100 1100 12 3
0100100 111 7 0100101 0101 5 10
1001101 111 7 1001100 0100 4 2
1001001 010 2 1101001 0001 1 8
1001100 000 0 1001100 0100 4 2
0100100 111 7 0100101 0101 5 10
0011010 100 1 1011010 1010 10 5
0110100 001 4 0111100 1100 12 3
1011011 111 7 1011010 1010 10 5
0010011 010 2 0110011 1011 11 13
1011001 100 1 0011001 1001 9 9
0010110 000 0 0010110 1110 14 7Missing bits padded with 1
Code:
Input Syn E Output Data BE LE
1001001 010 2 1101001 0001 1 8
0010010 101 5 0010110 1110 14 7
1001011 001 4 1000011 0011 3 12
0100101 000 0 0100101 0101 5 10
0110011 000 0 0110011 1011 11 13
1101010 100 1 0101010 0010 2 4
0011010 100 1 1011010 1010 10 5
1001011 001 4 1000011 0011 3 12
0110101 110 3 0100101 0101 5 10
0110110 010 2 0010110 1110 14 7
1001001 010 2 1101001 0001 1 8
0110101 110 3 0100101 0101 5 10
0100100 111 7 0100101 0101 5 10
1001011 001 4 1000011 0011 3 12
1100110 000 0 1100110 0110 6 6
1010010 001 4 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
0011001 000 0 0011001 1001 9 9
0110011 000 0 0110011 1011 11 13
0110110 010 2 0010110 1110 14 7
0100111 011 6 0100101 0101 5 10
0101011 111 7 0101010 0010 2 4
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
0100101 000 0 0100101 0101 5 10
1001010 110 3 1011010 1010 10 5
1001001 010 2 1101001 0001 1 8
0101011 111 7 0101010 0010 2 4
0010011 010 2 0110011 1011 11 13
1100100 011 6 1100110 0110 6 6
1010110 100 1 0010110 1110 14 7
1001001 010 2 1101001 0001 1 8
0110111 101 5 0110011 1011 11 13
1100101 100 1 0100101 0101 5 10
0101101 001 4 0100101 0101 5 10
0100101 000 0 0100101 0101 5 10
1001101 111 7 1001100 0100 4 2
0110101 110 3 0100101 0101 5 10
0101100 110 3 0111100 1100 12 3
0110010 111 7 0110011 1011 11 13
1100110 000 0 1100110 0110 6 6
1010011 110 3 1000011 0011 3 12
0101100 110 3 0111100 1100 12 3
0011011 011 6 0011001 1001 9 9
1011001 100 1 0011001 1001 9 9
0010010 101 5 0010110 1110 14 7
0010100 011 6 0010110 1110 14 7
0100110 100 1 1100110 0110 6 6
0110011 000 0 0110011 1011 11 13
1001011 001 4 1000011 0011 3 12
1010110 100 1 0010110 1110 14 7
0011001 000 0 0011001 1001 9 9
0011001 000 0 0011001 1001 9 9
0011010 100 1 1011010 1010 10 5
1010110 100 1 0010110 1110 14 7
0010011 010 2 0110011 1011 11 13
0110110 010 2 0010110 1110 14 7
1001101 111 7 1001100 0100 4 2
1001100 000 0 1001100 0100 4 2
1001011 001 4 1000011 0011 3 12
1010011 110 3 1000011 0011 3 12
0110101 110 3 0100101 0101 5 10
0101011 111 7 0101010 0010 2 4
0101011 111 7 0101010 0010 2 4
1001001 010 2 1101001 0001 1 8
0110010 111 7 0110011 1011 11 13
1010110 100 1 0010110 1110 14 7
1010101 000 0 1010101 1101 13 11
0110011 000 0 0110011 1011 11 13
0101001 100 1 1101001 0001 1 8
1101001 000 0 1101001 0001 1 8
0101101 001 4 0100101 0101 5 10
0011011 011 6 0011001 1001 9 9
1010110 100 1 0010110 1110 14 7
1001101 111 7 1001100 0100 4 2
0011001 000 0 0011001 1001 9 9
0110101 110 3 0100101 0101 5 10
0011001 000 0 0011001 1001 9 9
0110100 001 4 0111100 1100 12 3
1010011 110 3 1000011 0011 3 12
0100110 100 1 1100110 0110 6 6
0110111 101 5 0110011 1011 11 13
0110010 111 7 0110011 1011 11 13
1001101 111 7 1001100 0100 4 2
1001011 001 4 1000011 0011 3 12
0100101 000 0 0100101 0101 5 10
0010110 000 0 0010110 1110 14 7
1101001 000 0 1101001 0001 1 8
0010110 000 0 0010110 1110 14 7
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
1010100 111 7 1010101 1101 13 11
1011001 100 1 0011001 1001 9 9
0100110 100 1 1100110 0110 6 6
1001101 111 7 1001100 0100 4 2
1010011 110 3 1000011 0011 3 12
1010111 011 6 1010101 1101 13 11
1101100 010 2 1001100 0100 4 2
0110011 000 0 0110011 1011 11 13
1100110 000 0 1100110 0110 6 6
1001101 111 7 1001100 0100 4 2
0101101 001 4 0100101 0101 5 10
0101011 111 7 0101010 0010 2 4
1010011 110 3 1000011 0011 3 12
0110011 000 0 0110011 1011 11 13
1001100 000 0 1001100 0100 4 2
1001011 001 4 1000011 0011 3 12
1010011 110 3 1000011 0011 3 12
0100100 111 7 0100101 0101 5 10
1011011 111 7 1011010 1010 10 5
0110011 000 0 0110011 1011 11 13
1100100 011 6 1100110 0110 6 6
1001001 010 2 1101001 0001 1 8
0100101 000 0 0100101 0101 5 10
1100111 111 7 1100110 0110 6 6
1011001 100 1 0011001 1001 9 9
1001011 001 4 1000011 0011 3 12
0110011 000 0 0110011 1011 11 13
0010010 101 5 0010110 1110 14 7
0110011 000 0 0110011 1011 11 13
0101100 110 3 0111100 1100 12 3
0110110 010 2 0010110 1110 14 7
1001010 110 3 1011010 1010 10 5
1100100 011 6 1100110 0110 6 6
0100110 100 1 1100110 0110 6 6
0101001 100 1 1101001 0001 1 8
0101001 100 1 1101001 0001 1 8
1010101 000 0 1010101 1101 13 11
1011011 111 7 1011010 1010 10 5
1101100 010 2 1001100 0100 4 2
1001101 111 7 1001100 0100 4 2
0010100 011 6 0010110 1110 14 7
1010011 110 3 1000011 0011 3 12
0110110 010 2 0010110 1110 14 7
1010100 111 7 1010101 1101 13 11
1011001 100 1 0011001 1001 9 9
0110111 101 5 0110011 1011 11 13
0110011 000 0 0110011 1011 11 13
0110011 000 0 0110011 1011 11 13
1100111 111 7 1100110 0110 6 6
1100101 100 1 0100101 0101 5 10
0010011 010 2 0110011 1011 11 13
1101100 010 2 1001100 0100 4 2
0110100 001 4 0111100 1100 12 3
1100110 000 0 1100110 0110 6 6
0101010 000 0 0101010 0010 2 4
0010101 100 1 1010101 1101 13 11
0110111 101 5 0110011 1011 11 13
0110011 000 0 0110011 1011 11 13
1001100 000 0 1001100 0100 4 2
1001100 000 0 1001100 0100 4 2
0101101 001 4 0100101 0101 5 10
1010010 001 4 1011010 1010 10 5
1001001 010 2 1101001 0001 1 8
1001010 110 3 1011010 1010 10 5
0011001 000 0 0011001 1001 9 9
0110010 111 7 0110011 1011 11 13
1001011 001 4 1000011 0011 3 12
0110101 110 3 0100101 0101 5 10
0110011 000 0 0110011 1011 11 13
0010100 011 6 0010110 1110 14 7
1001001 010 2 1101001 0001 1 8
1011010 000 0 1011010 1010 10 5
1010111 011 6 1010101 1101 13 11
1100110 000 0 1100110 0110 6 6
1001100 000 0 1001100 0100 4 2
1011010 000 0 1011010 1010 10 5
1001001 010 2 1101001 0001 1 8
1011011 111 7 1011010 1010 10 5
0110011 000 0 0110011 1011 11 13
1101101 101 5 1101001 0001 1 8
0100101 000 0 0100101 0101 5 10
1001011 001 4 1000011 0011 3 12
0011001 000 0 0011001 1001 9 9
1011010 000 0 1011010 1010 10 5
1011011 111 7 1011010 1010 10 5
0110101 110 3 0100101 0101 5 10
0011001 000 0 0011001 1001 9 9
0100100 111 7 0100101 0101 5 10
1001101 111 7 1001100 0100 4 2
1011001 100 1 0011001 1001 9 9
0110011 000 0 0110011 1011 11 13
0010011 010 2 0110011 1011 11 13
1001101 111 7 1001100 0100 4 2
0100101 000 0 0100101 0101 5 10
1011001 100 1 0011001 1001 9 9
1001101 111 7 1001100 0100 4 2
0110101 110 3 0100101 0101 5 10
0110110 010 2 0010110 1110 14 7
1100101 100 1 0100101 0101 5 10
0100110 100 1 1100110 0110 6 6
0100110 100 1 1100110 0110 6 6
1100101 100 1 0100101 0101 5 10
1001001 010 2 1101001 0001 1 8
0011011 011 6 0011001 1001 9 9
1001101 111 7 1001100 0100 4 2
0011011 011 6 0011001 1001 9 9
0110110 010 2 0010110 1110 14 7
0100110 100 1 1100110 0110 6 6
1001100 000 0 1001100 0100 4 2
1010110 100 1 0010110 1110 14 7
0101101 001 4 0100101 0101 5 10
1001001 010 2 1101001 0001 1 8
0010010 101 5 0010110 1110 14 7
1001011 001 4 1000011 0011 3 12
0101101 001 4 0100101 0101 5 10
1101101 101 5 1101001 0001 1 8
0101101 001 4 0100101 0101 5 10
1011011 111 7 1011010 1010 10 5
1011010 000 0 1011010 1010 10 5
1010101 000 0 1010101 1101 13 11
1001101 111 7 1001100 0100 4 2
0100110 100 1 1100110 0110 6 6
1001010 110 3 1011010 1010 10 5
0011001 000 0 0011001 1001 9 9
0010111 111 7 0010110 1110 14 7
1100100 011 6 1100110 0110 6 6
1101011 011 6 1101001 0001 1 8
1101001 000 0 1101001 0001 1 8
1001100 000 0 1001100 0100 4 2
1010010 001 4 1011010 1010 10 5
0110110 010 2 0010110 1110 14 7
1011001 100 1 0011001 1001 9 9
0010011 010 2 0110011 1011 11 13
0011001 000 0 0011001 1001 9 9
0110101 110 3 0100101 0101 5 10
1101011 011 6 1101001 0001 1 8
1010010 001 4 1011010 1010 10 5
0100110 100 1 1100110 0110 6 6
1010110 100 1 0010110 1110 14 7
1001001 010 2 1101001 0001 1 8
0100101 000 0 0100101 0101 5 10
0011011 011 6 0011001 1001 9 9
0110011 000 0 0110011 1011 11 13
1010101 000 0 1010101 1101 13 11
1010111 011 6 1010101 1101 13 11
0101011 111 7 0101010 0010 2 4
0011010 100 1 1011010 1010 10 5
0100110 100 1 1100110 0110 6 6
1010100 111 7 1010101 1101 13 11
0110010 111 7 0110011 1011 11 13
0110100 001 4 0111100 1100 12 3
0100100 111 7 0100101 0101 5 10
1001101 111 7 1001100 0100 4 2
1001001 010 2 1101001 0001 1 8
1001100 000 0 1001100 0100 4 2
0100101 000 0 0100101 0101 5 10
0011010 100 1 1011010 1010 10 5
0110100 001 4 0111100 1100 12 3
1011011 111 7 1011010 1010 10 5
0010011 010 2 0110011 1011 11 13
1011001 100 1 0011001 1001 9 9
0010110 000 0 0010110 1110 14 7
To me, this is an indication that a range of numbers is encoded, but not the entire available range is used.
For example, if I encode the letters A - Z in 32 bit numbers, the highest bit will be a 0 sixteen times, and a 1 only 10 times.
I've been doing a bit more digging in the raw data. This time looking at adjacent n-grams. This approach has some limitations when applied to something as simple as a (potentially) binary string, but it's kept me busy during a couple of lunch hours =)
This is looking at the data within each 6-7 bit clause, tests that include an unknown digit in either the initial or following n-gram are dropped and not included in the results.
Starting with the simplest analysis. How does an initial n-gram of length 1 influence a following n-gram of length 1? We see that roughly 2/3 of the time the 'opposing' digit is chosen, this is the start of a pattern that continues as we look at longer n-grams.
Code:
0 0 32.26%
0 1 67.74%
1 0 67.33%
1 1 32.67%Moving forward to look at an initial n-gram length of 2 and a following n-gram length of 1 we see a very clear pattern, one that was noticed already very early on; there are no cases where we see the same digit three times in a row. Notice that where the digits in the initial n-gram are not the same the distribution of the following n-gram is pretty much even.
Code:
00 0 0.00%
00 1 100.00%
01 0 49.75%
01 1 50.25%
10 0 49.11%
10 1 50.89%
11 0 100.00%
11 1 0.00%We see the same pattern when looking at an initial n-gram length of 3 and a following n-gram length of 1.
Code:
001 0 50.68%
001 1 49.32%
010 0 48.47%
010 1 51.53%
011 0 100.00%
011 1 0.00%
100 0 0.00%
100 1 100.00%
101 0 49.69%
101 1 50.31%
110 0 49.64%
110 1 50.36%When we look at the longer n-gram pairs we see the same thing; wherever the distribution is not even it's because it's being influenced by the 'no digit repeated three times' rule. For some of the longer n-grams the distribution is less even, most likely because of the smaller sample size for each n-gram pair.
Initial n-gram length 1, following n-gram length 2
Code:
0 00 0.00%
0 01 31.87%
0 10 33.90%
0 11 34.23%
1 00 33.74%
1 01 34.96%
1 10 31.30%
1 11 0.00%Initial n-gram length 2, following n-gram length 2
Code:
00 00 0.00%
00 01 0.00%
00 10 50.68%
00 11 49.32%
01 00 25.00%
01 01 26.58%
01 10 48.42%
01 11 0.00%
10 00 0.00%
10 01 48.42%
10 10 25.63%
10 11 25.95%
11 00 49.64%
11 01 50.36%
11 10 0.00%
11 11 0.00%Initial n-gram length 2, following n-gram length 3
Code:
10 000 0.00%
10 001 0.31%
10 010 26.02%
10 011 25.08%
10 100 10.97%
10 101 15.05%
10 110 22.57%
10 111 0.00%
11 000 0.00%
11 001 47.55%
11 010 29.37%
11 011 23.08%
11 100 0.00%
11 101 0.00%
11 110 0.00%
11 111 0.00%
00 000 0.00%
00 001 0.00%
00 010 0.00%
00 011 0.00%
00 100 27.43%
00 101 26.86%
00 110 45.71%
00 111 0.00%
01 000 0.00%
01 001 24.56%
01 010 12.13%
01 011 15.68%
01 100 22.19%
01 101 25.44%
01 110 0.00%
01 111 0.00%Initial n-gram length 1, following n-gram length 3
Code:
0 000 0.00%
0 001 0.00%
0 010 16.16%
0 011 15.73%
0 100 17.03%
0 101 18.10%
0 110 32.97%
0 111 0.00%
1 000 0.00%
1 001 33.63%
1 010 17.80%
1 011 18.02%
1 100 15.16%
1 101 15.38%
1 110 0.00%
1 111 0.00%Initial n-gram length 2, following n-gram length 3
Code:
10 000 0.00%
10 001 0.31%
10 010 26.02%
10 011 25.08%
10 100 10.97%
10 101 15.05%
10 110 22.57%
10 111 0.00%
11 000 0.00%
11 001 47.55%
11 010 29.37%
11 011 23.08%
11 100 0.00%
11 101 0.00%
11 110 0.00%
11 111 0.00%
00 000 0.00%
00 001 0.00%
00 010 0.00%
00 011 0.00%
00 100 27.43%
00 101 26.86%
00 110 45.71%
00 111 0.00%
01 000 0.00%
01 001 24.56%
01 010 12.13%
01 011 15.68%
01 100 22.19%
01 101 25.44%
01 110 0.00%
01 111 0.00%Initial n-gram length 3, following n-gram length 3
Code:
001 000 0.00%
001 001 21.74%
001 010 11.59%
001 011 15.94%
001 100 24.64%
001 101 26.09%
001 110 0.00%
001 111 0.00%
010 000 0.00%
010 001 0.00%
010 010 24.32%
010 011 25.68%
010 100 12.16%
010 101 10.81%
010 110 27.03%
010 111 0.00%
011 000 0.00%
011 001 48.68%
011 010 25.00%
011 011 26.32%
011 100 0.00%
011 101 0.00%
011 110 0.00%
011 111 0.00%
100 000 0.00%
100 001 0.00%
100 010 0.00%
100 011 0.00%
100 100 24.68%
100 101 28.57%
100 110 46.75%
100 111 0.00%
101 000 0.00%
101 001 22.22%
101 010 12.50%
101 011 23.61%
101 100 20.83%
101 101 20.83%
101 110 0.00%
101 111 0.00%
110 000 0.00%
110 001 0.00%
110 010 22.81%
110 011 28.07%
110 100 10.53%
110 101 15.79%
110 110 22.81%
110 111 0.00%The simplest explanation that would produce these distributions is this pseudocode.
Code:
If previous two digits are the same {
next digit must be different
}
else {
next digit picked randomly
}This doesn't mean that this the algorithm that is used, but simple explanations tend to be correct more often thean complicated ones.
As a result I stand by my initial conclusion that there is no message encoded in the data in this way.
I don't like the conclusion, but that is fair evidence for it, and I'm forced to consider the possibility much more likely. (I was toying with ideas of a repeating message encoded within a non-repeated string, but I suspect there would be more deviation in distribution than what you see). Nice work. A deception also dovetails with the tone's volume-envelope cutting off bits and corrupting the data through what (for a signal) looks more like sloppiness rather than anything useful or meaningful.
If so, then this was huge missed opportunity for Elite, and too big to just leave on the table - if nothing comes of it with 1.3, then I'll look into implementing this gameplay properly for the game franchise I design for
Last edited:
There has been some speculation in the main relic thread that Manchester encoding may be a factor, but this has issues with the transcripted data as it stands. I have a couple of ideas related to this which I intend to look at over the weekend. We'll see what those turn up. Interestingly I was drinking with a couple of my ore security-minded friends last night and this topic came up. It lead to some interesting discussion about the characteristics of encoded/enciphered information when viewed as raw data. So that was interesting...
It's also worth noting that I've been working off the great transcription work done by other people. I'm not listening to the sounds myself. These transcriptions are just one way of looking for information within one aspect of the sounds, high and low tones, and interpreting them in one manner, a binary string or strings. It's still possible that information could be present within other aspects of the sounds as a whole or when viewed in a different manner. What we're doing here isn't science, but in scientific investigation looking for something and not finding it can be equally as informative as finding it.
I've certainly enjoyed taking this raw data and looking for for indications that information is present without trying to impose a context on it (which is what I'm at least trying to do here), and I intend to carry on doing that as long as it's still fun and interesting. This whole thing has also made me much more aware of just how good the sound work is in this game. Sitting wearing the Rift with headphones on the middle of nowhere (I'm off exploring at the moment) and just parking the ship and listening to things actually made me physically shiver at one point this week. Great stuff! =)
This is the same conclusion as I've been coming to over the last few days; despite the work of the people who transcribed the recorded audio, I don't think we've been transcribing all or the right information.
I also think Xakaz may have been onto something here
Perhaps, for now, this is a better way of visually representing the audio without potentially ignoring two thirds of the data it may contain.
There is an additional (unlikely) possibility - if the data is compressed, then we would expect to see the same distribution as random (because random distribution cannot be compressed further, meaning that compression algorithms produce output closer and closer to random distribution the more effective they become) however I think a compression scheme is unlikely because it would be bad puzzle design (unless the plan is to hand us decompression-method clues at a later date), so I'm not going to investigate that avenue, I'm just assuming that it's not the case.
Unless the 'compression' is something similar to 8B/6T which uses three states (- 0 +) rather than two for binary:
http://www.wildpackets.com/resources/compendium/fast_ethernet/signal_4b5b
This is also the case if the information is enciphered in a way the meets the criteria for Perfect Secrecy such as using a one-time-pad. This was something I wasn't aware of until the discussion I had in the pub last night.
However, if there's information there but enciphered then we're really up against it, but I don't think that this is the case.
Hi Guys,
Been a bit out of the loop recently however I've been busy behind the scenes.
I've been putting together a bit of software that will model the recorded audio sequences in order to facilitate any kind of patterns analysis we want. It imports label exports from Audacity to model the sequences in a structured (and most importantly queryable format) with start and end time stamps, along with other metadata, for all elements.
It is built in .Net and the object model and import process is almost complete. I haven't started on a UI yet, but Linq can be used to analyse the sequences and output the results to console. I will get it on Git as soon as I am able. Anyone any good at .Net UI design, as this is likely to be the bit that takes me longest to get in place!
Another benefit of this approach is that we can establish a set of transcriptions using audacity projects and an agreed set of label tags. We can use these to discuss/agree message content and I hope can develop into an approach to establishing a standard data-set as there seems to be quite a lot of inconsistency in the current transcriptions.
Marking up a recording in Audacity using labels is pretty straight forward, so as new recordings come it it will be easy to add them to this common dataset.
Anyone got any views on useful properties to capture for the audio elements in addition to start and end timestamps? At the moment I have sequences, which are formed of Segments. Segments start with a chitter and and with a Howl. They contain a collection of Purrs which can be any length.
Sequences are flagged with Recorder, System, Capture Time, and a decription
Howls just capture whether they are type 1 or 2.
Purrs capture an indentifer for the bit value, whether they are clipped/quiet and can also capture the frequency (if anyone can be bothered to work that out)
Chitters are just time stamps
Does this sound useful?
Cheers
Cmdr Phyl
Been a bit out of the loop recently however I've been busy behind the scenes.
I've been putting together a bit of software that will model the recorded audio sequences in order to facilitate any kind of patterns analysis we want. It imports label exports from Audacity to model the sequences in a structured (and most importantly queryable format) with start and end time stamps, along with other metadata, for all elements.
It is built in .Net and the object model and import process is almost complete. I haven't started on a UI yet, but Linq can be used to analyse the sequences and output the results to console. I will get it on Git as soon as I am able. Anyone any good at .Net UI design, as this is likely to be the bit that takes me longest to get in place!
Another benefit of this approach is that we can establish a set of transcriptions using audacity projects and an agreed set of label tags. We can use these to discuss/agree message content and I hope can develop into an approach to establishing a standard data-set as there seems to be quite a lot of inconsistency in the current transcriptions.
Marking up a recording in Audacity using labels is pretty straight forward, so as new recordings come it it will be easy to add them to this common dataset.
Anyone got any views on useful properties to capture for the audio elements in addition to start and end timestamps? At the moment I have sequences, which are formed of Segments. Segments start with a chitter and and with a Howl. They contain a collection of Purrs which can be any length.
Sequences are flagged with Recorder, System, Capture Time, and a decription
Howls just capture whether they are type 1 or 2.
Purrs capture an indentifer for the bit value, whether they are clipped/quiet and can also capture the frequency (if anyone can be bothered to work that out)
Chitters are just time stamps
Does this sound useful?
Cheers
Cmdr Phyl