SGP
Halo ODST => SGP Task Force: 7 => Topic started by: Dunder Moose on October 25, 2009, 23:12:28
-
There is a J sign that has me puzzled. In sector 5 there are J signs with code above the J. Here's a couple looks:
http://www.bungie.net/Online/Halo3UserContentDetails.aspx?h3fileid=98938695
And here's another sign in the same sector
http://www.bungie.net/Online/Halo3UserContentDetails.aspx?h3fileid=98938339
At first I thought it was definitely a Baudot code, but it only has four bits instead of the standard five. I though it might be Braille, but those are only three deep and two wide. So, one less than Braille, one more than Baudot. Is it significant that Baudot's code is based on an earlier code developed by Carl Friedrich Gauss and Wilhelm Weber in 1834? There is a thank you to the Gauss Ammunition people in the end credits.
I'll keep on working on it. I wonder if the barrier codes might provide the missing dots in the Baudot somehow? Has anyone seen any more of these dot images in the other sectors? And does anyone else have the UNSC Eagle image with Spoiler instead of the picture in your Bungie.net gallery for in game photos?
Here is a web page with a Baudot code key about halfway down.
http://www.diycalculator.com/popup-h-paper.shtml
-
We knew (http://subculturelifestyle.com/forum/index.php?topic=28.0) about the banners and the possible codes therein, but we didn't know about gauss code.
Brilliant connection with the credits!
My initial research is a bit disheartening. Gauss' system (actually Gauss/Weber) was originally based on needle rotations (left or right). A sequence of rotations (between 1-4) encoded a letter or number. Thirty different codes were possible. The disheartening part is that there are essentially three possible values in each position (nothing, left, right). With only dots and blanks, the banners don't convey the difference between the "left" and "right" of the Gauss/Weber code. You cannot assume that blank spaces are one or the other, because you won't be able to discern the difference between a series of two blanks and a series of three!
The system is still in use for historical reasons- a laser beam sends Gauss/Weber encoded messages by mapping left and right to short pulses and long pulses in Goettingen (http://www.measurement-valley.de/mv/index.php?id=308).
I also compared the banner codes against the Baudot encoding in case it was possible to drop one of the five lines and still convey a full alphabet. Using the standard encoding, it is not.
I wouldn't expect the barrier codes to come into play here. But expectations are dangerous! Thanks for revitalizing this topic! And welcome to the SGP!
-
Thanks ColdGlider. Good to be here!
-
I noted a similar thing with the J code but didn't get as far, this prompted me to do some reading on Gauss(bored...)
well after some fruitless reading, I followed a link to a link and found this...
http://www.mathsisfun.com/numbers/convert-base.php?to=quaternary
there's a pulldown to switch between any given number of a code
The thing about the J code...it could be 5 bits 01234, or the empty column could just be a space and end up being 0123 for the rest.
I posted this in that other thread somewhere, but it could be useful also:
http://home2.paulschou.net/tools/xlate/
That's a code convertor, but not numbers, it handles text as well, IE
Omeganuepsilon = 01001111 01101101 01100101 01100111 01100001 01101110 01110101 01100101 01110000 01110011 01101001 01101100 01101111 01101110
The J code, could be the conversion of a conversion, but I'm too tired.
-
Thanks for dropping in again, Omega! And thanks for the base conversion link!
I'm going to lock this thread since the older one has more info at this point. So please resume the conversation over here (http://www.gruntspajamas.com/forum/index.php?topic=28.15).
-
I changed my mind. I would like to maintain a topic in Task Force: 7 for each of the "J Banners". This post is a perfect place for discussion on the banner that Dunder Moose provided image links to (now also in the gallery (http://www.gruntspajamas.com/gallery/v/SGP/companies/Jotun/).)
Each of the three banners discovered so far has a unique signature in the fourth column. (They differ elsewhere as well, but that is the first column which differs among all three.) If you decode these columns into binary (Top-Bottom, Little Endian, Dot=1) then you can obtain three unique numbers which can be used to identify the banners easily: 7, 0, 15. This is especially nice since each of the extremes (as well as the number 7!) is represented.
Based on this unique property, I'm naming the banner cataloged by Dunder Moose "J Banner 7" and renaming this topic's subject accordingly. I'm also happily unlocking it.
:)
-
(http://www.gruntspajamas.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=1218&g2_serialNumber=2)
Way back on October 8th I described (http://www.gruntspajamas.com/forum/index.php?topic=28.msg482#msg482) what I thought would be the required specifications for an analysis tool capable of tackling the binary data we assumed to be contained in the J-Banners. I began writing that tool the other night and the current version allowed me to produce the following data for J Banner 7.
I wanted to catalog the 4-bit values encoded in the banner columns. For each of four possible interpretations, a comma separated list is provided. There are three fields: column index, binary representation, decimal value.
Please note that maintaining Little Endian bit order while changing scan direction (top-bottom / left-right) is functionally equivalent to maintaining the scan direction and altering the endianness.
J Banner 7
Transcribed Data (Dot=1)
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
Decoded by Columns: Dot=1, L-R, T-B, Little Endian
01,0011,3
02,0100,4
03,0011,3
04,0111,7
05,0010,2
06,1011,11
07,1110,14
08,1001,9
09,0101,5
10,0110,6
11,1111,15
12,0101,5
13,1011,11
14,0011,3
15,0000,0
16,1011,11
17,1110,14
18,0101,5
19,0010,2
20,1011,11
21,0111,7
22,1101,13
23,1011,11
24,0111,7
25,1001,9
26,0110,6
27,1101,13
28,1010,10
Decoded by Columns: Dot=1, L-R, B-T, Little Endian
01,1100,12
02,0010,2
03,1100,12
04,1110,14
05,0100,4
06,1101,13
07,0111,7
08,1001,9
09,1010,10
10,0110,6
11,1111,15
12,1010,10
13,1101,13
14,1100,12
15,0000,0
16,1101,13
17,0111,7
18,1010,10
19,0100,4
20,1101,13
21,1110,14
22,1011,11
23,1101,13
24,1110,14
25,1001,9
26,0110,6
27,1011,11
28,0101,5
Transcribed Data (Dot=0)
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
Decoded by Columns: Dot=0, L-R, T-B, Little Endian
01,1100,12
02,1011,11
03,1100,12
04,1000,8
05,1101,13
06,0100,4
07,0001,1
08,0110,6
09,1010,10
10,1001,9
11,0000,0
12,1010,10
13,0100,4
14,1100,12
15,1111,15
16,0100,4
17,0001,1
18,1010,10
19,1101,13
20,0100,4
21,1000,8
22,0010,2
23,0100,4
24,1000,8
25,0110,6
26,1001,9
27,0010,2
28,0101,5
Decoded by Columns: Dot=0, L-R, B-T, Little Endian
01,0011,3
02,1101,13
03,0011,3
04,0001,1
05,1011,11
06,0010,2
07,1000,8
08,0110,6
09,0101,5
10,1001,9
11,0000,0
12,0101,5
13,0010,2
14,0011,3
15,1111,15
16,0010,2
17,1000,8
18,0101,5
19,1011,11
20,0010,2
21,0001,1
22,0100,4
23,0010,2
24,0001,1
25,0110,6
26,1001,9
27,0100,4
28,1010,10
-
:( What ever happened to good old fashion Tea Bagging?
(I will not be offended if you delete this, for the sake of keeping the it clean)
-
So I can't program like ColdGlider, but I know how to use Excel a little. Here is a graphic representation of the banners. The composite at the bottom shows the frequency for dots in each position. It seems likely since the codes are so similar that they do not give meaning when concatenated (Unless it is a repetitious phrase). Also by comparing them one to another, each position now has a possibility of 4 values: 0-3. I don't know what that opens up decryption-wise. I got excited when I learned that there is an 8-dot Braille that would match the depth of the banners, but I haven't found a good ascii mapping for it online.
If the attachment doesn't work you can view it here - http://www.flickr.com/photos/61024371@N00/4100217777
-
Well done, Mr. Moose! The image has been added to our gallery here (http://www.gruntspajamas.com/gallery/v/SGP/numbers/SL-2009_11_13-DunderMoose-JBannersExcel.jpg.html).
I've often wondered what binary operations on the data (i.e binary AND, binary OR) would produce. Moose's result set provides the equivalent of an AND operation if you convert the blocks equal to 3 to a 1 and the rest of the values to 0. To get the equivalent to an OR operation, you convert all non-zero values to 1. Performing each of these operations produces two "new" sets of banner dots:
J-Banners - Binary AND
1010010110011101010100000010
1010011001000001101010110101
0000001010010000110011010110
0000011000001001100101101011
J-Banners - Binary OR
1011010110111101010111111110
1011111111111111111111111101
0101001011111100110011011110
1111111100101101101101101011
I dropped these new data sets into my binary analyzer program and set it to decode ASCII in four different ways. I turned up a big nothing, but am providing the data for proof/reference. I also grouped the data into 7-bit bytes, performed the same eight decodings, and also got nothing. If anyone really wants that data I can provide it as well.
Below are the ASCII conversions of these new sets of dots, assuming 8-bit bytes. Values in brackets ([]) are within the ASCII range but are non-printable. Values in braces ({}) are outside the valid ASCII range. Valid ASCII values are decoded into the corresponding ASCII character.
J-Banner Binary AND ASCII Conversion: Dot=1, L-R, T-B, Big Endian
01,10100101,{165}
02,10011101,{157}
03,01010000,P
04,00101010,*
05,01100100,d
06,00011010,[26]
07,10110101,{181}
08,00000010,[2]
09,10010000,{144}
10,11001101,{205}
11,01100000,`
12,01100000,`
13,10011001,{153}
14,01101011,k
J-Banner Binary AND ASCII Conversion: Dot=0, L-R, T-B, Big Endian
01,01011010,Z
02,01100010,b
03,10101111,{175}
04,11010101,{213}
05,10011011,{155}
06,11100101,{229}
07,01001010,J
08,11111101,{253}
09,01101111,o
10,00110010,2
11,10011111,{159}
12,10011111,{159}
13,01100110,f
14,10010100,{148}
J-Banner Binary AND ASCII Conversion: Dot=1, L-R, T-B, Little Endian
01,10100101,{165}
02,10111001,{185}
03,00001010,[10]
04,01010100,T
05,00100110,&
06,01011000,X
07,10101101,{173}
08,01000000,@
09,00001001,[9]
10,10110011,{179}
11,00000110,[6]
12,00000110,[6]
13,10011001,{153}
14,11010110,{214}
J-Banner Binary AND ASCII Conversion: Dot=0, L-R, T-B, Little Endian
01,01011010,Z
02,01000110,F
03,11110101,{245}
04,10101011,{171}
05,11011001,{217}
06,10100111,{167}
07,01010010,R
08,10111111,{191}
09,11110110,{246}
10,01001100,L
11,11111001,{249}
12,11111001,{249}
13,01100110,f
14,00101001,)
J-Banner Binary OR ASCII Conversion: Dot=1, L-R, T-B, Big Endian
01,10110101,{181}
02,10111101,{189}
03,01011111,_
04,11101011,{235}
05,11111111,{255}
06,11111111,{255}
07,11111101,{253}
08,01010010,R
09,11111100,{252}
10,11001101,{205}
11,11101111,{239}
12,11110010,{242}
13,11011011,{219}
14,01101011,k
J-Banner Binary OR ASCII Conversion: Dot=0, L-R, T-B, Big Endian
01,01001010,J
02,01000010,B
03,10100000,{160}
04,00010100,[20]
05,00000000,[0]
06,00000000,[0]
07,00000010,[2]
08,10101101,{173}
09,00000011,[3]
10,00110010,2
11,00010000,[16]
12,00001101,[13]
13,00100100,$
14,10010100,{148}
J-Banner Binary OR ASCII Conversion: Dot=1, L-R, T-B, Little Endian
01,10101101,{173}
02,10111101,{189}
03,11111010,{250}
04,11010111,{215}
05,11111111,{255}
06,11111111,{255}
07,10111111,{191}
08,01001010,J
09,00111111,?
10,10110011,{179}
11,11110111,{247}
12,01001111,O
13,11011011,{219}
14,11010110,{214}
J-Banner Binary OR ASCII Conversion: Dot=0, L-R, T-B, Little Endian
01,01010010,R
02,01000010,B
03,00000101,[5]
04,00101000,(
05,00000000,[0]
06,00000000,[0]
07,01000000,@
08,10110101,{181}
09,11000000,{192}
10,01001100,L
11,00001000,[8]
12,10110000,{176}
13,00100100,$
14,00101001,)
-
Great work ColdGlider!
Does Octadecimal binary map to ASCII very well? I'm still kinda new to this process but. If 0=00 and 1=01, 2=10 and 3=11 then we have doubled the length of the binary we can get from each column.
the first column being a 3,3,0,1 would come out: 11110001, and so on. I can try to hand-crunch these here at work on lunch, but CG's program would probably be a reliable backup to check my work.
-
I have no idea what you guys are talking about in this thread. That just makes it more awesome though.
-
:-\ ??? same lol. im twelve and what is this? XD
-
It's okay if you don't totally get it, just nod you'r head and say something like "Good working" and stuff. ;D
-
Basically I took all three known J Banners and overlaid them. I added the dots for each position. The top left position has a dot in all three J Banners, so I gave it a value of 3. If there was a position that had no dots in any banner I gave it a 0. If there was a position that had only 1 or 2, I gave those numbers. So for example if you look at the first column of all three banners, they all have the first two positions dotted, the third empty and only the 3rd J Banner has a dot in the bottom. So it goes 3,3,0,1 for the combined first column.
Proceeding thusly I converted 0 to 00, 1 to 01, 2 to 10, and 3's became 11's.
Here are the bytes that the overlay gives.
11110001
00001001
11110001
10101001
00100001
11110011
00111111
11010010
11011100
00111000
10101010
11011100
11100111
11101001
00010000
11110011
00111111
11011100
00110001
11100011
10111100
10011111
10110011
10111100
10010111
01111100
11001111
00110011
I am not skilled to convert this binary into text. Everything I've seen online is for hexadecimal (six positions) binary conversion and not full 8. So I get a differing final three numbers all preceeded by &#. Hopefully somebody with the skills ta pay da bills can crunchify this. Otherwise I am educating myself as fast as I can.
-
Converting to ASCII is a trivial task with my analyzer. The processed data is at the end of this post in both bitwise endian decodings. I really need to work on getting the UTF decodings in there, too.
Anyhow, I'm sorry to report that the results of your transformation are not ASCII, Moose. :( A quick way to tell if a byte is not ASCII is to check the most significant bit (MSB). Conventionally, this is the leftmost bit. If it is a 1, then the value of the byte will be 128 or greater. Values this large are outside of the range of ASCII. This does not mean that a value of 0 in the leftmost position will be ASCII- just that it can be decoded as such whether the original data was meant to be ASCII or not.
The hunt continues...
Big Endian Bit Order
01,11110001,{241}
02,00001001,[9]
03,11110001,{241}
04,10101001,{169}
05,00100001,!
06,11110011,{243}
07,00111111,?
08,11010010,{210}
09,11011100,{220}
10,00111000,8
11,10101010,{170}
12,11011100,{220}
13,11100111,{231}
14,11101001,{233}
15,00010000,[16]
16,11110011,{243}
17,00111111,?
18,11011100,{220}
19,00110001,1
20,11100011,{227}
21,10111100,{188}
22,10011111,{159}
23,10110011,{179}
24,10111100,{188}
25,10010111,{151}
26,01111100,|
27,11001111,{207}
28,00110011,3
Little Endian Bit Order
01,10001111,{143}
02,10010000,{144}
03,10001111,{143}
04,10010101,{149}
05,10000100,{132}
06,11001111,{207}
07,11111100,{252}
08,01001011,K
09,00111011,;
10,00011100,[28]
11,01010101,U
12,00111011,;
13,11100111,{231}
14,10010111,{151}
15,00001000,[8]
16,11001111,{207}
17,11111100,{252}
18,00111011,;
19,10001100,{140}
20,11000111,{199}
21,00111101,=
22,11111001,{249}
23,11001101,{205}
24,00111101,=
25,11101001,{233}
26,00111110,>
27,11110011,{243}
28,11001100,{204}
-
Rats, that really felt like I was on to something there... Why would they have such a tantalizing code, and three versions of it if it didn't add up to something more than "it looked cool"?
Other options that present: 8 Dot Braille, Binary clock, I don't know. Might take some pajama time before I build up the steam to tackle this again.
-
I had another take on the dots. There are 28 in a row. What if they represent the alphabet with two other characters, maybe space and period, or paragraph? Then maybe the missing dots highlight the letter for that position.
if
abcdefghijklmnopqrstuvxyz_.
* ** * ** **** * * ****** * (the first line in the J Banner 7 code.)
So the standouts are: be gj oq sz.
So I don't have my alphabet lined up right.
OR
I'm not using the right alphabet. Swahili is a Bantu Language, but it only has 23 letters. But it is based in part on Arabic and that has.....28 letters.
Could the spaces in the dot codes highlight arabic letters that might spell something????
Am I going insane????
What do y'all think?
-
I think this is a very very good point Dunder. If I recall correctly the code for IWHBYD was in Greek or some such thing?
Especially since Arabic has 28 letters and there's 28 dots. I'm off to find an Arabic alphabet and online translator!
EDIT: A stumbling block: Arabic letters can be written in several ways depending on whether they are represented alone or at the beginning, middle, and end of words. The possibilities are staggering unless we had someone who could read Arabic.
-
I would say you'r going crazy, Moose, but if Apollos reference to IWHBYD is real, it could just be something like that.
-
I'm sorry but I cannot confirm this. I have tried searching for people referencing the search for IWHBYD but I can only find people who talk about looking at the code to discover it. Consider it hearsay for now.
-
This morning it occurred to me that the J banners may be part of a citywide sanitation system. The Oliphaunts that Jotun Heavy Industries manufacture seem to be unmanned in Sadie's story. Perhaps the dot code banners are there for them to scan, like checkpoints on their circuit of the city. A way to test this would be to map out which codes are in which locations and see if they are just Jotun code for the different Districts.
-
That's a good thought, but there are only three J-Banners and they often appear clustered in the same area. Squidhenge has two different ones, for example. I believe all three can be found in D01 at the beginning of CM:CH. But I like your theory!
-
This puts Braille back on the table..
Braille computer monitor
The Braille computer monitor has rows and columns of rectangular cells. The cells include four rows and two columns of dots that can be felt for interpretation by the user. "The pins are driven by electromechanical impact drivers and are held in position by resilient elastomeric cords. The impact drivers are carried on a bi-directional printhead which travels beneath the movable pins. An erasing mechanism is provided to positively drive the pins downwardly to erase the characters produced by the printhead."[3] The Braille computer monitor is under the United States Patent 6700553.[3
http://http://en.wikipedia.org/wiki/Refreshable_Braille_display
-
Good show Engima! I know a quy in my judo team who's blind, and he has told me about that. Never realized it would pop back up here. :D
So...How do we translate it then? ???
-
Well, I guess we would need that machine... or at least a detailed picture of the machine.
-
Well, I guess we would need that machine...
No, we shouldn't need the machine. We only need the specification for the 8-dot braille it produces.
Braille has been extended to an 8-dot code, particularly for use with Braille embossers and refreshable Braille displays. In 8-dot Braille the additional dots are added at the bottom of the cell, giving a matrix 4 dots high by 2 dots wide. The additional dots are given the numbers 7 (for the lower-left dot) and 8 (for the lower-right dot). Eight-dot Braille has the advantages that the case of an individual letter is directly coded in the cell containing the letter and that all the printable ASCII characters can be represented in a single cell. All 256 (28) possible combination of 8 dots are encoded by the Unicode standard. Braille with six dots is frequently stored as Braille ASCII.
So I thought perhaps we might simply drop the fourth row if all it contains are alternate versions (e.g. capitals) of the text.
Using the following links:
http://www.brailleauthority.org/
http://www.dotlessbraille.org/displays.htm
http://en.wikipedia.org/wiki/Braille_Ascii
http://acharya.iitm.ac.in/disabilities/asc_brl.php
...I converted J-Banner 7 to text as if it were normal 6-dot braille. I simply dropped the fourth row of dots. Here is what results:
| GLYPH # | DOTS | BRAILLE ASCII | LITERARY BRAILLE |
| 01 | 1-2-6 | < | gh |
| 02 | 1-2-4-5-6 | ] | er |
| 03 | 2-4-5 | J | j |
| 04 | 2-3-4 | S | s |
| 05 | 1-3-5-6 | Z | z |
| 06 | 1-2-3-4-6 | & | and |
| 07 | 1-2-4-5 | G | g |
| 08 | 4-5 | ^ | [contraction prefix] |
| 09 | 2-3-4-6 | ! | these |
| 10 | 2-4-5 | J | j |
| 11 | 1-2-3-4-6 | & | and |
| 12 | 1-2-4-5-6 | ] | er |
| 13 | 1-5-6 | : | wh |
| 14 | 1-3-5 | O | o |
So, yeah. "Gherjsz" to you, too, Bungie. Onward to the 8-dot spec.
any computer braille displays indicate capital letters with a dot on the lower left (dot 7). There is little consistency in the use of the lower two dots. There are a number of official and semi-official 8 dot Braille codes, mostly in Europe. Very little literature has been reproduced in any 8 dot Braille code.
The website quoted above (http://8dotbraille.com/) provides a totally different method of decoding. The bad news is that the very first 8-dot glyph we find in J-Banner 7 has no mapped value in 8-dot unified! (3-4-7 according to their dot numbering scheme) Maybe I missed something, but I did not find this pattern in their lists of letters (http://8dotbraille.com/8-dot-braille-capital-letters.html), numbers (http://8dotbraille.com/8-dot-braille-numbers.html), punctuation (http://8dotbraille.com/8-dot-braille-punctuation.html), or symbols (http://8dotbraille.com/8-dot-braille-symbols.html).
I looked further. This article (http://www.brailleauthority.org/eightdot.pdf) (http://www.brailleauthority.org/eightdot.pdf) is a great overview of Braille development and includes 8-dot systems. From this, I discovered ISO/TR 11548 which defines an 8-dot braille encoding method for computer use. And you know what? For $157 you can purchase that document from Amazon.com. The ISO site itself has a shopping cart application which they use to SELL you the standards documents.
So let me get this straight. If I'm a blind person who wants to read 8-dot braille, I need to pony up over $100 to purchase a document NOT PRINTED IN BRAILLE in order to figure it out?!? Sure, I understand that these documents cover the technical aspects of character encoding, blah, blah, blah, and the standards folks have to be reimbursed for their fancy French lunches as they sit in board rooms and vote on the acceptance of new proposals. That doesn't stop me from wanting to whack them upside their heads with a stale loaf of french bread for selling overpriced digital copies of what are supposed to be International Standards. Ok, flame off.
Anyhow, we're striking out (once again) on decoding this banner.
I might as well toss out another idea I had while we're on the topic of failed attempts: "BARRETT" as the key. I reasoned that since the J-Banners have rows of 28 dots and this number is evenly divisible by 7 (go figure) then the seven-character string "BARRETT" (revealed to us in-game though binary) might be the first word encoded in one of the J-Banners. This seemed logical to me: use an established encoding method (binary) to give us the means to decode your own custom encryption method (J-Banner dots.) Sounds fun and exciting, right?
If this were true, then perhaps the first group of four dots in one of the J-Banners is meant to decode to a "B", the second four an "A", the third and fourth four an "R", etc. If there was indeed a one-to-one mapping between the dot patterns and the characters, then we should observe a repeating dot pattern for the doubled letters in "BARRETT", namely the "RR" and the "TT". I don't see this pattern in any of the banners. Please feel free to check into this yourself and prove me wrong.
Finally, in case anyone is wondering "Why does he bother posting all of this stuff if it turns up nothing?"
1: To record that it has already been tried, how it was attempted, and what the results were so that others can save time.
2: For others to correct if I made a mistake or otherwise misinterpreted the data or the decoding processes.
There are still some leads in the Braille department. As the aforelinked article (http://www.brailleauthority.org/eightdot.pdf) describes, there are Braille encoding methods for music and math, for example. For now, though- I am spent. My shift is over. Next?
-
I am deleting my horrible paint shop graph now!.. I spent all day on it, that and going back and forth on Morse code, Braille and AC. I hate the J banners....makes me want to smoke a "J"( and no, I don't do drugs)
CG did you notice if language mattered?
...Maybe its just like the Pedestals.. just a couple of different versions of the same thing with no rhyme or reason. Maybe its purely aesthetic. :-[
-
Great idea Enigma and well done CG. Seriously impressed over here. Excellent.
The J Banners have resisted yet another attempt. Gherjsz-it!
I wonder if it isn't time to try brute force cryptography. I mean if there is a column to letter ratio we could see which patterns repeat most often and try to map them to the most often occurring English letters - ESTONIA-R. I am super busy this weekend, but if I get some time I will try to map this out and show my work, following ColdGlider's sterling example here. :)
Edit: Let's not toss out Braille hastily. Remember the J looks like an upside down seven. Tomorrow I will try to recreate ColdGlider's Braille work for the upside down version as well as forward and backward like we did with the binary attempts. Of course this will be between trying ideas on the Ubiquitous Number, of course...
-
...Maybe its just like the Pedestals.. just a couple of different versions of the same thing with no rhyme or reason. Maybe its purely aesthetic. :-[
Hmm, three J Banners and three Pedestals...and two J Banners in squidhenge. Hmmm. Maybe one under each J and one in the henge?
-
i wish... but we cant get the third one due to deloading. :-[
-
Okay, so two j banners in the henge and only two plinths at a time. Let's look at those Plinths and see if there's any marking on them that might associate them with the banners. If there is a connection we could get the correct plinths into the henge and see....
-
I still haven't figured out why you call it J Banner 7. Could someone please explain?
-
Due to being stuck in a place without my laptop, I can't make a nifty explanatory image. So I'll have to make a cheesy one (see attached).
The Problem: How to easily recognize the different banners?
The Solution: Find something unique among the three known versions.
My Implementation: Look at the dot codes in columns. The fourth column is different in all three J-Banners.
The Problem: How to name the banners?
The Solution: Make the names indicative of each banner's unique quality.
My Implementation: Use the fourth column. Map the dots to bits. This will produce a unique number.
Here is what the fourth column looks like in all three banners if you substitute 1 for a dot and 0 for a blank:
J-Banner 7:
1
1
1
0
J-Banner 0:
0
0
0
0
J-Banner 15:
1
1
1
1
Now, convert these columns to rows:
J-Banner 7: 0111
J-Banner 0: 0000
J-Banner 15: 1111
We now have binary data. Last step- convert to decimal:
J-Banner 7: 0111 = 7
J-Banner 0: 0000 = 0
J-Banner 15: 1111 = 15
-
So the two J Banners in Squidhenge are identical. J-0's over the doors leading into the courtyard.
That squashes my idea of the banners being scan points for the Oliphaunts as they go around town collecting garbage. Maybe they signify the clearance codes for the different entry points the Jotun employees may use. At work I have to badge into a turnstyle to get in the door, maybe these codes indicate which employees can go where. Probly not. They are Asklon buildings and Jotun just builds the Oliphaunt, it doesn't run the trash services. Vergil would. Or maybe Asklon owns the building and leases offices to Jotun. Hmmm.
(Sigh) Maybe they just look cool...
-
Oh, thanks! :)
I actually always read the J Banners like the dots made images, like the V in the beginning of your attachment's code. (was that understandable? o.O)
-
Oh, thanks! :)
I actually always read the J Banners like the dots made image, like the V in the beginning of your attachment's code. (was that understandable? o.O)
Actually....It was.
-
Still been hammering away on this. Looking at the images I was put in mind of the quipu (http://en.wikipedia.org/wiki/Quipu) that use knots in string to signify meaning. Unfortunately nobody has ever really deciphered the quipu. but they most likely use a 7 bit arrangement for knots/no knots. That lead me down a trail to a place we've already been.
The next thing I am excited about is the ARK01 code that team vergil uses. The layout got me thinking, what if the j-banners work the same way? Each dot is a 1 and strings of them add up with the open space being the separators.
So J-15 would give us
1 2 1 2 4 1 1 5 2
1 2 2 2 2 1 1 2 1 1
1 1 1 4 2 2 4
7 1 1 2 2 2 1 2
J7 is:
1 2 1 2 4 1 1 6 1
1 5 2 2 2 3 2 1 1
1 1 1 4 2 2 1 2
3 1 1 2 1 2 1 2
And J0 would be
1 1 1 2 3 1 1 1 1 1
1 1 6 15 1
1 1 2 1 1 2 2 1 2
3 2 2 1 2 1 2
I'm going to be working on these numbers but I don't want to be the only one. Come on team. Do they map to letters? To a polybius square (bungie's would probably be 7 by 7). Let's knock the cobwebs off and get those labcoats back on!
-
I'm going to be working on these numbers but I don't want to be the only one. Come on team. Do they map to letters? To a polybius square (bungie's would probably be 7 by 7). Let's knock the cobwebs off and get those labcoats back on!
Yeah, that's like it!
C'mon people, we used to rock in this! We still can, but you just have to straighten up, get those brains running good and type! There's lots of doing here, just pick one case and start working!
-
Hmmmm...Why can't I find more than one of the Banners in our gallery, or anywhere in the forums? ???
-
Try here (http://halo.wikibruce.com/Image:ODST-J-Signs.jpg).
-
Try here (http://halo.wikibruce.com/Image:ODST-J-Signs.jpg).
Ah, yes. Thanx.
I have been wondering this: From the Genet Coup Register plate Egg, we know that The Halo 3: ODST team was co-operating with The Halo: Reach team. Therefore, could it be possible that there are more things that relate somehow to Reach?
Could this and/or some else of our mysteries have some kind of counterpart in Reach that helps us solve the mystery?
Speaking of Reach, perhaps we should get a Reach forum here...
-
I may be off a little... Thought it might be worth doing
-
The red is were the 2 over lay each other. There is only 2 J-banners...right?
-
The red is were the 2 over lay each other. There is only 2 J-banners...right?
No...? (http://halo.wikibruce.com/Image:ODST-J-Signs.jpg)
-
I am not sure where I am going with this....
-
Well, that last one could have something.
Could maybe perhaps possibly. ;)
-
May be nothing... buy does anyone have pictures of the scrolling names from Alpha site..(I think).. Had a thought that maybe using my templet to block out certain letters...just a thought. I am looking at all possibility.
-
Here you go! (http://www.gruntspajamas.com/forum/index.php?topic=492.msg3544#msg3544)
-
[EDIT BY CG: Not sure how this data was converted from the J-Banners into binary. Enigma?]
J banners converted to binary:
01010010110011101010100001010
01010111111011111111111111101
00100001011010100110011010110 A
00000011100001101100101101011
01011010110111101010111110110
01011011001100001101010110101
00001001010111100110011011110 B
01111111000101001101101101011
01011010110111101010111111010
01011111001101101101110110101
00101001011110000110011010110 C
00000011100101001100101101011
01010010110011101010100000010
01010011001000001101010010101
00000001010010000110011010110 A-B-C meet
00000011000001001100101101011
= Nothing
-
had a theory that the banners were sideways, but they are one row short (when turned). Also, Braille music use the 2x3 pattern.
-
Guys... Your forgetting a banner! In the stonehenge there are 2, and one is diffrent!
It only has 2 lines though...
like
++ + + + + + + + +++ + + + +
+ + ++ + + + + + + + + + + + +
Instead of
++++++ + + + ++ +++ + + +
++ + + + + + + + + + ++ +
++ + + + ++ + + + + ++ + +
Or inform me if you knew this? I will get a picture soon.
-
I thought there were only three. POIDO.
-
Will do... Jeez i need bungie pro.
EDit: 8:11 Pacific going now. Be back in 10.
-
Well i found out its one of the three. But looking at it while on the sad mono glyph pillar it is diffrent... which makes me think, maybe we should try our binary and such with it too?
Maybe thats why the engineers are so mad...
-
forgot to mention: See picture from my point of view.
Blind skull was on.
-
Now that picture looks like actual brail
-
Now that picture looks like actual brail
Then maybe that means that it is part of a sentence, and we need to take the bottom half of all the J banners to make a full sentence.
Im going to try to convert this to english. Be right back.
-
Well, anyway braille has 3 lines. So i tried this, and it looks like you can get many diffrent things. Its all about how you use the spaces and if you combine or not.
I got this,
a ji(n)ae(n)bji(n)he(n)d,
And you can see there are some random N's in there. This was because there are 4 identical letters or combination. Braille for E, is the same as en, and Braille for I, is the same as In. The First one i showed you has those N's, the second one does not.
a jiaebjihed,
Yes, there is a comma at the end. The dot was this: . On the ground, so i looked for it and it can be many things, including capitalization or something, its complicated, with our Braille.
Dehijbeaij a That is it backwards, if you think the last dot did mean capitalization for the last letter.
Im sure if you try it you can come up with different results by seeing it differently, so go try it!
-
How exactly did you get the banner to do that??
I put the blind skull on but all of the J banners are still the same.
-
You have to go to the stonehenge, get the plinth, and stand on the pillar with the sad mono glyph. Then look in the direction of the banner withought crouching, and you will see it.
Its like my attached picture. I went up there with you right?
-
It's just the balcony above it covering half of the J banner.
-
I know... the only reason i am intrerested in this is because when you are in the position the engineers start scatting. Maybe it mean we are about to find something? Why else would the engineers do this?
-
OK i thought you were trying to say the J banner just appeared like that with blind on,my mistake.
Good find.
-
Cleaned up the bad links.
My posting in the Ubiquitous Number thread got me thinking: we have three J banners, and three sections to the UN. Might that be used as some sort of 'key' to the J Banners? I need to do some number crunching later, but if anyone wants to jump on this, be my guest.
-
Major Necropost, but we're back on topic.
Based on the punch card idea from Sep7ember (http://www.gruntspajamas.com/forum/index.php?topic=728.msg11852#msg11852), I decided to make some decoding attempts assuming that there are two 6-bit numbers in each column, once you stack all three J-Banners on top of one another.
Before you invest any time reading the below wall of text, note that I have not yet found any obviously meaningful translations. Gherjszed again, basically. Nevertheless, I wanted to document the attempt. I still feel that considering the banners as punch card data is worthwhile, as there are numerous historical punch card encoding methods. I have by no means exhausted the possibilities here.
Anyhow, if you're willing to dive back into J-Banner decoding a little refresher may be helpful:
- There are three unique J-Banners.
- Each J-Banner has 28 columns and 4 rows, for a maximum possibility of 112 dots
- The three banners are easily identified by looking at the fourth column of dots. One banner as all four dots, another has three dots, and another has none at all
- Various "translations" of the banners have been attempted by assuming that each dot is a binary 1 and each space is a binary 0 (or vice versa). Thus, each banner could encode 112 bits of binary data.
- If you translate each "4 bit" column into a binary value, you will arrive at a number from 0-15. The three unique J-Banners were named based on this conversion of the fourth column; thus, J-Banner 0 has no dots, J-Banner 15 has all four dots, and J-Banner 7 has three dots. (this last one could also have been considered J-Banner 14, but "Seven is Darker" (http://bungie.wikia.com/wiki/Seven))
- 112 is equally divisible by 8 and 7. Thus, each banner could encode 14 traditional (8 bit) bytes, or 16 7-bit numbers.
- The ASCII character set is encoded using 7-bit numbers. However, since computers store data in multiples of 8 bits, a full 8-bit byte of data is used to store one ASCII character. Therefore, the "high order bit" or "most significant bit" will always be zero in a byte used to encode ASCII. The point here is a J-Banner translation into ASCII is viable using either 7-bit or 8-bit numbers.
WikiBruce has a nice page showing all three banners over here (http://halo.wikibruce.com/File:ODST-J-Signs.jpg).
I attempted a meaningful translation by the merging all three J-Banners into a single 28x12 matrix. I did this by "stacking" one J-Banner on top of another. With 12 dots per column, the matrix will resemble a portion of a Hollerith Card (http://en.wikipedia.org/wiki/Hollerith_card) (or simply "punched card") since they contained 12 possible punch positions per column. The main difference is that Hollerith Cards were most commonly 80 columns, not 28.
I will be referring to these 12-dot-per-column stacked J-Banners as simply "stacks". There are six possible permutations (I've transcribed all six at the bottom of this post for reference.) I named each stack based on the order of the J-Banners from top to bottom. e.g "Stack 00-07-15" is J-Banner 0 stacked on top J-Banner 7 which is stacked on top of J-Banner 15.
Any way you stack them, you end up with 28 columns of 12 bits. Each column of 12 bits can be converted (in various ways) into two characters by assuming the characters are encoded using a 6-bit character set.
I used the following 6-bit character encoding resource:
http://en.wikipedia.org/wiki/Six-bit_character_code (http://en.wikipedia.org/wiki/Six-bit_character_code)
I have applied all of the encoding methods listed in the above article to my data. The DECSIX format (because it maps to printable ASCII) provides the most verbose results- but it's all seemingly random.
What I think will be most useful to future efforts down this particular path is to provide the translation of each possible stack into its 28 columns of two 6-bit decimal values. Anyone can then take those values and map them against any 6-bit character encoding method they wish to test.
For each of the six stacks, the bits can be read big-endian or little-endian. Also, the bits can be flipped (dot=1 or dot=0). So there are four versions of each stack, making 24 possible tables.
I've provided all 24 tables below. In addition, I give the DECSIX decoding of each. I could provide the others (and perhaps I will at some point) but they are likewise nonsensical and in many cases do not map to printable characters.
Enjoy! [evil laugh]
Transcribed J-Banners
J-Banner 0 (Dot=1)
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
J-Banner 7 (Dot=1)
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
J-Banner 15 (Dot=1)
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
J-Banner 0 (Dot=0)
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
J-Banner 7 (Dot=0)
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
J-Banner 15 (Dot=0)
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
Transcribed Stacks
Stack 00-07-15
Stack 00-07-15 (Dot=1)
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
Stack 00-07-15 6-Bit Conversion: Dot=1, Little Endian
51 04 51 48 34 59 46 27 23 38 48 23 59 63 02 59 46 23 34 59 54 30 58 54 27 38 29 42
44 33 44 61 32 46 59 06 21 09 63 21 54 20 00 46 59 21 40 38 29 55 46 29 50 29 55 42
S$SPB[N;7FP7[_"[N7B[V>ZV;F=J
LAL]@N[&5)_5V4 N[5HF=WN=R=WJ
Stack 00-07-15 6-Bit Conversion: Dot=1, Big Endian
51 08 51 03 17 55 29 54 58 25 03 58 55 63 16 55 29 58 17 55 27 30 23 27 54 25 46 21
13 33 13 47 01 29 55 24 42 36 63 42 27 10 00 29 55 42 05 25 46 59 29 46 19 46 59 21
S(S#1W=VZ9#ZW_0W=Z1W;>7;V9N5
-A-O!=W8JD_J;* =WJ%9N[=N3N[5
Stack 00-07-15 (Dot=0)
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
Stack 00-07-15 6-Bit Conversion: Dot=0, Little Endian
12 59 12 15 29 04 17 36 40 25 15 40 04 00 61 04 17 40 29 04 09 33 05 09 36 25 34 21
19 30 19 02 31 17 04 57 42 54 00 42 09 43 63 17 04 42 23 25 34 08 17 34 13 34 08 21
,[,/=$1DH9/H$ ]$1H=$)A%)D9B5
3>3"?1$YJV J)K_1$J79B(1B-B(5
Stack 00-07-15 6-Bit Conversion: Dot=0, Big Endian
12 55 12 60 46 08 34 09 05 38 60 05 08 00 47 08 34 05 46 08 36 33 40 36 09 38 17 42
50 30 50 16 62 34 08 39 21 27 00 21 36 53 63 34 08 21 58 38 17 04 34 17 44 17 04 42
,W,\N(B)%F\%( O(B%N(DAHD)F1J
R>R0^B(G5; 5DU_B(5ZF1$B1L1$J
Stack 00-15-07
Stack 00-15-07 (Dot=1)
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
Stack 00-15-07 6-Bit Conversion: Dot=1, Little Endian
51 04 51 48 02 59 46 27 23 38 48 23 27 31 02 59 46 23 34 27 54 30 58 54 11 54 29 42
14 18 14 31 10 46 59 36 21 24 63 21 47 13 00 46 59 21 10 46 29 55 46 29 39 25 55 42
S$SP"[N;7FP7;?"[N7B;V>ZV+V=J
.2.?*N[D58_5O- N[5*N=WN=G9WJ
Stack 00-15-07 6-Bit Conversion: Dot=1, Big Endian
51 08 51 03 16 55 29 54 58 25 03 58 54 62 16 55 29 58 17 54 27 30 23 27 52 27 46 21
28 18 28 62 20 29 55 09 42 06 63 42 61 44 00 29 55 42 20 29 46 59 29 46 57 38 59 21
S(S#0W=VZ9#ZV^0W=Z1V;>7;T;N5
<2<^4=W)J&_J]L =WJ4=N[=NYF[5
Stack 00-15-07 (Dot=0)
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
Stack 00-15-07 6-Bit Conversion: Dot=0, Little Endian
12 59 12 15 61 04 17 36 40 25 15 40 36 32 61 04 17 40 29 36 09 33 05 09 52 09 34 21
49 45 49 32 53 17 04 27 42 39 00 42 16 50 63 17 04 42 53 17 34 08 17 34 24 38 08 21
,[,/]$1DH9/HD@]$1H=D)A%)T)B5
QMQ@U1$;JG J0R_1$JU1B(1B8F(5
Stack 00-15-07 6-Bit Conversion: Dot=0, Big Endian
12 55 12 60 47 08 34 09 05 38 60 05 09 01 47 08 34 05 46 09 36 33 40 36 11 36 17 42
35 45 35 01 43 34 08 54 21 57 00 21 02 19 63 34 08 21 43 34 17 04 34 17 06 25 04 42
,W,\O(B)%F\%)!O(B%N)DAHD+D1J
CMC!KB(V5Y 5"3_B(5KB1$B1&9$J
Stack 07-00-15
Stack 07-00-15 (Dot=1)
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
Stack 07-00-15 6-Bit Conversion: Dot=1, Little Endian
51 04 51 07 34 59 46 57 53 38 15 53 59 51 32 59 46 53 34 59 39 45 43 39 57 38 29 42
44 33 44 60 32 46 59 06 21 09 60 21 54 23 00 46 59 21 40 38 29 55 46 29 50 29 55 42
S$S'B[NYUF/U[S@[NUB[GMKGYF=J
LAL\@N[&5)\5V7 N[5HF=WN=R=WJ
Stack 07-00-15 6-Bit Conversion: Dot=1, Big Endian
51 08 51 56 17 55 29 39 43 25 60 43 55 51 01 55 29 43 17 55 57 45 53 57 39 25 46 21
13 33 13 15 01 29 55 24 42 36 15 42 27 58 00 29 55 42 05 25 46 59 29 46 19 46 59 21
S(SX1W=GK9\KWS!W=K1WYMUYG9N5
-A-/!=W8JD/J;Z =WJ%9N[=N3N[5
Stack 07-00-15 (Dot=0)
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
Stack 07-00-15 6-Bit Conversion: Dot=0, Little Endian
12 59 12 56 29 04 17 06 10 25 48 10 04 12 31 04 17 10 29 04 24 18 20 24 06 25 34 21
19 30 19 03 31 17 04 57 42 54 03 42 09 40 63 17 04 42 23 25 34 08 17 34 13 34 08 21
,[,X=$1&*9P*$,?$1*=$8248&9B5
3>3#?1$YJV#J)H_1$J79B(1B-B(5
Stack 07-00-15 6-Bit Conversion: Dot=0, Big Endian
12 55 12 07 46 08 34 24 20 38 03 20 08 12 62 08 34 20 46 08 06 18 10 06 24 38 17 42
50 30 50 48 62 34 08 39 21 27 48 21 36 05 63 34 08 21 58 38 17 04 34 17 44 17 04 42
,W,'N(B84F#4(,^(B4N(&2*&8F1J
R>RP^B(G5;P5D%_B(5ZF1$B1L1$J
Stack 07-15-00
Stack 07-15-00 (Dot=1)
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
Stack 07-15-00 6-Bit Conversion: Dot=1, Little Endian
51 04 51 55 02 59 46 25 21 38 63 21 27 19 00 59 46 21 34 27 55 29 59 55 09 54 29 42
14 18 14 03 10 46 59 44 29 24 03 29 47 61 08 46 59 29 10 46 25 59 42 25 47 25 55 42
S$SW"[N95F_5;3 [N5B;W=[W)V=J
.2.#*N[L=8#=O](N[=*N9[J9O9WJ
Stack 07-15-00 6-Bit Conversion: Dot=1, Big Endian
51 08 51 59 16 55 29 38 42 25 63 42 54 50 00 55 29 42 17 54 59 46 55 59 36 27 46 21
28 18 28 48 20 29 55 13 46 06 48 46 61 47 04 29 55 46 20 29 38 55 21 38 61 38 59 21
S(S[0W=FJ9_JVR W=J1V[NW[D;N5
<2<P4=W-N&PN]O$=WN4=FW5F]F[5
Stack 07-15-00 (Dot=0)
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
Stack 07-15-00 6-Bit Conversion: Dot=0, Little Endian
12 59 12 08 61 04 17 38 42 25 00 42 36 44 63 04 17 42 29 36 08 34 04 08 54 09 34 21
49 45 49 60 53 17 04 19 34 39 60 34 16 02 55 17 04 34 53 17 38 04 21 38 16 38 08 21
,[,(]$1FJ9 JDL_$1J=D(B$(V)B5
QMQ\U1$3BG\B0"W1$BU1F$5F0F(5
Stack 07-15-00 6-Bit Conversion: Dot=0, Big Endian
12 55 12 04 47 08 34 25 21 38 00 21 09 13 63 08 34 21 46 09 04 17 08 04 27 36 17 42
35 45 35 15 43 34 08 50 17 57 15 17 02 16 59 34 08 17 43 34 25 08 42 25 02 25 04 42
,W,$O(B95F 5)-_(B5N)$1($;D1J
CMC/KB(R1Y/1"0[B(1KB9(J9"9$J
Stack 15-00-07
Stack 15-00-07 (Dot=1)
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
Stack 15-00-07 6-Bit Conversion: Dot=1, Little Endian
59 08 59 15 40 59 46 49 53 34 15 53 61 53 32 59 46 53 42 57 39 45 43 39 60 39 29 42
12 17 12 28 08 46 59 38 21 25 60 21 46 15 00 46 59 21 08 46 29 55 46 29 38 25 55 42
[([/H[NQUB/U]U@[NUJYGMKG\G=J
,1,<(N[F59\5N/ N[5(N=WN=F9WJ
Stack 15-00-07 6-Bit Conversion: Dot=1, Big Endian
55 04 55 60 05 55 29 35 43 17 60 43 47 43 01 55 29 43 21 39 57 45 53 57 15 57 46 21
12 34 12 14 04 29 55 25 42 38 15 42 29 60 00 29 55 42 04 29 46 59 29 46 25 38 59 21
W$W\%W=CK1\KOK!W=K5GYMUY/YN5
,B,.$=W9JF/J=\ =WJ$=N[=N9F[5
Stack 15-00-07 (Dot=0)
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
Stack 15-00-07 6-Bit Conversion: Dot=0, Little Endian
04 55 04 48 23 04 17 14 10 29 48 10 02 10 31 04 17 10 21 06 24 18 20 24 03 24 34 21
51 46 51 35 55 17 04 25 42 38 03 42 17 48 63 17 04 42 55 17 34 08 17 34 25 38 08 21
$W$P7$1.*=P*"*?$1*5&8248#8B5
SNSCW1$9JF#J1P_1$JW1B(1B9F(5
Stack 15-00-07 6-Bit Conversion: Dot=0, Big Endian
08 59 08 03 58 08 34 28 20 46 03 20 16 20 62 08 34 20 42 24 06 18 10 06 48 06 17 42
51 29 51 49 59 34 08 38 21 25 48 21 34 03 63 34 08 21 59 34 17 04 34 17 38 25 04 42
([(#Z(B<4N#404^(B4J8&2*&P&1J
S=SQ[B(F59P5B#_B(5[B1$B1F9$J
Stack 15-07-00
Stack 15-07-00 (Dot=1)
1011010110111101010111110110
1011011001100001101010110101
0001001010111100110011011110
1111111000101001101101101011
1011010110111101010111111010
1011111001101101101110110101
0101001011110000110011010110
0000011100101001100101101011
1010010110011101010100001010
1010111111011111111111111101
0100001011010100110011010110
0000011100001101100101101011
Stack 15-07-00 6-Bit Conversion: Dot=1, Little Endian
59 08 59 63 40 59 46 17 21 34 63 21 61 53 00 59 46 21 42 57 55 29 59 55 28 39 29 42
12 17 12 01 08 46 59 46 29 25 03 29 46 60 08 46 59 29 08 46 25 59 42 25 46 25 55 42
[([_H[N15B_5]U [N5JYW=[W<G=J
,1,!(N[N=9#=N\(N[=(N9[J9N9WJ
Stack 15-07-00 6-Bit Conversion: Dot=1, Big Endian
55 04 55 63 05 55 29 34 42 17 63 42 47 43 00 55 29 42 21 39 59 46 55 59 14 57 46 21
12 34 12 32 04 29 55 29 46 38 48 46 29 15 04 29 55 46 04 29 38 55 21 38 29 38 59 21
W$W_%W=BJ1_JOK W=J5G[NW[.YN5
,B,@$=W=NFPN=/$=WN$=FW5F=F[5
Stack 15-07-00 (Dot=0)
0100101001000010101000001001
0100100110011110010101001010
1110110101000011001100100001
0000000111010110010010010100
0100101001000010101000000101
0100000110010010010001001010
1010110100001111001100101001
1111100011010110011010010100
0101101001100010101011110101
0101000000100000000000000010
1011110100101011001100101001
1111100011110010011010010100
Stack 15-07-00 6-Bit Conversion: Dot=0, Little Endian
04 55 04 00 23 04 17 46 42 29 00 42 02 10 63 04 17 42 21 06 08 34 04 08 35 24 34 21
51 46 51 62 55 17 04 17 34 38 60 34 17 03 55 17 04 34 55 17 38 04 21 38 17 38 08 21
$W$ 7$1NJ= J"*_$1J5&(B$(C8B5
SNS^W1$1BF\B1#W1$BW1F$5F1F(5
Stack 15-07-00 6-Bit Conversion: Dot=0, Big Endian
08 59 08 00 58 08 34 29 21 46 00 21 16 20 63 08 34 21 42 24 04 17 08 04 49 06 17 42
51 29 51 31 59 34 08 34 17 25 15 17 34 48 59 34 08 17 59 34 25 08 42 25 34 25 04 42
([( Z(B=5N 504_(B5J8$1($Q&1J
S=S?[B(B19/1BP[B(1[B9(J9B9$J
-
Aw man. AW MAN.
My brain hurts. This means we still have a multitude of ways to decode the J-Banners, and, gah, just shows once more how much can be done!
Before you invest any time reading the below wall of text, note that I have not yet found any obviously meaningful translations. Gherjszed again, basically. Nevertheless, I wanted to document the attempt. I still feel that considering the banners as punch card data is worthwhile, as there are numerous historical punch card encoding methods. I have by no means exhausted the possibilities here.
Are we talking about the different kind of punch cards listed on the wikipedia page, or just different encoding methods? I'd like to think that Bungie had some kind of logic in choosing the system for J-Banners. What could it be though? Do we have any methods that could be connected to Jotun, the Elephants or anything concerning ODST?
In any case, your work is awesome CG, and I thank you for it. Gjerzhed I'm not sure anymore what the right term is) or not, we might very well be one step closer to encoding them. :)
-
Well, it sure has been a while, hasn't it?
With the launch of Destiny 2's Solstice of heroes, I found this
(https://i.imgur.com/DINKjDt.jpg)
Look familiar to anyone? :)
-
I've been having a further look at this area, it's something that has really interested me being a Computer Science student.
It looks like some interesting avenues have been explored already.
- The punch card theory is interesting as punch cards have 12-rows and so do the 3 banners when stacked on top of eachother.
- 7-bit or 8-bit encoded binary values are an interesting theory as each banner of 112 values is divisible by 7 and 8. If this is the case, what techniques have been used on the text to "encrypt" it? Caesar Ciphers? XOR? Enigma Machine? There's no limit on the complexity that could have been used for the ciphers.
- Braille seems less and less plausible the more I look into it. Braille characters are typically 2x3, but the banners are each 4-high, making 2x4 braille characters. Some 8-bit Braille conversions exist online, but these appear to be very niche. One thing that might not have been explored is how braille is affected by differences in language, I can see that New Mombasa is supposed to be based in Kenya. Has anyone investigated how languages like Swahili affect Braille encoding? I see there have been attempts to standardise Braille characters, but there are still some individual language quirks. https://en.m.wikipedia.org/wiki/International_uniformity_of_braille_alphabets
There are useful tools that could simplify the process of checking out different encoding techniques to help crack this challenge that did not exist when this thread was first made.
One of the most versatile and easy to use is CyberChef. This is a handy code cracking tool created by the UK's GCHQ. You can access it at https://gchq.github.io/CyberChef/
Here are some pointers for getting started with CyberChef:
You will want to use the transcribed banners from ColdGlider a few posts back http://www.gruntspajamas.com/forum/index.php?topic=228.msg12195#msg12195 (http://www.gruntspajamas.com/forum/index.php?topic=228.msg12195#msg12195). You can arrange these in any way you see fit, then enter them into CyberChef.
The easiest way is to paste the banners into a text splitting tool like https://onlinetexttools.com/split-text (https://onlinetexttools.com/split-text), remove any line breaks so it's all on one line, and set the text split length (7 or 8 bits).
In CyberChef add the "From Binary" recipe to convert this input into text. After this a number of further "recipes" can be used to try to decode the message. An interesting technique is called "Magic", which tries a series of transformations and looks for how this affects the Entropy (a lower entropy score means it closer resembles natural language).
One disadvantage of CyberChef (which should be fixed in a future update with the addition of a "Byte Length" option) is that it expects 8-bit binary segments. This means if you have 6 or 7 bit splits you will need to pad them with zeroes at the beginning to make 8-bit segments.
Eg: 101010 111111 would need to be padded to 00101010 00111111 to make 8-bit segments.
I hope this makes sense and helps people get started into using different code cracking techniques. CyberChef has loads of techniques, the vast majority I have not yet explored. If you find anything worthwhile, you should be able to share a link which contains your input, "recipe", and output. This should make sharing any discoveries easier.
-
Look familiar to anyone? :)
God damn it. I knew this day would come. :D While for the longest time I had decided that it'd only be a decoration with a code-y look, there really is no reason for an artist to use some actual code generator for their props. This seems great! I wonder how much many who worked on ODST still are at Bungie?
I've been having a further look at this area, it's something that has really interested me being a Computer Science student.
Great stuff man! I also study cs - it could be an interesting project to write up something that could really crank through all those different comninations of transforms on this.
-
I made this to show that it really looks like something that’s been copy pasted 2 times and altered; they all share the same basic structure. This compounded by the fact that the Destiny J banner is 4x17 makes it pretty clear to me that it’s just aesthetic.
-
Hmm, I can't quite read that picture...? What do the colors correspond to here again?
And, yeah, might definitely be just aesthetics. Wonder what sort of design process brought it back to Destiny. Just a shoutout to a game that marketed Destiny before it was a thing?
-
Each J banner is colored unique, and then some of the areas that are consistent on all three are marked gold. For there to be this much replication in the banners means that either someone copy pasted and edited them to look different, or the data is 3 different things that all say almost the same thing (given that the structure does not vary by much at all). We’re talking letter differences by 1 and such. The former seems more reasonable at this point considering everything has been exhausted now. I think it honestly might just be a cool futuristic look, much like greebles on a spaceship.
-
Hmm, ok, you are right about the similarity (though it doesn't seem that all the gold dots are present in all of the banners?). Supposedly it could decode into something else than letters, but I agree that the simple explanation of "looks cool" is enough, though of course the world building angle of "This is something that the trash trucks [Elephants] could read" is also nice.
-
Hmm, ok, you are right about the similarity (though it doesn't seem that all the gold dots are present in all of the banners?). Supposedly it could decode into something else than letters, but I agree that the simple explanation of "looks cool" is enough, though of course the world building angle of "This is something that the trash trucks [Elephants] could read" is also nice.
Oops, I hope they do, I just pulled the banners from the discord :o. I think the Oliphant angle is bunk, as these doors are everywhere and anywhere, including balconies on Kikowani and inside Squidhenge. I believe it is just a technique to use the same assets on multiple doors and not have them all (appear to) say the same thing. For example, if the sign read "Cafeteria" instead, you'd think "Hmm. Why are there hundreds of cafeterias in Mombasa streets?". Instead, you see a complex barcode, and you assume them to be different, providing the illusion of many different labels throughout the city. Now, we of course just see 3 J banners, but the average player is going to see them as unique to the door they are over, which provides great immersion.
Now, could a dev use this opportunity to shove an egg in there? Sure, but at this point I doubt that's the case.
-
I think the Oliphant angle is bunk, as these doors are everywhere and anywhere, including balconies on Kikowani and inside Squidhenge.
I mean, yeah, why would exactly autonomous vehicles need to read barcodes anyway in the year 2552? : D Or any tech for that matter? In many ways the angle doesn't make sense, but then again it might not need to make sense to create an atmosphere of "these people have barcode on their walls 'cause they're so automated and sci-fi".
I believe it is just a technique to use the same assets on multiple doors and not have them all (appear to) say the same thing. For example, if the sign read "Cafeteria" instead, you'd think "Hmm. Why are there hundreds of cafeterias in Mombasa streets?". Instead, you see a complex barcode, and you assume them to be different, providing the illusion of many different labels throughout the city. Now, we of course just see 3 J banners, but the average player is going to see them as unique to the door they are over, which provides great immersion.
Now, could a dev use this opportunity to shove an egg in there? Sure, but at this point I doubt that's the case.
Oh, that's a great point!! When you make a texture that seems complex enough, even if they're mundane and used a lot people might not notice. Also, it again is creating atmosphere in the sense that something could be written there that has a function, and that something is going on in the city, instead of just using "blank_concrete_slab_19" -texture and making a bustling city's walls look empty. Nice.
-
I don't think I ever posted my thanks to Nightcrafter for bringing the J-Banner to our attention two years ago as it appears in Destiny 2- thank you, Nightcrafter! I'm so happy to see my old friends still working together with new members here at the SGP!
Well done keeping things going... I told you this place wasn't going anywhere. Looking at the news on our website, it appears that Destiny 1 is coming soon. [facepalm] I suppose those are examples of two interpretations of "not going anywhere"... but clearly it's all going somewhere if you are still here!
I had a reaction similar to Imppa's when I passed under the J Banner near the Drifter in Destiny 2... an unresolved puzzle from the past, rearing its bedotted face. Since Umbral Engrams were added in Season 11, a lot more in-game traffic is passing under that sign. Image attached.
Ark told me to tell you all that she sends her love.