Blocking Steganosonic Data In Phone Calls

psyced writes "Steganography is a technique to encode secret messages in the background noise of an audio recording or photograph. There have been attempts at steganalysis in the past, but scientists at FH St. Pölten are developing strategies to block out secret data in VoIP and even GSM phone calls by preemptively modifying background noise (link is to a Google translation of the German original) on a level that stays inaudible or invisible, yet destroys any message encoded within. I wonder if this method could be applied to hiding messages in executables, too."

Not going to work....

[dgatwood (11270)]

That's completely pointless. All it does is create an arms race. Any amount of noise you add can simply be dealt with by including the stego data more than once or using checksums or whatever. Any amount of damage sufficient to prevent any possibility of hidden messages would result in significant audible alteration of the sound to the point of unusability....

Re:Not going to work....

[Brian Gordon (987471)]

Obviously if they modify the background noise then no amount of ECCs could recover anything from it since they're modifying all of the ECCs too.. unless you knew exactly what shifting frequencies they were using or something, but that's just reversing the damage, not working through it.

Re:Not going to work....

[Jah-Wren Ryel (80510)]

Obviously if they modify the background noise then no amount of ECCs could recover anything from it since they're modifying all of the ECCs too

Who says that the people with secrets will even try to encode them in the background noise?

Maybe they will use the foreground noise -- for example, they could alter the pitch of the speaking voice to precisely fall into certain discrete frequency ranges, and then they occasionally bump a couple of samples into an 'unused' range and use those as a simple binary encoding of the secret data.

If they use enough discrete frequency ranges, the general tone of the speaker's vioce won't be noticeably different and the occasional minor shifts in frequency for the encoded data will hardly stand out.

That is just one example that I literally thought up in 30 seconds. I'm sure someone who was really concentrating could come up with much better ways to defeat the described countermeasures.

Re:Not going to work....

[badfish99 (826052)]

More likely, the people with secrets would just use some other method to communicate them.

Given that this project is (according to TFA) partnered by the Ministry of Defence, this smells to me like someone spending a lot of money defending against a non-existent threat. What's the betting they used the magic word "terrorism" in their grant application?

Re:Not going to work....

[ZeroExistenZ (721849)]

this smells to me like someone spending a lot of money defending against a non-existent threat

It's against the people itself. It's propaganda to keep the "terror" alive in memory, generating visions of terrorist so advanced we have to process and inspect all telecommunication, so you can feel safe.

Please, have a look at this documentary: The century of the self [bbc.co.uk].

Re:

[fizze (610734)]

Just two facts, that noone has seemed to mention here:
a.) The project is a feasibility evaluation, and as such doesn't have to produce results.
b.) The Austrian Ministry of Defence is supporting this project.

This isn't even remotely like DARPA, so chill out ;-)

Re:

[severoon (536737)]

Just as a degenerate example of a little thought experiment...what if I use PKE to encrypt a message to the person on the other end of the line and then write it out in hex? The conversation would go something like: "A! F! 3! 8! 8! 4! 9! BEEEEE!!!"

This is, of course, the most trivially stupid possible way to do this. Much better would be to write a small program that translates the encrypted message to, say, base-256 and then bleeps short tones in one of 256 different frequencies to the receiver program,

Re:

[Lumpy (12016)]

Why waste the time. hook the cellphone to a PC, take a photo with the camera, load photo the pc, add your stenao message and then send it as a sms to the intended recipient.

Far far easier than trying to secretly encode a message in the background of my audio phone call, and no special gear needed.

Wow are the "spies" of the world getting incredibly lazy? I can come up with at least 30 ways to get around this, one of which is having several prepay disposable cellphones to get around them even tapping my pho

Re:

[QRDeNameland (873957)]

Who says that the people with secrets will even try to encode them in the background noise?

Maybe they will use the foreground noise --

I couldn't read TFA as Google translation was hung, but I question the summary's definition of steganography as hiding data in the "background noise".

If you read wikipedia's steganography entry [wikipedia.org], you'll see no mention of background/foreground noise in the definition. My understanding is that steganography generally alters the lowest order bits in a audio/video/image fil

Re:Not going to work....

[Zemran (3101)]

would result in significant audible alteration of the sound to the point of unusability....

Sounds like an average mobile phone call to me...

Re:Not going to work....

[StuckInSyrup (745480)]

(since the signals used to transmit the secret message are the same as the ones transmitting the public message, and they do not have permission for destroying the public message)

Did you just call a phone call a "public message"? Man, you are even more cynic about privacy than I am.

Re:Not going to work....

[jd (1658)]

You're probably right. Block-length FEC and Turbo Codes allow you to fix errors assuming bursty data corruption of exactly this kind, which is why NASA uses them for deep space missions. You can't exactly ask a probe on the edge of the solar system or skimming geysers to repeat itself. With sound, there's also the fact that you've multiple parameters - delay, amplitude and frequency. Unless they plan to randomize all three, you can use any of the others for covert data. Data compression isolates anything either side, so whatever they are "protecting" is limited to that one side. Shouldn't be hard to use the other.

Re:

[ColoradoAuthor (682295)]

The jamming will also easily be defeated by an entirely new branch of coding theory using the BBC algorithm ( http://crisp.cs.du.edu/frisc/baird.pdf [du.edu]). Error correction is distributed throughout the data stream, so even if the jammer completely obliterates parts of the signal--to the point that the original signal is unintelligible--the coded message will still get through.

This coding theory is handy for all sorts of stuff, from military comms to cell phones to MIMO access points. And unlike most crypto

Re:

[The Clockwork Troll (655321)]

Any amount of noise you add can simply be dealt with by including the stego data more than once or using checksums or whatever

Yes, but how to do this in real-time in a cryptographically secure manner is the subject of much ongoing research.

The feeling in the research community at the moment is that efficient stego-redundancy requires a working database of discovered steganographic synonyms, i.e. a stegosaurus [wikipedia.org].

Re:

[CastrTroy (595695)]

On top of this, if you have a VOIP/GSM phone, you probably have email. Why not just send encrypted email? Why jump through hoops trying to send stenographic data through the phone system.

Re:Not going to work....

[cnettel (836611)]

On top of this, if you have a VOIP/GSM phone, you probably have email. Why not just send encrypted email? Why jump through hoops trying to send stenographic data through the phone system.

(More) deniability.

Re:

[Ortega-Starfire (930563)]

If you want that, just post a one time pad code on a popular public website. I mean, that way people could post links to instructional manuals for covert materials creation for example and not get caught. Try to imagine the manpower involved to go through each lead.

I doubt the CIA will investigate every no carrier joke on slashdot, and if they di^H^H^H^H^H^ 01101000 01110100 01110100 01110000 00111010 00101111 00101111 01110111 01110111 01110111 00101110 01111001 01101111 01110101 01110100 01110101 01100010

Re:Not going to work....

[gstoddart (321705)]

On top of this, if you have a VOIP/GSM phone, you probably have email. Why not just send encrypted email? Why jump through hoops trying to send stenographic data through the phone system.

Because, they can tell when you send an encrypted e-mail.

The whole point of steganography is to embed the secret message in something you broadcast in the clear, and have nobody be any the wiser that you are, in fact, sending hidden data. You give up your covertness when you observably send something secret. If nobody knows you sent it, they're not looking for it. They just think you were talking about your aunt's petunias.

Think of it as analogous to fieldcraft for spies -- you're supposed to be able to do something completely innocuous so that they can't ever confirm that you've actually done something nefarious.

This system is trying to preemptively just eliminate the ability to send something embedded in a clear-channel communication. Basically, take away your ability to send an encrypted sub-channel in your normal conversation.

Cheers

Re:

[CastrTroy (595695)]

But by completely removing the ability for them to transmit the data, they've also lost the ability to catch people who want to transmit data this way. If you know how to break their codes, don't tell them, because they will find some other way of transmitting the data more securely.

Re:Not going to work....

[SQLGuru (980662)]

Unless it's a foot race.

Layne

Subliminal white-noise?

[Zymergy (803632)]

Could this just be subliminal white noise? (as opposed to superliminal).

I guess its one way to prevent getting the alien infection from over the phone (anyone remember Threshold)... might mitigate some people's fears of harmful sensation. http://en.wikipedia.org/wiki/Motif_of_harmful_sensation [wikipedia.org]

I wonder if it will foil over the phone lie-detectors like this one: http://www.liarcard.com/ [liarcard.com] ?

Not a secret message.

[Creepy Crawler (680178)]

The butterfly flaps its wings twice.

I repeat, the butterfly flaps its wings twice.

Re:

[Thanshin (1188877)]

The butterfly flaps its wings twice.

Oh dear God no! Quick, everybody to the shelter. Micky, take the tinfoil, Becky, the red ink. John, the condenser and the racket.

Just hope we're not too late.

Re:

[CmdrGravy (645153)]

Are you Mr Xaviar, Mr Francis Xaviar of Mile End Road, London ?

Just need to check because that looked like one of the secret code words I am supposed to be watching out for, anyway if that is you Francis ( there is a code word for this somewhere but I think it's in the basement somewhere so I'll just give you the gist ) the "materials", you know what I mean eh - one ends the barrel and they're "easily triggered", anyway the "materials" will be loaded onto the Builders Merchants truck which will then be park

Re:

[Alsee (515537)]

After anti-steganographic transformation:

I saw a bug.

-

Re:Not a secret message.

[Chrisq (894406)]

The butterfly flaps its wings twice.

I repeat, the butterfly flaps its wings twice.

Please clarify immediately. Is that just a repetition or does the butterfly flap its wings four times. This could be the difference between a gang of naked teenagers invading Prime Minister's question time and the defacing of Nelson's column.

Re:

[baboonlogic (989195)]

What is this a reference to? Whatever it is, it doesn't appear to be that popular [google.co.in]. If this is a sci-fi movie quote... I wanna see the movie it's in...

I like parent's sig

[aepervius (535155)]

"though the parent's sig is annoying, hackneyed, stupid, redundant, and (did I already say this?) annoying."

I see the parents sig as a sort of darwinian filter on how careful one is the slashdot reader at clicking link.

Can I add random noise to a .exe file...?

[Joce640k (829181)]

Um, no.

Yes, you can, sort of.

[archeopterix (594938)]

You can add "random noise" to an .exe file - most processors have at least some opcodes with "don't care" bits. You can alter those bits without affecting the semantics of the code.

Or..

[Chrisq (894406)]

Or just jump over a group of random bytes that will never be executed. In a high level language have some unused variable

myString = "FooFoogh234h2j4hj23hj";

search the executable for FooFoo then read the following bytes.

Re:Or..

[kvezach (1199717)]

Or perturb the logic. The easy way is just to look at how polymorphic viruses did it. The hard way is to get out your disassembler and change

cmp eax, edx
jle offset
to
cmp edx, eax
jae offset
(insert your own variation here). Have a program read all cmp eax, edx (or cmp edx, eax) opcodes and output 0 for the first and 1 for the second.

Re:

[by Anonymous Coward]

Um, yes you can. Many instruction combinations are interchangeable. You merely need to be certain the result is same in all relevant cases for both instruction sequences. In the easy cases it might mean just to swap two instructions. See polymorphic viruses.

Additionally you can use empty areas in executable formats, in the headers or padding. Or even add an extra data segment... If file size is no issue, you can typically just concatenate some extra data in the end of file.

However, instruction sequenc

Re:

[utnapistim (931738)]

Of course you can. Just don't expect it to still execute. If we're talking here about steganography (stegano-[something else?]), you can still carry the file as an .exe file and go: "I don't know why it doesn't execute ... I guess it's corrupted!"

Re:

[hairyfeet (841228)]

Uh, they actually had an article on slashdot a few years back about a program that would let you hide stuff in executables. And they still worked fine. Here is the article [slashdot.org] and the link to where you can get the code still works.

I personally think this is just another government handout. There are so many much easier ways to hide a secret message than using a phone. Hell, they could just post one of those stupid lolcat pictures on the web with the message inside. The operative would only have to know somet

Re:Can I add random noise to a .exe file...?

[yoris (776276)]

Yes you can. Some examples: - replace "add 1024" with "substract -1024" - replace "if greater then 100" with "if greater then or equal to 99" - replace "copy a to b, copy c to d" by "copy c to d, copy a to b" Just have a look at any assembly language and use your imagination. To make matters even simpler, there are operators which completely ignore certain parameters (e.g. a JUMP operator which only takes 1 parameter leaves room for hidden data in the 2nd and 3rd operator field). There are plenty of instructions or combinations of instructions which leave room to such minor changes without any difference in execution. So for the steganographers, the goal would be to look for all of such instances in an executable, then agree on some kind of code (for example "add n" is a 1, "substract -n" is a 0). Semantically there is no difference, both codes will result in the exact same execution, but you found some wiggle room to leave a message. It was reported on Slashdot a few years ago.

As the tag says: encryption.

[Rah'Dick (976472)]

I wonder if we will ever have widespread end-to-end encryption for all of our private communication, so that "service providers" cannot mess with our actual message and/or data stream. I guess there will always be someone making a profit by preventing this on a legal level, sadly. When will the "mindless consumer" finally wake up and kick the government that allows all this?

Re:

[monsted (6709)]

You can use SRTP [networksorcery.com]. It's been available in many VoIP implementations for years. For lawful interception, the call controller (cisco call manager or such) usually holds the key to the stream, but if you're in control of both ends and the controller, you're safe.

Microsoft uses that.

[SharpFang (651121)]

I wonder if this method could be applied to hiding messages in executables, too.

Yes, a similar method has been employed by Microsoft to all the executables it ever released, ever since the times of MS-DOS.
After compilation they run the program through a special utility that modifies a few bits in the executable at random. Then they run the resulting executable through some tests and if it passes, they release it, if it crashes, they try with a different random bits.

Arrogant bastards!

[pla (258480)]

scientists at FH St. Polten are developing strategies to block out secret data in VoIP and even GSM phone calls by preemptively modifying background noise

...And once again, they treat all of us like criminals for the sake of annoying (not even preventing or catching) the 0.0001% that really pose a threat.

Good work, guys - Even a classic BOFH has higher efficacy and useability standards than anything related to the War on Non-Western, Non-Irish, Non-Russian (and "non-former-Soviet") Terror. At least the BOFH's systems work for him, you asshats can't even manage that despite taking all that daaaaaaangerous toothpaste away from us.

However, even I overstate the case here - Encoding data in background noise doesn't break any laws!

We all have every right to send hidden data, or even to use hard encryption right in plain sight. However, exercising that right may lead to some undue scrutiny, and thus we expose the real reason for techniques like this... Erosion of plausible deniability, which The Powers That Be loathe far, far more than any actual threat. It looks bad to just deport and torture someone with no evidence. But if you can demonstrate that he had (gasp!) something he didn't want the whole world to know about (because only criminals have secrets, of course), well then the sheep will approve of going all Jack Bauer on him.

Fundamentally flawed

[by Anonymous Coward]

Data can only be defined as varying bits of a defined pattern. So if the pattern is defined as 'a bunch of numbers that are either 0s or 1s', then the data stored within it is defined as varying the positions of 0s and 1s.

Obscuring data equals obscuring the patterns. So, to obscure the data within a 0 and 1 pattern, you might switch around the 0s and 1s.

For a message embedded in the background noise in a phone call, data may be modulated as 'loudness of background noise within a certain frequency range' or whatever. Obscuring this would be to add random data in the frequency range or whatever.

But that actually takes knowledge of the pattern used. If the pattern is rather the speaker knocking on a table, then any method designed to obscure background noise wouldn't register it or obscure it. It's similar to a scrambling technique that randomizes the 0s and 1s on a diskette sent in the post, while the actual message may be morse code holes punched in the plastic.

Conclusion: To void steganographic data, you need to know the method used to embed it.

Hiding information in an executable is easy

[Terje Mathisen (128806)]

They key to hiding data in executables is to realize that there are many instructions with multiple possible encodings.

You can also reverse the order of many comparison operations as long as you also modify the following branch/set instructions.

If you want to jam such a channel you would have to do the same job, first identifying all the possible locations for such transformations, then randomly flip half of them.

(Un?)fortunately neither the encoding nor the jamming process can be totally secure, because you can check (or know up front) which compiler had generated the original executable, then decompile/recompile and check which encodings the compiler tend to use.

Terje

Wow, more money spent on foolishness

[kurt555gs (309278)]

This could be better spent on more cell towers, or not allowing bastard fone companies to charge $200.00 termination fees.

Stopping secret messages? , puleeese.

"John has a long mustache"
"The chair is against the wall"

Stop that!

Steganography and watermarking.

[MartinG (52587)]

I'm sure someone will correct me if I have missed something, but it seems to me that the desire by some to hide irremovable watermarks within digital streams is a similar technical challenge to adding steganographic content. Similarly, those attempting to destroy watermarks will face the same problems as those wishing to remove or destroy steganographic content.

The interesting thing is who is on which side of the battle.

Generally it's corporations who like the idea of watermarks, and individuals who don't. Individuals do however like steganography, but the authorities don't. It will be interesting to see who develops what technologies and who, if anyone, wins this arms race.

The real question is..

[lakiw (1039502)]

How often do people hide data in the background noise of their phones? Is this a big enough problem that we should care about solving it? I mean, first of all you need a program to do the stego, (short of having someone talk really softly in the background). Then you would need to play back the recording during your conversation. Wouldn't it be easier for the criminal to send an encrypted e-mail instead? Given a choice, I'll pick strong crypto over stegonography any day. The only good thing about stego is it's useful if whatever authority in charge blocks all unauthorized messages.

It's along the lines of "How do you tell if there are stego images on someone's computer?"

Answer:You find the stego converter tool on their harddrive.

Re:

[smellsofbikes (890263)]

I don't know how often people have done this with phones. I've done stego in noise in pictures, when I was exchanging email with a friend who was living in China; we used a Matlab function. (It relied on her getting pictures from me and comparing them to the originals posted on a US-based website.)
The nice thing was precisely that it wasn't encrypted so the messages didn't just disappear, as so many others we sent did. (We started serializing our messages so we could tell when ones were going missing.)
So

Snoops

[Detritus (11846)]

How about not monitoring my calls in the first place? I am at a loss to understand the mindset of a person who thought that this was a problem that needed a solution.

I want end-to-end encryption on all my calls. This could be added to cell phones with some modest changes. Not having it on VOIP is just inexcusable. If the FBI wants to tap my phone, why don't they get off their lazy asses, obtain a warrant, and do some actual work, rather than expecting everything to be handed to them on a silver platter, complete with booze and hookers. I'm under no obligation to make it easy for them.

Why block?

[redelm (54142)]

First and foremost, I'm not sure it is moral or ethical to block any form of communications, crypto or stego. One might well claim certain communications are illegal and facilitate harm. But that is for already-illegal and incontrovertibly harmful activities apart from the communications. Police authorities are grasping at communications because they are otherwise impotent (by design). Fighting against stego or crypto seriously risks causing greater, even if less-spectacular, harm. Baby out with the bathwater.

That said, it is relatively easy to disrupt stego by lossy compression/decompression or vice-versa if the source is compressed. Low-order bits will get stripped in JPEGs & MP3s. This obviously doesn't work for loss-less compression as is needed for binaries. If hash or other non-compressibles found, just rehash. Once you've decided to meddle inthe datastream, some eggs will get broken. You'll have both alpha and beta errors (misses and false postives).

Sounds impossibly

[MobyDisk (75490)]

If you could detect and modify the background noise, then you could simply eliminate it. But I don't think that is possible, since what makes something "background noise" is the fact that it can't really be removed without damaging the foreground signal. If it could, you would have a perfect signal-to-noise ratio. Such a technology could be used to improve the bandwidth, compression ratios, etc. - which is something far more useful than fearmongering.

Unfortunately, I don't real have anything to go on other than a Google translated abstract, a Slashdot headline, and armchair knowledge of electronics. Anyone care to correct me?

Steganography in program files

[Beryllium Sphere(tm) (193358)]

It doesn't have white noise, but a program has enough places where you can replace code by a functional equivalent that you can pass messages in modified executables: http://www.crazyboy.com/hydan/ [crazyboy.com].