I was just wondering considering those watermarked librairies which still get pirated none the less...

 

There was a book on watermarking posted on the sister site, I don't remember the name but the page count is about 950-1000, so it might be more complex than a ''hearable noise'' thing... It's mostly a mixture of time-related and frequency-related watermarking but it's way more deep than something which us mortal can understand (and hear...)

 

I don't understand how it can withstand processing with FX and the like. Isn't the watermarking lost somewhere in the process?
i will look after the book you mentioned

 

My guess is most watermarks affect the frequencies above and beneath human hearing. E.G., beneath 20hz or above 20khz.

 

Frequencies that usually cuted. Very smart move. :D

 

I'm pretty sure that's not how it works. The watermarking is in the nki, nkc, nkr, nkx (KONTAKT format) files, not the audio files themselves. I'm not a 100% certain, but to me it's highly unlikely they would be able to watermark the audio in any way that would be so strong to survive dithering / fx chains / noise reduction, you get the point. If it's in the audio there's a ways to remove it.

Unless it's not in the audio itself but rather in the file. I think I read something a long time ago about hiding information in audio files, without it being in the playback audio.

They say that you if you resave a library it will still be watermarked. I've looked through a few KSP script and almost all of them are obscured so that one can't read it easily. There's pretty much just a bunch of numbers and letters, nothing coherent. That might be where the information is stored.

Or I might just be creating theories and talking out of my ass.

 

From what I understand (and it isn't a lot), yes, it is for identifying the uploader. But (although not used due to impracticality--you could cut everything beneath or above a frequency range for example) audio watermarking within an audio signal itself IS possible. Just not practical. At all.
I googled "audio watermarking" and got this:

An audio watermark is a unique electronic identifier embedded in an audio signal, typically used to identify ownership of copyright. It is similar to a watermark on a photograph. Watermarking is the process of embedding information into a signal (e.g. audio, video or pictures) in a way that is difficult to remove.

Edit: BTW, I like your avatar DFox. ;)

 

What do you mean? That companies don't care about the use of their librairies as long as they can identify the uploader?
That makes no sense at all.
Here's why:
Legally, you are NOT responsible if somebody stole your goods and shared it, and it's a real legal pain in the ass to try tp prove it.
Second, when you're a company you mean business. What is the greatest asset? Having "caught" the uploader? Or stopping people from using your librairies without having bought them?
So if watermarking is useless from stopping the sharing and only dedicated to identifying the uploader then it's actually the stupidest method ever used to protect a sound.

And it's not because people have said the same stuff a thousand times that is is more true...

 

Steganography is useless once you modify the image containing the cipher. So it's the same for audio. Basically. Once processed, you cannot reconstruct the watermarking. Or i need a more precise explanation...

 

I found the book. I'll read it and i'll come back if i find some real insight

 

ok. No need to actually read the whole book. The answer was easy to find and has nothing to do in just identifying the uploader. The goal was obviously to keep the watermarking inside the audio files as long as possible.
It seems that the watermark is "heavily threatened" (sic) by time related manipulations. It is fragile and makes heavy use of psycho acoustic algorythms for its implementation. The more you process the sound, the less you can be caught.

Thanks guys for your cooperation

 

tone2 has watermarking in their products.

 

Some companies claim that they have methods to watermark something without changing changing the binary integrity of a file, which imo is bullshit. Everything digital is either a 1 or a 0, and if you add watermarking it's going to be a different combination of 1 and 0 that make up that file compared to one with another watermarking. There's just no way to have different information and still have the same binary code of a file.

Anyway... I found this
http://audiowatermarking.info/awt2_main.php

Yabiss, it better for companies to kill the source rather then go after everyone using their library. Sure their library is already shared and will not be unshared just because they go after the unloader, but at least he won't be uploading anymore of their stuff. It's easier to go after the ones sharing the library rather than the ones using it. If they were to go after me they would be making a mistake.
Yes, maybe I've downloaded their library and used it in a few of my tracks, but if I really like it and use it, then I will buy it. So if they go after me they lose a customer.
Also when they convict someone for uploading a library (I only know of this happening once, and that was not really thanks to any watermarking iirc) they go based on the number of downloads (potential sales). But if I download a library and use it what are they supposed to base their sentence on?
Please read the book, I would love to know what it says. But from what you wrote it sounds like it way over my head. Also sound like something that would change the audio.

 

Btw do anyone of you know of anyone getting caught uploading libraries or using them?

 

are you people paranoid or something ?

 

I think Kontakt libraries have to do with the library files or audio containers themselves(not sounds) in order to identify the customer whose library was uploaded and shared on the net. Most of what I've seen as far as watermarking audio files deals with data embedded in the container and not the sound. And usually when it does involve the sound it is deliberately audible such as those used in sample/loop pack demos. However, I can tell anyone for a fact that undetected audio watermarking(the sound) is possible because I have my own method of watermarking all my audio material although it isn't always necessary. I haven't seen any other example of it being done the way I do it. I call it fogging the spectrum. My signature is written directly into the audio spectrum and repeated throughout the duration of the material. It is mixed with the audio material just enough to be visible, It spans the audible frequency ranges, but is not loud enough to be heard. And the best thing is I can create the signature at a specific 3D viewing angle so it can only be seen using a 3D analyzer set at precise degrees. But again, it's so easy for one to prove they are the creator of their material nowadays that you don't need a paid copyright unless you decide to sue someone later down the line.. what I do is not necessary but it serves well for public arguments sake should their be any rumors or discrepancies about "who did what".

 

Last edited: Nov 19, 2015

@korte1975 ,

Yes, I am. I know it's stupid but I really am paranoid about things, and this is one of them.

@dbmuzik ,

That's not really the same watermarking as what the companies do. They do it, not to prove that they are the creators of the material, but to see who shared it to begin with. Also the way you do it is actually audible, just not distinguishable from the actual music. And I really doubt companies would do something like this. Because if they watermark every single audio file and you play a chord of let's say 5 notes then that watermark would be 5 times louder (if it's in exactly the same location on each file). Also they pride themselves sooo much on their "pristine-ultrarealistic-amazing-sounding" libraries that I really hope they are not stupid enough to add something like that to their audio files.

But the way you do it is pretty cleaver and also pretty common.

 

Do a simple DA-AD conversion are you're done.
Watermarking exists only in the digital domain. Once the (audio) file is converted from digital to analogue it looses ALL digital informations.