I don't know a lot about how ebook files are constructed but if they're not text files like simple HTML it might be possible to include identifying information in binary portions.
Once, before Audible was bought by Amazon I participated in an experiment with them. They gave a number of us in an Audible forum Audible audiobooks with identifying information and asked us to see if we could decrypt them in some way and rid the books of their watermarking. We each encrypted our book. We then exchanged decrypted copies and did whatever we thought might destroy the watermarks. I converted mine from MP3 to 2 or 3 other formats and then back to MP3. Others did similar things. Then we all returned our files to Audible and they told us exactly which file had been originally given to whom. They got it 100% right which speaks well of the durability of watermarking.
If there are binary portions, or maybe even text portions, of an ebook that a program like Calibre doesn't concern itself with, those could be where the watermark is placed. I don't know that there are such areas in an ebook file but if there are that might be the answer to the question.
Barry
|