Slashdot Mirror
Watermarking to Replace DRM?
An anonymous reader writes "News.com has an article on the announcement of Microsoft and Universal to introduce watermarking technology into audio files. The technology could serve several purposes including tracking file sharing statistics and inserting advertisements into audio tracks. The article goes on to suggest that watermarking could possibly replace DRM in the near future."
Which is precisely why it won't work. What one tool can detect, another can circumvent.
Oh, and it's detectable and not detectible. Don't know what moron at news.com.com hired Taco...
This message is brought to you by the Bureau of Massively Distributed Peer Review, Department of Free Culture.
The Banjo Players Must Die!
Yes, it is even more trivial than you think : simply re-encoding the whole song (from mp3 to aac for example) will remove the carefully placed watermark.
It's the same as with invisible watermarks in pictures. Just save the image (normally a jpg) again, with recompressing, and the watermark will be mostly gone.
You can argue that re-encoding songs like this diminishes quality, which is true, but it is so little that it's not audible. A minor, inaudible quality loss is a small price to pay for DRM/Watermark free songs.
"Some days you just can't get rid of a bomb."
You have absolutely no fscking idea what you're talking about, do you? ... It has absolutely nothing to do with advertising whatsoever.
I'm guessing that he actually read the article.
FTFA:
Activated Content hasn't explained exactly how it'll use the Microsoft technology, but the company's Web site promotes a very interesting service called ActiveNow. The idea: whenever a watermarked file is played on an ActiveNow-enabled device, the service could dynamically insert some sort of advertising--presumably audio, but perhaps video or text depending on the device being used.
Douchebag.
"Tell me doctor, with all of your defenses, are there any provisions for an attack by killer bees?"
It's not absurd. Welcome to the article:
"FTFA:
Activated Content hasn't explained exactly how it'll use the Microsoft technology, but the company's Web site promotes a very interesting service called ActiveNow. The idea: whenever a watermarked file is played on an ActiveNow-enabled device, the service could dynamically insert some sort of advertising--presumably audio, but perhaps video or text depending on the device being used."
-- "It's not stalking if you're married!" My Wife.
The RIAA did go after used CD shops and lately they are winning.
In Florida, the new legislation requires all stores buying second-hand merchandise for resale to apply for a permit and file security in the form of a $10,000 bond with the Department of Agriculture and Consumer Services. In addition, stores would be required to thumb-print customers selling used CDs, and acquire a copy of state-issued identity documents such as a driver's license. Furthermore, stores could issue only store credit -- not cash -- in exchange for traded CDs, and would be required to hold discs for 30 days before reselling them.
This is now in Florida and Utah trying for Rhode Island and Wisconsin (and potentially elsewhere).
It's drafting us for market research, not preparations for lawsuits.
I posted a comment at news.com with basically the same idea.
If the bits and bytes can be adjusted in an undetectable manner to put a watermark on, say, an audio or video file, why can't someone just come along after and adjust the bits and bytes again in some random manner to effectively erase the watermark? I mean, if they can't read the bits and bytes that they put on the media because they've been altered, they wouldn't be able to track it, and the watermark would pretty effectively be broken.
It just seems to me that although having a bit-for-bit identical copy of the original would be nice, they've already altered it so that we can't get that. Altering it a bit more (no pun intended) wouldn't really be harmful, and it would still meet the end goal of distributing the media untraceably.
But you're right, another option would be to have two (three? four?) accounts get multiple copies of the same file and do a bit-by-bit comparison, either averaging the differences or picking from one of the two copies at random. If you have multiple copies, you might even be able to derive a highly probable copy of the original.