From: Nathan Syfrig <nsyfrig@condor.depaul.edu>
Date: Tue, 23 Jul 1996 17:01:16 +0800
To: cypherpunks@toad.com
Subject: Re: Digital Watermarks for copy protection in recent Billboard (fwd)
Message-ID: <Pine.SOL.3.91.960722164534.22991B-100000@condor.depaul.edu>
MIME-Version: 1.0
Content-Type: text/plain



I got the following from the e$pam service.

---------- Forwarded message ----------
Date: Mon, 22 Jul 1996 14:56:43 -0400 (EDT)
From: e$pam <e$pam@intertrader.com>
To: Multiple Recipients of e$pam <e$pam@intertrader.com>
Subject: Re: Digital Watermarks for copy protection in recent Billboard

Forwarded by Robert Hettinga

-----------------------------------------------------------------------
Comments: Authenticated sender is <alexf@[204.241.60.5]>
 From: "Alex F" <alexf@iss.net>
 Organization: Internet Security Systems, Inc.
 To: cypherpunks@toad.com, "Deranged Mutant" <WlkngOwl@unix.asb.com>
 Date: Mon, 22 Jul 1996 11:19:17 +0000
 MIME-Version: 1.0
 Subject: Re: Digital Watermarks for copy protection in recent Billboard
 Reply-to: alexf@iss.net
 Priority: normal
 Sender: owner-cypherpunks@toad.com
 Precedence: bulk


> Paged through a recent (June or July 13) edition of Billboard
 > magazine yesterday.  There was an article about the music industry,
 > the internet, and copyright issues.  Didn't have a chance to read in
 > thoroughly, but it mentioned using digital watermarks which contained
 > info on to who (CC number) and when the material was sold... the
 > watermarks allgedly could survive if a CD was taped, copied several times
 > and redigitized.
 >
 
 Easy enough.

- Unless somebody reversed-engineered it, filtered it, and re-stamped it.
 
 > The anti-piracy scheme is only useful for direct sale to a customer
 > though.  If you buy music anonymously, how is it traced?  This only
 > works for pirating on-demand purchases.
 
 This is probably yet another case of people not thinking ahead.  As
 usual.  People buying CDs at a garage sale & getting arrested for
 piracy.  Wonderful.

- The entertainment industry has a reputation of being paranoid, ever 
  since individual cassette duplication became popular, albeit with
  reduced signal quality (a lot of people don't care as much about the 
  signal quality and the industry knows that).  Some of you may recall 
  the flap over DAT, which significantly reduced the consumer market 
  penetration (the industry itself uses them all over the place).

  The industry is also not known for forward-thinkers, even though 
  they can and do hire them on occasion.

  While they aren't going to be worried about $8.00 at a garage sale, if 
  they see mass single-copy distributions going on at enough garage 
  sales, but they only sold 30 "master" copies, they might get concerned.
  Hey, judging by the announcements of how much piracy costs the 
  industry, they probably don't rule this out (again, being all digital, 
  under the current copyright scheme, there might even be a point).  Yup, 
  time to rethink the whole concept of copyright, intellectual property, 
  et al, although I have no idea how to approach the issue.
 
 >
 > Other issues: what if an eavesdropper steals the music or video? It's
 
 If they steal it, well, who cares?  If there is something worked out
 so that they could trace STOLEN (not traded or sold) CDs then fine,
 arrest them.  Do you really think though that anyone would waste so
 much time over $8?

- if it's too easy, then tools to do it will become so widespread that 
  even the average user will engage in such practices.  This time, 
  being digital, the reduced-quality incentive doesn't hold (you still 
  have the even-less-effective argument of the associated cover art not 
  being included or being scanned and duplicated with reduced signal 
  quality, unless the distribution is all on-line).

  Bottom line:  There really is no way around this in the long run, but 
  there's a l-o-t of money at stake.  Therefore, delaying tactics are 
  worth something to the big players in the industry, which is what we 
  are seeing (some people might not 'get it' but there's enough money 
  to where people who 'get it' can be and are hired to gum up progress).
 
 > If it uses a credit-card number as (part of) an ID, that's pretty
 > bad.  Someone can sniff for CC numbers if they know how it's stored.
 
 Probably not done that way.  My guess is that the disk ID is assigned
 to the disk at the time of manufacturing.  At the point of purchase
 the customer is forced to give name, address, ID, whatever.  This is
 then stored in a database along with the disc ID (serial num) which
 is prolly printed in the ISBN number or cross referenced with that in
 a national database or something, or just printed right on the disc.
 Anyway, a number is given to you from the CD, and not vice versa, I
 would imagine.

- Would YOU want to be responsible for maintaining that database?  It's 
  like maintaining a hardware store trying to maintain an ID on every 
  single screw and nail in inventory.
 
 >
 > The system will have to rely on proprietary tech and security through
 > obscurity.  Even know how watermarks are stored without understanding
 > the math, one must be able to somehow garble the sound without
 > distorting it, but which renders the watermark useless.
 
 Actually, this would be quite easy.  The "watermark" would be a
 signal that plays inband, but out of our hearing range during the
 entire CD.  The human ear can only hear in the 20-20,000 (Hz, KHZ?,
 whatever) range.  It would be trivial to add a digital ID signal at,
 say 30,000 or 15 or something like that.  This could then be decoded,
 if need be.  This seems the easiest and most efficient way.  This
 could also be defeated with a lot of $$ (and/or a LOT of HD space).
 If the frequecy is known (it can be found out) it can easily be run
 through recording studio eqipment that can very effectively isolate
 the frequency and cut it out.  If you have a LOT of HDD space
 (digital audio at 2 stereo tracks, not sure of the sampling rate or
 bit resolution, takes about 20MB of HDD space per minute (2 tracks,
 good sampling and bit rate) ) you could probably find the freq.
 fairly easily by isolation and just edit it out, and write the new
 stuff to a CD-R.  If the signal is purely digital, I would imagine
 that it might be even easier that if it were an analog signal (?).
 Someone w/ good equipment (Digital Labs' stuff, or SAW (Software
 Audio Workshop) would be able to do this w/o much problem.  The
 question is is the price/effort worth it?  In quantity maybe.  On an
 individual basis, only if you already happen to have the erquipment.

- Nobody's going to try and do a higher-frequency encoding (I HOPE).  While 
  the human ear cannot hear those frequencies directly, we have found out 
  that those higher-frequencies interact in such a way to influence the 
  sound waves that influence what the user can hear.  This is the reason 
  there's still a debate between digital and analog recordings, and is 
  still a big reason a lot of artists still record on analog equipment 
  (in musical "fuzzy" terms, it's equated with the warmth of the sound, 
  sort of like the tube-amp vs. solid-state amp debate among some guitar 
  players, etc.)  If somebody deliberately played with such frequencies, 
  the journalistic media would probably have a field day.  Yes, there are 
  audio cancelling and other tricks that could be deployed, but no matter 
  what, you're still deliberately introducing signal noise (I wonder this 
  influenced the non-acceptance of "minidisks" from a few years back - 
  aside from it's incompatibility with anything else around)

  If I remember correctly, there is plenty of room in the design of the 
  audio CD protocal to embed such information, just like you can embed 
  the timing and track number information.  Some might remember the 
  sort-of craze of embedding stupid "graphics" and words to audio CD's 
  which special players could read and display on a monitor but didn't 
  affect normal audio CD players (Lou Reed's "New York" was one of the 
  few releases that I saw which advertised this "feature").  It turned 
  out to be too hokey even for the consumers of the time.  In other 
  words, there are plenty of ways of achieving this.  However, my guess 
  would be to use up 650Meg of a hard drive, copy the CD byte-by-byte, 
  and reverse-engineer away.  Then you could easily stamp a "clean" master.
  (DVI could change the game - I don't know what the status of this 
  battle is, other than it's shades of the DAT battle all over again)
 
 I have a suspiscion that this type of thing will not really come to
 any kind of fruition due to not only the ability to defeat this, but
 mainly due to things like buying at a garage sale, etc.  If it did,
 only MASS market piraters would be investigated.  (Another example of
 a law creating it's own violators.  Don't make the law, there won't
 be mass piratingof "clean CDs"

- Well, the MASS market piraters are exactly the point.  Well, let's face 
  it, if the industry controllers got their way, there would be no 
  second-hand market like garage sales - there IS money involved here 
  (witness the bizarre dealings with CD-rental stores that have shown up 
  over the years).  However, they are counting on the majority of their 
  customers not having the equipment to easily defeat this, which up 
  until now, has been the case.  However, recordable CD's have come down 
  dramatically, along with hard-disk prices, and all the tools required 
  are much more available than most people outside this list would have 
  predicted.  And from an industry perspective, as this list already 
  knows, it ain't gonna get any better.

- Bottom line:  I expect things are going to get pretty bone-headed.  
  Wow, such insight!
 
 Alex F
 =-=-=-=-=-=-=-=-=-=-=-=-=-
 Alex F    alexf@iss.net
 Marketing Specialist
 Internet Security Systems
 =-=-=-=-=-=-=-=-=-=-=-=-=-

Nathan F. Syfrig
(views are my own standard disclaimer)