Re: hacksdmi?

[Steve Mosher <farq () KILN ISN NET>]

On Thu, 12 Oct 2000, Bluefish (P.Magnusson) wrote:

> > P.S. Yes, the whole premise of "secure music" is fundamentally
> > broken.  Yes, the minute someone figures the algorithm, the
> > watermark is gone.  Yes, converting it to an MP3 would hopelessly
>
> This was one of the subjects in my computer security class, actually. A
> secure watermark is to survive any filter and AD->DA or DA->AD conversion,
> compression and [continue pasting in any common standard operation on
> images and audio] *and* be impossible to remove without mayor
> sound/picture degradation (even if the algorithms in use are known)

        I'm thinking that secure DA->AD->DA watermarking is an unholy
graal of sorts, i.e. not something I'd ever expect to really work. The
watermark, in this case, has to exist in the sound, somewhere. If you know
the process, what's to stop you from mucking with that portion of the
sound?

        The only way to really make it work (as I see it) is to make the
watermark a *fundamental* portion of the sound, but that would be entirely
moot since there's no way to apply the watermark to existing sound, thus
it's useless.

        Of course, if you don't care about DA->AD->DA (or even DA->AD)
then it's less of a problem, at which point you only need to ensure that
decompression + recompression is as (or more) costly (qualitywise) than
DA->AD->DA.

> Ps. Thanks to everyone who hasn't participated in the hacksdmi contest.
> Your efforts to defend civil rights & fair use all over the earth is much
> appriciated :)

        The funny thing is that we're telling technology that it's "not
allowed to go there, at least for now," by boycotting the contest. I don't
really mind in this case, though.