Everyone -
We have a possible breakthrough regarding the message format from Unknown Artifacts... please see the reddit thread here!
https://www.reddit.com/r/EliteDangerous/comments/34wdza/unknown_artifact_decryption_breakthrough/
63 Bits...
Although I've yet to solve this mystery, I think I've figured out how to decrypt the artifact signals, and the message packet format.
Looking at the bit transcriptions on the other thread, something immediately jumped out at me. Take the following transmit burst:
Not all the transmission bursts have this exact format, but I'll assume this is the most correct at present (I'll explain why later). I believe that people have correctly identified the first part of the message as a header -- let's look at that:
Translated into decimal, those are
Hmm... not terribly useful at a glance. But let's examine the rest further. The most common case of what follows involves a series of nine 7-bit sub-bursts, which is what I believe can be proven to be a correctly transcribed message. Let's count the total bits:
**And there it is. 36=63 right in the header!!!** I appears that the actual decimal is reverse encoded by order of magnitude -- just reverse the numbers!
Additionally, 63 = 3 x 21, which together with the initial header digit (3) would appear an immediate indicator that the message is in fact an encoded 3-space coordinate value
Given this, here is the complete data, with each 7-bit value raw converted, followed by the reverse:
One question is whether the individual numbers above should be left as individual numbers, or taken as a whole. If left as individuals, then one question is whether, like their digits, each sequence of 3x7 bits is also reverse encoded.
Alternatively, here it is decoded, with each set of 21 bits taken together:
This assumes non-signed numbers, which may not be useful. Instead, we may need to play with the first or last bits as sign bits, making each digit 20 bits long + sign. Also, the values are rather large (if they in fact represent coordinates in LY) so perhaps the last digit (or more) are fractional?
I haven't gotten that far yet myself, I got too excited and get this online... **And that's why I'm posting, because we'll get there faster all working together!**
* *Important: A final bit of business, the header format should allow us to go back into the audio samples, and correctly retrieve the necessary contents -- I noticed that some bursts indicated 63 bits when in fact fewer were transcribed. If all of this is correct, we should be able to verify the number of message bits from the header.*
We have a possible breakthrough regarding the message format from Unknown Artifacts... please see the reddit thread here!
https://www.reddit.com/r/EliteDangerous/comments/34wdza/unknown_artifact_decryption_breakthrough/
63 Bits...
Although I've yet to solve this mystery, I think I've figured out how to decrypt the artifact signals, and the message packet format.
Looking at the bit transcriptions on the other thread, something immediately jumped out at me. Take the following transmit burst:
Code:
011
100100
0110101
0100100
1001011
1100110
1010010
1010110
0011001
0110011
0110110
Not all the transmission bursts have this exact format, but I'll assume this is the most correct at present (I'll explain why later). I believe that people have correctly identified the first part of the message as a header -- let's look at that:
Code:
011
100100
Translated into decimal, those are
Code:
3
36
Hmm... not terribly useful at a glance. But let's examine the rest further. The most common case of what follows involves a series of nine 7-bit sub-bursts, which is what I believe can be proven to be a correctly transcribed message. Let's count the total bits:
Code:
7 x 9 = 63
**And there it is. 36=63 right in the header!!!** I appears that the actual decimal is reverse encoded by order of magnitude -- just reverse the numbers!
Additionally, 63 = 3 x 21, which together with the initial header digit (3) would appear an immediate indicator that the message is in fact an encoded 3-space coordinate value
Given this, here is the complete data, with each 7-bit value raw converted, followed by the reverse:
Code:
011 3 3
100100 36 63
0110101 53 35
0100100 36 63
1001011 75 57
1100110 102 201
1010010 82 28
1010110 86 68
0011001 25 52
0110011 51 15
0110110 54 45
One question is whether the individual numbers above should be left as individual numbers, or taken as a whole. If left as individuals, then one question is whether, like their digits, each sequence of 3x7 bits is also reverse encoded.
Alternatively, here it is decoded, with each set of 21 bits taken together:
Code:
011 3 3
100100 36 63
0110101 873035 530378
0100100
1001011
1100110 1681750 0571861
1010010
1010110
0011001 416182 281614
0110011
0110110
This assumes non-signed numbers, which may not be useful. Instead, we may need to play with the first or last bits as sign bits, making each digit 20 bits long + sign. Also, the values are rather large (if they in fact represent coordinates in LY) so perhaps the last digit (or more) are fractional?
I haven't gotten that far yet myself, I got too excited and get this online... **And that's why I'm posting, because we'll get there faster all working together!**
* *Important: A final bit of business, the header format should allow us to go back into the audio samples, and correctly retrieve the necessary contents -- I noticed that some bursts indicated 63 bits when in fact fewer were transcribed. If all of this is correct, we should be able to verify the number of message bits from the header.*
Last edited: