Decoding the Station Transmissions

[John Stoutbeard]

The station transmission signals have been heard for many years. As far as I know we have not been able to decode these.

History​

Back in 2016 we did observe the signals. This was short time after v1.5 which made some changes to some of the other mysteries (Raxxla missions had just been removed, Early 2016). A twitch-recording from 2016, in v2.1.0.5 (Horizon). The recording is of bad quality, you're excused to ignore it. It does demonstrate the principle.

Spoiler: old sample

Sample of raw Audacity Spectrogram

What is this project?​

It is an attempt to make sense of what the Transmissions are, if they are saying anything useful or interesting.
The transmissions coming from the towers and the specific symbols being used is the scope for this project.

Goals:​

- Map stations with signals, without signals​

- Map relation to state (either for faction or system)​

- Detect repetition in the transmissions​

- Detect enough symbols​

​

Stretch goals:​

- Transcribe messages if possible.​

- Decode messages if possible.​

What are "The Transmissions"?​

We don't know. They can be recorded and made visual using Audacity a Windows app that can record, edit and show sound patterns.
Other examples are the FD-easter egg, the Unknown Probe (Thargoid Probe), and many more examples.

Where is it?​

In many stations having a Transmission Tower inside. In the past we did find some stations not doing transmissions.
This is part of what we're trying to map.

• Stations without transmission towers may not produce the signals.
• Outposts have no towers, no signals detected.
• 20-80% Coriolis, Orbis, Ocellus may produce signals. (Guesswork based on 2016 experience)

How can you record, hear and view this?​

You can follow these steps (work in progress):

Spoiler: Full instructions in spoiler

Audacity: Download link
Windows Audio: Use Headphones if you desire to listen into it. External (gaming) speakers may not be suited for listening purposes.
Elite:

• In ED Audio settings:, set "Audio mode" to "Night mode". Disable & Mute things like Music and unwanted sounds.
• In game, find a landing PAD near the back of the station. Use the external free-cam and move the view closer to the transmission tower
• TIP: have the tower (sound source) coming left or right side, this way the remaining noise from the rest of the station is played in the other channel.

In Audacity, before you record:

• Set the Waveform settings to something like this:
  
  
• Select the source where the game-sounds are coming from.
• Start recording
• Switch to Spectrogram display.

How to adjust Audacity​

Key points:

• Use "non-linear" scale, the linear gives distorted images due to how it's represented. I prefer "Mel" scale, but you may use "logarithmic" or "Bark"
• The Very High-range mirrors the Hi/medium range, but we can filter this out later.
• The low-frequency range is hard to visualize due to the nature of the sounds.
• Window Size and Window type matter. 2048 Window Size, play around with Window type.
• Play with Gain, Range and High-boost for best effects. See video.

Source: https://www.twitch.tv/videos/1413230442

How to position yourself in ED​

Source: https://www.twitch.tv/videos/1413209838
Note: The video is made darker just so the Text can be viewed. The colors ingame is of no importance, the position to the relay tower is.

---

Old symbol collections​

This is a table of symbols collected back in 2016, I'll try to categorize these again using new recordings in 2022.
Note: To get this kind of SHAPE you must use filtering for High-frequencies, the correct Scale in the Spectrogram settings. (Please see instructions above).

On trying to build a alphanumerical table:

• It is hard to decide which symbol represented which alphanumerical, some "i", "t", "L", "J" appears similar. Same problem for "o", "0 (zero)" and "Q". Some of the symbols are just guesswork based on elimination. It is flawed for now.
• New map in progress in post#3

Baseline​

A 3 hour session with live 'output' from Audacity onscreen can be viewed here. The output is being used to map the symbols to characters further down in post#3.

• The recording started at ~18:04 ingame
• System, Station: Zeaex, Tryggvason Orbital (Boom state)
• Intentionally using the Left audio channel to record, leaving traffic more on the right audio channel

Spoiler: Video

Source: https://www.twitch.tv/videos/1410986356

 

[John Stoutbeard]

Symbol analysis​

The symbols, when viewed properly filtered, appear to be Alphanumerical characters in handwriting.

• Most appear to be Sans Serif.
• Most appear to be capital letters (block).

Interval between symbols: ~4.6875 secs

Spoiler: 16 symbols interval

Example block letters handwriting​

​

 

[John Stoutbeard]

Symbol to characters - Assertion by deduction​

1. Included the nearest timestamp from the 'baseline' recording, so it is possible to go back and recheck.
2. Multiple sample of each symbol, so you can detect each individual characteristics side by side.

WARNING: Tall page - lots of images inside the Spoilers.

Table 1 - Numbers​

Spoiler: Numbers Table

Symbols​	Notes​	Possible classification​
	One. A slight arc on the left side of the number	1
	Two. By deduction, there are nothing else that can be 2.	2
		3
		4
		5
		6
		7
		8
	Nine. Some could argue it's a lowercase 'g', but most letters are uppercase/block letters	9
separate discussion #4	separate discussion #4 Zero, must decide between the correct 'o', zero and 'Q'	0

Table 2 - Letters deduction​

There are certain problems to identity E,F,G,I,J,L,O,Q. Separate discussions for these below.

Spoiler: Letters

Symbols​	Notes​	Possible classification​
	More likely to be A, notice img2 and img4 which seems to have a horizontal bar	a
	Most likely 'B', due to 'R' is ok. Less likely 'R' as when you see the leg of the R, it cannot be 'B'.	b
	One of these can be 'C' (appears as sunken), while the top one is a 'G'?	c
	Most likely D, less likely to be A	d
separate discussion #1		e
separate discussion #1		f
separate discussion #2		g
	Most likely H by deduction. The only other candiate is 'U', but it has more curve at the bottom. This one does not curve at the bottom.	h
	Confident this is 'i', the bottom line is straight enough for it NOT to be 'J' Signature: Upwards slant at the top bar. Different from 'T' (down) which can be mistaken in some cases.	i
separate discussion #3	separate discussion #3	j
		k
separate discussion #3	separate discussion #3	l
		m
		n
	separate discussion #4	o
	Must be 'P' not 'R', no leg. Signature: Much wider than 'R'	p
	separate discussion #4	q
	Can be mistaken for 'P', but img1 and img4 shows the outgoing leg. Signature: Much narrower and 'pointy' than 'P'	r
		s
	Separate discussion #3, but it most likely to be 'T'.	t
	Most likely 'U'. Could be 'H', but missing horizontal bar while it curves at the bottom. The 'H' is a more likely candidate for 'H'.	u
	By deduction: 'V'. See 'W'. Signature: Notice the 3-smudges and 'person in the background'.	v
	See the 'inner pyramid'. Signature: Has 4-smudges, different from the 3 smudges of 'V'.	w
	Strange one, but this is not 'A'. It could be 'V', but our current 'V' makes more sense. There is a background 'arc' behind top-half of a 'X'.	x
	Not 'X'. Is 'Y'	y
	No other character have this angle. Could be '2', but the current '2' we have is so much more distinctive. Must be 'Z'.	z

Discussion #1 - E,F problem​

J could also be part of this discussion

Spoiler: Comparison

Symbols​	Notes​	Possible classification​
	E or F. Signature: Top line slants down.	E,F
	There is a 'smudge' which could be a line. But the top line could also be a 'semi serif' J if someone wanted to make J stand out a bit.	J,E,F
	E or F. Signature: Top line rises and there's a 'tiny hole' at the start of the top horizontal line.	E,F

Discussion #2 - G or C?​

Spoiler: Comparison

Symbols​	Notes​	Possible classification​
	Could this be G? Unable to see any inside 'curl' ending of the G. img1 MAY have a horizontal line inside the curve. Signature: Looks like a C with a ghosted-V inside.	G or C
	The symbol 'sits lower' than the above one, and it's one of the last candidates for 'C'. Signature: Sits very low compared to the other 'C'.	C or G

Discussion #3 - J or L or something else​

Spoiler: Comparison

Symbols​	Notes​	Possible classification​
	Cannot be '1' as it is found. Cannot be 'i' as it is found. Can be 'J'. Can be 'L'.	J or L
	It cannot be 'I', see 'i'. Can be 'J'. Could be 'L'. Most likely 'T' due to the top.	T or L or J

Discussion #4 - The o,0,Q problem​

Spoiler: Comparison

Symbols​	Notes​	Possible classification​
	Signature: Has a smaller 'ghost o' in the background.	o,0,Q
	Signature: Emphasize top left corner is sharper.	o,0,Q
	Signature: There is a dash both in top right corner which could mean 'Ø' as in Zero. The dash at the bottom could also indicate it is a 'Q'	o,0,Q

 

[John Stoutbeard]

A slight apology, I keep expanding to more posts so I can edit+reference more easily in the future

Tests and comparisons​

• Multiple tests at same location (system and station). Any repeating patterns?
• Tests from specific places such as Sol, Shinrarta Dezhra.

Comparison test from baseline​

Test1: See if any patterns from 'day3' (2 days after baseline) with a 15 min recording appeared in the baseline.​

• Chose 5 easily recognizable (starting with a T or I or J) symbols to find the same set in the baseline.
• Pattern not found

Test2: See if any patterns in baseline repeated itself.​

• Selected 5 easily recognizable symbols to find the same set in the baseline.
• Pattern not found
• Repeated with another set of 5 symbols. Still not found.

Temporary conclusion​

1. In a 3 hour recording there was no obvious repeating pattern.
2. In a 15 min recording, 2 days later, same place (at the same timeframe ~18:05), there was no identical pattern in them.

 

[John Stoutbeard]

There is now a almost 15 hours stream from a new station, smallest system we could find which have a supporting station.
Note: The cam drifts ~1hour, but I was not able to attend the cam every hour, sometimes we have larger drift away from the antenna.

System: Teshub​

Station: Brereton Orbital​

Allegiance: Independent​

GameTime: 8.March3308, ~16:45​

State: None​

@Vildu has been kind enough to transcribe a 3 hours session.
No repeating sequences or repeating (double) characters as far as we have detected.
This unique signature may be very important to understand what we're dealing with.

One automatic cryptographic analysis service assumed it was Pollux-Cipher.
Another service drew a lottery between 'unknown' and 'base64'. It is definitively not base64.

 

[Vildu]

Transcriptions

Spoiler: 15-hour stream from Cmdr John Stoutbeard (above)

System: Teshub
Station: Brereton Orbital
Allegiance: Independent
Time: ~16:45, 8. March 3308
State: None
(link)

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

EDIT 01:19, 14. March 3308: Sorry! There was an error in the concatenation of my transcriptions, so a repeated segment of length 4001 was in the data. This has now been fixed along with the analysis below. Happy Pi day!
EDIT 11:45, 14. March 3308: The transcription of the symbol L was double checked from index 6077 and forward (only one mistake). Also an accidental double letter 'aa' was checked and fixed.

Spoiler: 3-hour stream from Cmdr John Stoutbeard

System: Zeaex
Station: Tryggvason Orbital
Time: ~18:10, 28. February 3308
State: Boom
(Link)

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

Both transcriptions were done by hand, and the stream of chirps continued without spacing differences or new characters. While transcribing I got a suspicion that a character is more likely to repeat shortly after it is observed in the stream, but such a pattern is pretty much the only one a human would think they saw writing down apparent randomness.

Preliminary analysis

To establish candidates for any possible meaning in the data, I have so far done some very basic tests for randomness. The normalized coincidence indices for the samples are very much like random: 0.999 for the 15-hour and 1.005 for the 3-hour sample.

Spoiler: 1. Statistical hypothesis testing

Hypothesis 1: Data is uniformly distributed.
The two images below show frequency analysis for both samples with corresponding 95% Wilson score confidence limits (red and blue lines) with continuity correction for proportions with p = 1/36. Wilson score is used because it is more accurate than normal approximation for extreme probabilities. What to expect: About 95% of character frequencies should be between the two lines.

Results: The hypothesis seems to hold for the samples. We expect to see on average about 0.05·36 = 1.8 characters outside the intervals, and while the count is slightly more than that for the shorter sample, the observation is not significant. However the symbol for 'l' Is not only outside the confidence interval for both samples, it also has the highest frequency.

Hypothesis 2: The two samples have the same distribution.
This is tested by comparing the difference of the proportions of each character in the two samples. Significance is measured by 95% Newcombe score confidence intervals with continuity correction for a difference of proportions. Under the hood, this method uses the Wilson score with continuity correction. The image below shows difference of proportions for each symbol, marked with an x, and the associated confidence interval as a vertical line. What to expect: The value 0 should be within the interval for about 95% of symbols.

Results: The hypothesis holds, and the two samples from two different stations have at least a very similar distribution. Only the symbol 't' doesn't have the value 0 within the confidence intervals. This is what we expect for 95% confidence.

Spoiler: 2. Bigrams with sliding window

At this point, it is clear that double letters are missing from the text, as the probability of seeing none at a chance in both of the two samples of 11395 and 2293 symbols would be less than 10⁻¹⁶⁷.

It does raise the question if some other bigrams would be missing, too. This does not seem to be the case, since no bigrams (other than doubles) are missing from the 15-hour transcription.

The sorted frequency analysis of bigrams yields a curve similar to random data with no double letters, although the maximum is somewhat lower.

The frequencies in the images are calculated by dividing the counts by the number of all bigrams N-1.

So far, the results show that the symbols are fairly random. The absence of a character repeating itself is clear, but it was a surprise that the letter 'L' seems to be more frequent than others. It is possible that the transcriptions have errors, with many characters being misclassified as an L (to be verified). The above analyses rule out substitution ciphers and some polygrammic ciphers. Also no significant difference in the character frequencies of the texts from the two stations was found.

I believe further analysis is needed to rule out more encryption schemes. I've been thinking of doing Kasiski's examination or coincidence counting on it. Also if anyone knows more distinguishers (randomness tests) that we could run on this amount of observations with some confidence, please let me know or try them yourself.

Testing of cipher candidates

The Dcode cipher identifier suggests to investigate the Pollux cipher when given any of the samples (or a randomly generated sample with repeat exclusion). This cipher does indeed produce a nearly uniform character distribution if the key is selected with proper proportions that have to be tailored for a given message. The Pollux cipher works by first encoding a message with Morse code having a third symbol to represent pause between characters and words, and then encrypting that by representing each Morse symbol with a character selected at random from a corresponding set. For plaintext English, the sets will be with 36 characters approximately of size 14, 10, 12, for dits, dahs and silence, correspondingly.

Spoiler: 1. Pollux on a repeated message

In this method we assume that the ciphertext is the result of encrypting the same message over and over and concatenating the resulting ciphertexts. Then no matter at which part of the ciphertext stream you start accumulating characters in message length sized blocks, the characters at the same indices in different blocks will be different representations of the same Morse symbol. Moreover, if these blocks are arranged by rows into a matrix, each column will have a subset of characters from the same set. Thus, it is possible to take unions of overlapping sets to form the hopefully entire sets in the key. If the block size is not the message length or there was an error in the transcription, it is very probable that the final accumulated list of sets will contain only one set.

Results: No key candidates can be found in parts of the transcriptions. Transcriptions were analyzed in sections to decrease probability of error. If there is a repeating message, it maybe is longer than 4000 characters. The eventual coincidence analysis will collaborate this result.

Spoiler: 2. Pollux spaces set deduction

Here we abuse the fact that the Morse code encoding can never have three or more pause characters one after another, since one pause is used to separate characters and two pauses are used to separate words. Thus we will attempt to find the maximal set of characters such that this constraint still holds. While the number of combinations even in this reduced case would be still quite large, the search can be done efficiently by breadth-first search of the combinations (Use numbers for the characters. Store the set with the number it was last expanded with. When this pair is dequeued, expand the set with only values larger than the stored number! Would there be a similar method to brute all three sets?).

Results: Inconclusive so far. You get 5 to 7 character sets with 4000 character segments of text, which is similar to applying the method on random text. It could also be that most of the message consists of long Morse encodings and there are no word boundaries. I have not yet used the results to try to decode or use Morse encoding lengths to verify any sets.

Running key cipher was suggested by CMDR ScolioTheMost, and to me, an interlaced keystream+ciphertext stream could be possible. I have only tried key-ciphertext pairs, so other interlacings are still on the table.

I have so far found only one cipher that has no double letters, albeit within bigram blocks. This is the Playfair cipher, but apparently it should be identifiable by bigram frequency analysis. I suspect it could be modified to match our situation by using sliding window bigrams instead. The properties of such a cipher would have to be discovered, as the security of such a cipher is not obvious to me.

Do you know of any encryption schemes that have no double letters in the ciphertext? Try to crack the code yourself or let us know!

 

[StefanOS]

Impressive, didnt know this. 2016? 6 years now.... could this still be relevant to the story today? Or is it a cypher that was not picted up and therefore abandoned in the story?

 

[Ian Doncaster]

Very interesting!

It's surprising that it can go on for so long without any repeating sequences at all. Assuming there is some meaning to it, that would make it quite difficult to decode even knowing the scheme and key for a lot of cipher types, just because of the difficulty of identifying the message boundaries.

Question: do two people recording the same beacon at the same time get the same letters?

 

[electro]

The station transmission signals have been heard for many years. As far as I know we have not been able to decode these.

History​

Back in 2016 we did observe the signals. This was short time after v1.5 which made some changes to some of the other mysteries (Raxxla missions had just been removed, Early 2016). A twitch-recording from 2016, in v2.1.0.5 (Horizon). The recording is of bad quality, you're excused to ignore it. It does demonstrate the principle.

Spoiler: old sample

Sample of raw Audacity Spectrogram

What is this project?​

It is an attempt to make sense of what the Transmissions are, if they are saying anything useful or interesting.
The transmissions coming from the towers and the specific symbols being used is the scope for this project.

Goals:​

- Map stations with signals, without signals​

- Map relation to state (either for faction or system)​

- Detect repetition in the transmissions​

- Detect enough symbols​

​

Stretch goals:​

- Transcribe messages if possible.​

- Decode messages if possible.​

What are "The Transmissions"?​

We don't know. They can be recorded and made visual using Audacity a Windows app that can record, edit and show sound patterns.
Other examples are the FD-easter egg, the Unknown Probe (Thargoid Probe), and many more examples.

Where is it?​

In many stations having a Transmission Tower inside. In the past we did find some stations not doing transmissions.
This is part of what we're trying to map.

• Stations without transmission towers may not produce the signals.
• Outposts have no towers, no signals detected.
• 20-80% Coriolis, Orbis, Ocellus may produce signals. (Guesswork based on 2016 experience)

How can you record, hear and view this?​

You can follow these steps (work in progress):

Spoiler: Full instructions in spoiler

Audacity: Download link
Windows Audio: Use Headphones if you desire to listen into it. External (gaming) speakers may not be suited for listening purposes.
Elite:

• In ED Audio settings:, set "Audio mode" to "Night mode". Disable & Mute things like Music and unwanted sounds.
• In game, find a landing PAD near the back of the station. Use the external free-cam and move the view closer to the transmission tower
• TIP: have the tower (sound source) coming left or right side, this way the remaining noise from the rest of the station is played in the other channel.

In Audacity, before you record:

• Set the Waveform settings to something like this:
  
  
• Select the source where the game-sounds are coming from.
• Start recording
• Switch to Spectrogram display.

How to adjust Audacity​

Key points:

• Use "non-linear" scale, the linear gives distorted images due to how it's represented. I prefer "Mel" scale, but you may use "logarithmic" or "Bark"
• The Very High-range mirrors the Hi/medium range, but we can filter this out later.
• The low-frequency range is hard to visualize due to the nature of the sounds.
• Window Size and Window type matter. 2048 Window Size, play around with Window type.
• Play with Gain, Range and High-boost for best effects. See video.

Source: https://www.twitch.tv/videos/1413230442

How to position yourself in ED​

Source: https://www.twitch.tv/videos/1413209838
Note: The video is made darker just so the Text can be viewed. The colors ingame is of no importance, the position to the relay tower is.

---

Old symbol collections​

This is a table of symbols collected back in 2016, I'll try to categorize these again using new recordings in 2022.
Note: To get this kind of SHAPE you must use filtering for High-frequencies, the correct Scale in the Spectrogram settings. (Please see instructions above).

On trying to build a alphanumerical table:

• It is hard to decide which symbol represented which alphanumerical, some "i", "t", "L", "J" appears similar. Same problem for "o", "0 (zero)" and "Q". Some of the symbols are just guesswork based on elimination. It is flawed for now.
• New map in progress in post#3

Baseline​

A 3 hour session with live 'output' from Audacity onscreen can be viewed here. The output is being used to map the symbols to characters further down in post#3.

• The recording started at ~18:04 ingame
• System, Station: Zeaex, Tryggvason Orbital (Boom state)
• Intentionally using the Left audio channel to record, leaving traffic more on the right audio channel

Spoiler: Video

Source: https://www.twitch.tv/videos/1410986356

What is the result if you align all the pictures on each other?

 

[SongOfTheVoid]

Good writeup and a well presented series of evidence with explanations of your current understanding.
Something that may help understand the purpose of what is being broadcast and the functions of these signals is to look for outlying oddities, the various mining outposts scattered in remote nebula and places across the galaxy may prove informative if not help you at least rule out or rule in such aspects as: 1) Difference's in repetition time period in relation to system distance to X locations within the primary bubble, 2) Changes in the order of symbology repeating cycles.

 

[Crimson Echo]

It seems odd that cannon has not managed to crack this in the last 6 years considering most puzzle don't last the weekend. Before the vanity camera was added how would players have been able to pull in such a seemingly long duration signal sequence?

When relogging in a station in a large ship the game seems to always spawn the player at the rear pad near the tower.

DB mentioned in an old video a fondness for surround sound/ directional audio. Perhaps all 3 of the glowing towers are emitting part of the audio in a directional manner that leads to the prescribed audio recording approach missing out on some of the audio. Even stations with out the large tower seem to have 3 shorter towers that glow similarly.

Has is it been determined if the color (red/blue) of the tower's glow changes the audio in any way?

 

[Leeloo Catfish]

Nice breakdown John ! Glad to see more interest in this phenomenon. Here is my own attempt at cleaning up the characters. I personally think that they function exactly like a 'Number Station' — constantly broadcasting random characters, never a repeat and waiting to either deliver a message amongst the chaos or potentially at a specific time convert the stream using a one time pad. Ive started looking at systems where key events are taking place. Recently in Novas I took a recording and it was like it was offline seemed kinda strange but proves nothing.
Ive tried all kinds of other tricks like byte swap and any cipher I could pump that data into, bupkis . My current plan of attack is to find a way to auto transcribe the audio and then go full dragnet on everything of interest 24-7 until something gives. Time is limiting so who knows when that will ever work, current version is flaky at best.

Novas signal offline

 

[D7]

Nice breakdown John ! Glad to see more interest in this phenomenon. Here is my own attempt at cleaning up the characters. I personally think that they function exactly like a 'Number Station' — constantly broadcasting random characters, never a repeat and waiting to either deliver a message amongst the chaos or potentially at a specific time convert the stream using a one time pad. Ive started looking at systems where key events are taking place. Recently in Novas I took a recording and it was like it was offline seemed kinda strange but proves nothing.
Ive tried all kinds of other tricks like byte swap and any cipher I could pump that data into, bupkis . My current plan of attack is to find a way to auto transcribe the audio and then go full dragnet on everything of interest 24-7 until something gives. Time is limiting so who knows when that will ever work, current version is flaky at best.

View attachment 307150
Novas signal offline
View attachment 307151

"waiting to either deliver a message amongst the chaos or potentially at a specific time convert the stream using a one time pad"

Yes, or perhaps the stations are all random except the one(s) that are supposed to be found with a message straight up. I could be wrong but, I am not convinced that ED really wants to require that a commander has to be a cipher expert to solve this. In the past I have tried basic cipher systems for ordinary folks and even more sophisticated ones with no results. There is an outside chance the right key for a keyed cipher is needed but, this type a coding seems either much too sophisticated for that or like I said, is actually pure random generated EXCEPT for certain stations of interest. A keyed cipher would be odd because every station has a different code. Meaning a key would unlock thousand of messages? I doubt that.

 

[Wily Rites]

I love this topic, and really want there to be an easter egg behind it, but the thing that causes me doubts is the absolute genius of the in game sound design. As such I'm kind of left with the hunch that we, rather you, have stumbled upon one of their secret techniques for generating such original and enthralling sounds; I genuinely hope that you prove me wrong though.

Great work though, I love it, keep it up!

Does anyone else here get the chills when there is a Roy Batty type howl when landing at planetary bases, such strong Bladerunner vibe in those moments, pure genius!

 

[Beng_Ammon]

Does the station Galileo at earth's moon have one?

I would suggest decoding it, and then decoding it again while earth is experiencing a full moon.

 

[selbie]

I decided to make a highlighted version of the character table produced by ScolioTheMost. Hopefully this helps with further recordings.

I also made an attempt at finding "companion" characters which seem to be merged with certain ones. There is potential here for substitutions or key in a cipher.
I have paid careful attention to isolating the parts that are distinctly different from the primary character shape.

Cross-reference compilation post in Quest for Raxxla

 

[SHEA PARDH]

Folks, those signals also exist in eligible stations that don't have the antennas (the sound source appears to be still there in the "empty" space where the antenna normally would be) - don't go to the back of the station - record where you'd expect the end of the antenna to be.

They also exist in the small outposts at the end of the long poles with blinking large red lights.