How does sample watermarking work?
-
- KVRian
- 996 posts since 14 Jun, 2012 from Toronto, Canada
I would love to know. Is there a paper I can look up? I wonder what methods devs use to create watermarks that would be robust enough to survive through dithering, etc.
-
- KVRist
- 33 posts since 16 Jan, 2012
If you're referring to what many Kontakt library developers do, the samples aren't watermarked. The library patch files are, however.
Audio watermarks don't work.
Audio watermarks don't work.
-
- KVRian
- Topic Starter
- 996 posts since 14 Jun, 2012 from Toronto, Canada
Hm, why is it interesting? I just don't know whether there is such audio watermarking because there doesn't seem to be any point otherwise.JD Gaffe wrote:Watermarks robust enough to survive through dithering. That's such an interesting sentence.
It's just a purely academic question, because I read things such as this: http://audiowatermarking.info/awt2_details.php Sorry if it seems such an ignorant question.
-
- KVRist
- 33 posts since 16 Jan, 2012
Nah, it's not ignorant to ask, but in a place such as this, one can only assume you're asking as it pertains to a sample library.
Audio watermarking was originally proposed for distributed music, as the average listener on average consumer equipment would not be able to hear what it does. In studio audio, however, it is glaringly obvious unless the samples were already of poor quality.
If your investigations prove fruitful in this regard, I suppose it would be a good idea to blacklist their customers, since they're sacrificing quality for supposed security.
Audio watermarking was originally proposed for distributed music, as the average listener on average consumer equipment would not be able to hear what it does. In studio audio, however, it is glaringly obvious unless the samples were already of poor quality.
If your investigations prove fruitful in this regard, I suppose it would be a good idea to blacklist their customers, since they're sacrificing quality for supposed security.
-
- KVRist
- 450 posts since 22 Aug, 2007 from Los Angeles
I think JD was just making a joke that "Watermarks robust enough to survive dithering" has kind of a sexual sound to it.schnapsglas wrote:Hm, why is it interesting?JD Gaffe wrote:Watermarks robust enough to survive through dithering. That's such an interesting sentence.
Your question isn't ignorant at all. Watermarking in the days of CD copy protection attempts (SCMS, for instance) were indeed modifications of the audio files themselves. Whether they really effected the audio quality, I don't know.
But as far as sample libraries go, I don't know any developer who watermarks the audio files themselves. We (Realitone) don't, for instance. Instead, the watermarks are hidden in the Instrument files and KSP scripts. They have no effect whatsoever on the sound or functionality of the instruments. They basically just amount to hiding a serial number in the code, so that if you illegally share your copy, we can see who the sharer was.
In the interest of full disclosure, with Realivox, we do also watermark the NKS sample monoliths, but it's in the header information, not the wave files. (It's not actually the header, but for obvious reasons, I can't say exactly where.) The audio itself is completely untouched.
