[Cryptography] "The Visual Microphone: Passive Recovery of Sound from Video"

[Jerry Leichter]

On Aug 6, 2014, at 2:51 AM, Tom Mitchell <mitch at niftyegg.com> wrote:
> This makes me wonder more and more what a soundproof wall or window 
> might be composed of.   With this type of effort the acoustic  attenuation
> may need to be astounding and futuresque....
The approach the spooks seem to have taken for years is "a room within a room":  Essentially a big box you set up inside some room somewhere.  It would be heavily soundproofed and mounted in such a way as to isolate any vibrations from the containing room.  There would be no reason for the box to have any windows.

While in theory no matter how well you insulate the walls of the box *some* internal vibrations are transmitted - but you don't need perfection.  Noise is inherently present, and if you can reduce the transmitted signal to be far enough below the noise, you can lower the available data rate below a point where you feel safe - basic information theory about channel capacity used in the reverse of the usual direction.  (Of course, you can also deliberately add noise to the out walls.)

It's an endless battle against new technologies.  Remember the famous "metal diaphragm embedded in the eagle seal and read by microwaves" from the annals of the cold war?  Given today's signal processing and RF technology, you can probably find *something* metallic in almost any room whose vibrations you can read off at some RF frequency and then beat into a source of audio with enough computation.  So now you need to screen or drown with noise a huge swath of frequencies.  Using a Faraday cage as part of the "room within a room" is probably standard, but practical Faraday cages (a) are typically built as meshes, which these days have to have very small openings as an opening comparable to the wavelength of incident RF makes the screen ineffective; (b) the materials of the Faraday cage itself are potentially great sources of reflected information modulated by sound from the room.

At the spook-against-spook level, this is probably a more or less even game, at least for well-funded spooks working for organizations that have access to leading-edge technology.  For everyone else - even including most of the world's security services - defense is probably effectively impossible against the high-end spooks.  And the trends seem to favor other attackers as well:   The attack technologies tend to be adaptations of common technologies (e.g., $1500 digital cameras) so are much more broadly available and affordable than the defensive technologies.

                                                        -- Jerry

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4813 bytes
Desc: not available
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20140806/155f92df/attachment.bin>