UNKNOWN ARTIFACT: Signal Fomat Decrypted?

[Ipecac]

If the message is in binary, we could get numbers from the signal.. The problem I'm having when I think about it is. What frame of reference could we possibly share with an alien species, in which to apply numbers?
Time and space could easily be experienced in a different way.

The first thing I was expecting was a countdown or count, since sequences of numbers require no unit measurement or human coordinate system.

Pulsars as landmarks in space is a great, universal feature which could be used to indicate locations (duh, thanks everyone who has come before, and Carl)

Is there a list of know pulsar EM burst intervals? That might be a good place to look for numbers that match (this i'm sure everyone has already looked for).

Also, i noticed that the big, less frequent honks are composed of a series of ticks that slow in frequency approaching the end of the tone. Could this be significant?
I'm just curious about features of pulsar EM transmissions like drifting modes.

Could the drift of a pulsar burst be used as an identifier if a species didn't have our same limited perception of time? If so, could something that be identified from ratios / fade-in or out of tones in the message?

Dont' make fun of me. I'm a Zoologist, not a Physicist!

P.S.
In looking for something about pulsar EM burst frequencies online, I found a paper which talked about "pulse nulls" in which bursts are still present but very faint. This makes me think of the audio file again, in the place where the binary notes become faint. Could the frequency of pulse nulls be represented in the signal and used as a pulsar identifier?

Don't make fun of me.

 

[retslag1]

I think it was supposed to be 42.

 

[Artie]

Is there a list of know pulsar EM burst intervals? That might be a good place to look for numbers that match (this i'm sure everyone has already looked for).

I tried it against set of 150 pulsars, no interesting results...

But, now I noticed another interesting thing - I filtered just full 7bit sequences and tried brief analysis. What I noticed is, that there is pretty limited set of numbers and are often repeating. More precisely, count of unique numbers is near [a-z0-9] set. Also, very frequent there is number 38 (0100110), which may lead to pretty simple substitution cipher (38 = letter "e"). I tried some initial deciphering, but it's time consuming (and I have no time ATM), so no results so far. Also, there is a question why Thargoid "beacon" should cipher english text. But, let's take it as a challenge, if anybody wants to try if it will lead to somewhere or not.

 

[Mysturji]

I think it was supposed to be 42.

You mean 101010

Why are some "bytes" 6 bits and others 7 bits? And shorter ones as well?
Has anyone considered just taking all the bits as a continuous stream and chopping them up into 8 bit bytes, or 4 bit nibbles? And reversing the whole thing?

 

[Ulysses_W]

It's at times like this that we need a mind like Alan Turing (RIP) to step up.

 

[Sentience]

You mean 101010

Why are some "bytes" 6 bits and others 7 bits? And shorter ones as well?
Has anyone considered just taking all the bits as a continuous stream and chopping them up into 8 bit bytes, or 4 bit nibbles? And reversing the whole thing?

You get r–•ž£Rš¶$‰%æ¥XË6 if it's reconstructed in 8 bit format.
Reversed it's lÓ¥g¤‘$mYJÅy©iHN.

 

[Ipecac]

So when i take a pen to marking the 1s and 0s of this message, with a little dot over the place where the digestive tone comes in, the big sound seems to come at the end of every binary phrase, but it oscillates (or seems to) between the second to last digit, the last digit and in between the two. Does anyone else think this is so? Also, it does not jump randomly in the sequence, it moves from the 6th digit, to 6.5, to 7th digit and back. Makes me think precession or wobble from a planet, or drift of a pulsar?
OR, since we've all thought plenty about C.Sagan's Contact signal, what if the location of the large tone was equivalent to the Contact page marker symbols and denoted some kind of dimensional orientation or direction for each phrase.. OR positive and negative?

Also, I marked 0100111 in the fourth complete phrase. does anyone else have this triple digit?

 

[Mr_Zeno]

When we have sent a transmission into space we've used a semi-prime as the coding mechanism, the factors in the semi-prime are used to determine the dimensions of the transmitted message grid.
All the blocks of code we have are different sizes and as such we have no idea on the sequence the blocks might be placed.

 

[Ipecac]

When I translate the sound file, my 6 digit phrases say "36" followed by six, seven digit phrases then "37"
011 (ignored)
100100 (36)
0010010
1001011
0100111
0110011
1101010
0011010
100101 (37)
0110101
0110110

also there's a triple digit in my accounting which nobody else seems to have, so I may be wrong about that.
UNFORTUNATELY, I don't know where to find more complete recordings to look for patterns. can you link a collection of recordings?

 

[Mr_Zeno]

When I translate the sound file, my 6 digit phrases say "36" followed by six, seven digit phrases then "37"
011 (ignored)
100100 (36)
0010010
1001011
0100111
0110011
1101010
0011010
100101 (37)
0110101
0110110

also there's a triple digit in my accounting which nobody else seems to have, so I may be wrong about that.
UNFORTUNATELY, I don't know where to find more complete recordings to look for patterns. can you link a collection of recordings?

If other number blocks can be identified we can find the start, have some idea on the number of blocks, and which blocks to search for.
If a block can be found with 38 it would be useful.

 

[Archvillain]

Also, I marked 0100111 in the fourth complete phrase. does anyone else have this triple digit?

You'll need to include info on which video or audio you're transcribing - the forum posts are happening too fast and thick for many of us to work it out from from context in previous posts. :/

I have seen triple-digits where not everyone agreed they were there. If we find a triple digit that is clearly there, that's definitely a point of interest.

 

[Petersen]

Yes when this unravles i hope the people that put time into all this analyzing dont have bad feeling if the solution is a simple one, or just not even implemented yet. I have really enjoyed this search and reading about ppls work and i think the comunity have never seemed so vital and alive, and its nice to see something productive going on in the forums instead of the usual isurance made me quit or lack of content threads. In short im greatfull even if there was never anything grand planned on FD's part, and i hope everyone els will be too.

I can only speak for myself, but I'm having fun in the now trying to analyze and solve the puzzle. There might not even BE a puzzle. I realize this, and I won't have a bad feeling if it turns out the sounds really are just random nonsense.

Right, monster post here

DATA! This is what we need more of! 24 samples is nice, but more is always better. We don't know how long the message actually is, so until we start actually finding repeating patterns (which I'm about to attempt with your data), we NEED. MOAR. DATA! =D

 

[Petersen]

Double-posting because for some reason editing will break all my line-breaks.

I took Digital scream's data into a spreadsheet program, put each string of data side-by-side, and applied some conditional formatting to shade cells containing d3 different 6-character strings and their Two's Complement (001001, 001100, 010101, and then 110110, 110011, 101010).

I used six-character strings because some entries are only six characters, and it's easy enough to look at the seventh character when there's a seventh and see which ones are the same.

Here's a link to the google doc.

https://docs.google.com/spreadsheets/d/1-mV_NxlUnoUvPOwkLb83NYMhZ89txIlgC4EPa81Dny4/edit?usp=sharing

I'm really not seeing any repetition in the sequences.

 

[Ipecac]

You'll need to include info on which video or audio you're transcribing - the forum posts are happening too fast and thick for many of us to work it out from from context in previous posts. :/

I have seen triple-digits where not everyone agreed they were there. If we find a triple digit that is clearly there, that's definitely a point of interest.

So the triple digit frame I was referring to was the fourth phrase in this initial MP3 posted on reddit.. Seems to be the one that "started it all"
https://www.dropbox.com/s/63xxqfopes427xh/unknown_artifact_audio_long-200pct.wav?dl=0

so, disregarding the first 3 digit piece as a fragment, I am referring to the fourth phrase, the first being the 6 digit 100100

011 (fragment)
100100 (36)
0010010
1001011
0100111 (triple digit)
0110011
1101010
0011010
100101 (37)
0110101
0110110

file interrupted by cargo doors

 

[Petersen]

0100111 (triple digit)

To my ears, that's incorrect, it's 0100101.

 

[Petersen]

I would very much like to know if the same artifact 'broadcasts' the same message to multiple players observing it at the same time. If it does not, I think this is compelling evidence that the sound is just noise.

I also think my inability to piece together even two of the twenty-four samples Digitalscream posted is evidence that it is just noise, or the message is very, very long. I'm not sure on the probability, but it sure seems that I should be able to match part of one to part of another in 24 samples, but I can't.

EDIT: HOLY CHEESE. I THINK I FOUND SOMETHING.

So I decided I'd take Digitalscream's data and look at the whole set of 1s and 0s as a single unit and try to match up patterns that way. And I think I found something.

"1001110110010010110" is found at the beginning of the 9th set, and at the end of the last set. I'd love to listen to these to see if it's all timed out the same (but I'm not sure which numbers correspond to which audio files), especially since the ninth set has a leading 1 to the specified sequence which I hope is in error, because the preceeding number in the last set is 0. =-\

Still investigating.

EDIT 2: Just look at the data first as it's broken up between howls. It's offset a bit. Still investigating this line, having a hard time finding any other long sequences that match up. tail-to-front. So could just be a 'coincidence'.

EDIT 3: So after spending an hour or more tinkering with this, I'm pretty confident that none of the broadcasts can be matched together. I wonder if perhaps the howls indicate the 'polarity' of the phrases...

 

[Artie]

The message is probably very long. Just some frequency analysis results I mentioned earlier:

Spoiler

just 7bit chunks used, ordered by converted number
format: converted bits to dec => number of occurences
10 => int 1
18 => int 4
19 => int 3
20 => int 3
22 => int 3
25 => int 8
26 => int 4
27 => int 4
36 => int 4
37 => int 4
38 => int 12
41 => int 2
42 => int 1
43 => int 5
44 => int 3
45 => int 4
50 => int 5
51 => int 9
52 => int 4
53 => int 6
54 => int 8
73 => int 7
74 => int 4
75 => int 8
76 => int 8
77 => int 11
82 => int 4
83 => int 3
84 => int 3
85 => int 1
86 => int 8
89 => int 9
90 => int 4
91 => int 5
100 => int 4
101 => int 3
102 => int 5
105 => int 3
106 => int 1
107 => int 1
108 => int 3
109 => int 2

These are definitely not a classic random numbers. So there are two options:
- it means something
- it means nothing and it is generated by pseudo-random algorithm, similar as stellar forge.

Character substitution still doesn't work, but it's also very hard to get some lead from it, as there are no long "sentences" in the samples and also with all those fuzzy parts. But, maybe someone get another idea based on this.

 

[bioid]

When I translate the sound file, my 6 digit phrases say "36" followed by six, seven digit phrases then "37"
011 (ignored)
100100 (36)
0010010
1001011
0100111
0110011
1101010
0011010
100101 (37)
0110101
0110110

also there's a triple digit in my accounting which nobody else seems to have, so I may be wrong about that.
UNFORTUNATELY, I don't know where to find more complete recordings to look for patterns. can you link a collection of recordings?

Here's some more recording from digitalscream: http://www.digitalscream.org.uk/audio/

It would be nice to find that the 6 bit chunks are part of a 'header'. This reminds me of the sequence number in TCP headers: http://www.freesoft.org/CIE/Course/Section4/8.htm

 

[Archvillain]

I would very much like to know if the same artifact 'broadcasts' the same message to multiple players observing it at the same time. If it does not, I think this is compelling evidence that the sound is just noise.

I also think my inability to piece together even two of the twenty-four samples Digitalscream posted is evidence that it is just noise, or the message is very, very long. I'm not sure on the probability, but it sure seems that I should be able to match part of one to part of another in 24 samples, but I can't.

Great work! I don't agree that it suggests noise, but it definitely suggests we may be barking up the wrong tree for how to find a consistent signal or start/finish.

Here is a long shot because even if it's true it should have still shown up with what you've done, but if you've sufficiently automated the process maybe it's worth it - perhaps there are matches between samples and inverted samples? For some kinds of encoding with this kind of structure, such as recording data to tape, a 1 or 0 = 0, and a 11 or 00 = 1, therefore a message could overlap regardless of whether the tones are inverted or not.


I'm starting to think about looking for a key. Eg. if we found that XORing the first block against the remaining blocks suddenly turned all UA sequences into the same sequence, we'd know we'd hit something. That particular example seems arbitrary and ugly and shoehorned to me, but maybe there is something similar yet more elegant... still thinking...

 

[bioid]

Great work! I don't agree that it suggests noise, but it definitely suggests we may be barking up the wrong tree for how to find a consistent signal or start/finish.

Here is a long shot because even if it's true it should have still shown up with what you've done, but if you've sufficiently automated the process maybe it's worth it - perhaps there are matches between samples and inverted samples? For some kinds of encoding with this kind of structure, such as recording data to tape, a 1 or 0 = 0, and a 11 or 00 = 1, therefore a message could overlap regardless of whether the tones are inverted or not.


I'm starting to think about looking for a key. Eg. if we found that XORing the first block against the remaining blocks suddenly turned all UA sequences into the same sequence, we'd know we'd hit something. That particular example seems arbitrary and ugly and shoehorned to me, but maybe there is something similar yet more elegant... still thinking...

I think the alternating tone theory is a good shot. A few people have suggested it sounds like like a tape or modem when played at higher speeds, which suggests it could be a carrier wave. Perhaps we can demodulate it somehow.