Digital Watermarks (slightly off-topic?)

[Alex F]

> In list.cypherpunks you write:
> 

> Tough to do.  A CD samples at 44,100 Hz.  Nyquist says you can only
> resolve the original frequencies up to 1/2 the sampling rate.  So a CD
> cuts off, of necessity, at 22,050 Hz.  Many people can hear beyond 22
> KHz, and can notice the CD cutoff effect.  (analog recordings taper off
> as the analog response diminishes)  Not a lot of room inband.

Actually (speaking from personal experience) many many recordings are 
now done at a sampling rate of 48 instead of 44.1 (actually it may be 
around 50/50 or so, from my experience).  Many CD replicators these 
days are thrilled to get recordings done at 48 instead of 44.1, which 
gives even more room.

Another possibility, if you want to get even more into detail is to 
encode the ID in digital format on the CD.  This will give an audible 
sound (ever put a CD-Rom in a regular CD player by mistake?), but a 
mirror of that sound played at the same time will effectively set 
that sound to nothing (cancelling it out).  You can't hear it, but a 
machine can still decode the digital info.  This way you can set it 
at a higher frequency where if it happens to cancell out a brief 
second of music, the listener won't notice (unless you are "Jamie 
Summers" :)  ).



Alex F
=-=-=-=-=-=-=-=-=-=-=-=-=-
Alex F    alexf at iss.net
Marketing Specialist
Internet Security Systems
=-=-=-=-=-=-=-=-=-=-=-=-=-