I'm not, but I'm serious about putting it to the test. I assume you have a working implementation of this? Why not put out a watermarked audio file, and let me hammer on it for a while? Let's find out if your watermark is as robust as you hope it is ;)
The point is not to make copying impossible, just to make HIGH-QUALITY copying SUFFICIENTLY difficult that people will buy the CDs instead, and this is certainly achievable.
Come, now - you're obviously smart enough to spot the flaw in that kind of plan. Namely, it only takes one person figuring out the magic formula for munging your watermark to kill your plan stone-dead, just like it only took one person figuring out the way the DVD CSS system works, and how to beat it, in order to render it totally ineffective at stopping anyone from copying DVD's at will. And that's because the next step is to write a program to automate the munging process, and make it available to everyone.
Just ask Jake Johanssen. High-quality from-scratch copies of DVD's are well beyond the abilities of most people, but pointing and clicking is not. And it doesn't matter that the vast majority of people have no idea how DeCSS works, or how later, cleverer solutions work, or what you'd have to do to go about solving the CSS problem for yourself, because they don't need to know those things. Some clever person did the hard work for them. You don't need a whole planet full of smart people to kill a scheme like this - you only need one ;)
I don't have this implemented, I just know that the theory of it works -- it's my own scheme, inspired by the postings on this thread. Trust me, I have the Computer Science background to know this is doable -- feel free to implement it yourself and try to patent it, I'm sure you will find someone has beaten you to it.
Come, now - you're obviously smart enough to spot the flaw in that kind of plan. Namely, it only takes one person figuring out the magic formula for munging your watermark to kill your plan stone-dead, just like it only took one person figuring out the way the DVD CSS system works, and how to beat it, in order to render it totally ineffective at stopping anyone from copying DVD's at will. And that's because the next step is to write a program to automate the munging process, and make it available to everyone.
There are actually two separate kinds of piracy problem this is trying to address.
The first is simply to make it difficult for people to untraceably copy their CDs and give them to their friends who then won't have to pay the record company and the artist. My scheme certainly does this, because it becomes provable who bought the original copy of the music. This is not really compromised by allowing people to pay extra for an anonymous cash purchase -- you're balancing a lot of little thefts against a lot of little extra bits of revenue. A pirate would always be able to get hold of the music without identifyng himself anyway, by stealing a CD.
The second problem is the large-scale commercial pirate who munges the watermark or provides software for others to do it. The beauty of my scheme is that even if he does this, it is still provable that the music has been tampered with because the watermark is munged. Many people will have illegal munged copies of the music, but they will all be identifiable as munged (someone who has an unidentified UNmunged copy will be able to say that he paid cash, so there is no evidence he did anything wrong, but someone who has a munged copy is in the position of a person who bought a paperback book with the cover missing, prima facie evidence of participation in an illegal activity).
The place public-key encryption comes in -- it allows the record company to provide free software which will make it easy for anyone to READ a watermark or see that the watermark has been munged, without allowing people to create their own watermarks or easily munge them. This could be a feature in apps like RealPlayer -- when you play the music file, the serial number of the copy it came from is displayed.