|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [WM]: Another use for steganography
Thanks for a very good message. Perhaps you can translate those papers into English.
Now, let's get down to work. Can anyone offer me a program that can generate a stream of white noise with a steganographic or other
watermark, and save it to an file that could be used to burn a 74-minute music CD?
I would also like an alternative to white noise. Some years back someone came up with a musical stream that had the weird effect of
seeming to be constantly falling or rising in pitch even though it doesn't actually do that. It seems to exploit some side-effect of
the human hearing system. The reason I mention that is that it would seem to be a sound that, if superimposed on an audio recording,
would make editing obvious to the ear without having to get into crypto analysis.
In the meantime, I may have to resort to playing some electronic music in the background that would make sampling and removal
difficult.
Vadym Fedyukovych wrote:
> hi Jon,
>
> it seems you did introduce a very interesting attack model that (I
> think) was never considered in academic papers.
>
> Given a message signal M, as recorded by other party' mic mixed with
> watermark signal W as played by the device while possible recording,
> convince some 3rd party that
> 1: watermark is present in message, in every time-frame of it
> 2: all watermark frames are there in place, without
>
> There was a public-key technique suggested for first task that include
> some random noise added to an artwork, homomorphic public-key
> encryption of that noise provided as a reference, producing encrypted
> correlation function of the mixed artwork and the noise and a
> zero-knowledge proof that correlation is higher than a threshold. Or
> maybe ZK proof replace by decryption of correlation only (but not the
> reference).
>
> I believe, this technique could be applied to every frame of recorded
> message very well. Someone using it would be making multiple statement
> on the watermark generated as long as he keeps his private key
> private.
> In case you're interested in details, please consider to take a look
> at my writeup at http://www.cryptography.ru/db/msg.html?mid=1169847
> Actually, that's in Russian so please consider to pickup references
> from there. Or maybe reading the math at that writeup would make you
> some idea. For your convenience, it is attached to the message..maybe
> the site will be fixed after NY eve
>
> Homomorphic encryption is the major part of that technique.
> Please consider a code at
> http://www.mail-archive.com/cryptopp-list@eskimo.com/msg02399.html
> that would unlikely result in any IP claims
>
> To accomplish the second task, ransom noise should be replace with
> something that could be proven a continuous stream.
> That will not harm technique based on encrypted correlation function
> in any way. I'm not quite ready with with all the details, just trying
> to combine some synchronization code and dependencies on previous
> frames with proof of knowledge of some secrets that could be
> demonstrated to a 3rd party.
>
> Please let me say your challenge looks very interesting for me.
> I'll write again, in case having something to fit the second task above.
> At this point, I'm curious whether you'd consider the scheme outlined
> good enough
>
> Best regards,
> Vadym Fedyukovych
>
> Jon Roland wrote:
>
>> Ron van Schyndel wrote:
>>
>>> A better option would be a pseudo-noise based WM in the microphone.
>>
>>
>>
>> For my purposes that won't work, because I won't control the other
>> party's microphone, which is the one whose recording is expected to
>> be altered. I might make my own recording that could be compared with
>> the other, and mine, having a watermark, would be more credible, but
>> I could be accused of adding the watermark afterward.
>> The solution I envision would add the watermark at the source, my
>> voice, not the pickup, the mike, and therefore to all recordings
>> using anybody's mike. It would be pervasive throughout the room,
>> including for any paraboloid mikes across the field.
>>
>> For this, you have to think spook. What would Tom Clancy's characters
>> need?
>>
>> -- Jon
>>
----------------------------------------------------------------
Starflight Corporation 7793 Burnet Road #37, Austin, TX 78757
512/374-9585 www.pynthan.com starflight@pynthan.com
----------------------------------------------------------------
______________________________________________________________________________
Watermarking Mailing List - http://www.watermarkingworld.org/ml.html
To unsubscribe send email to "majordomo@watermarkingworld.org" with
"unsubscribe watermarking YOURMAIL" in the body.
______________________________________________________________________________
|
