views:

41

answers:

1

I am trying to create an audio watermarking system. I figure the best solution is to create an audio file (WAV) based on a unique string of text and then combine this with the original wav.

The part that makes this tricky (for me anyway) is:

  • How do I convert the text string to a wav?
  • How do I ensure that the resulting WAV form is inaudible (or at least barely noticeable to the listener).

I would prefer this be done server side (via PHP, etc) but if the processing load isn't too much then would be ok with something in Flash or Javascript.

I'd be willing to pay someone to create me a workable solution (complete source code that functions as described).

Thanks, Scott!

+3  A: 

I think audio watermarking is a harder problem than you realize. You might want to consider using existing systems before trying to roll your own.

Karmastan
mp3stego (http://www.petitcolas.net/fabien/steganography/mp3stego) looks like it exactly what the author needs.
caspin
@Caspin thanks for the feedback; however, there are quite a few companies that offer watermarking that survives audio compression to MP3, etc. So I know it can be done, just trying to figure out how to do it using algorithms in the public domain.
Scott
@Karmastan Thanks for the links. My problem with those existing solutions is that I need this to be some kind of audio watermark and not data inserted into the file using steganography. The reason being I need to be able to encode the files so that if they are streamed, recorded, and then converted into MP3s, I can still pull out the unique audio watermark.I know this is possible - there are a number of research papers on it and several companies have commercialized the application of this, but I am hoping to find a way to code this myself.
Scott