Thought i'd chime in since I didn't see anyone answer this part...
Digital Files like Audio, digital photo's things like that. All have a lot of data within them aka "metadata" and other fancy words.
So if you take an audio file, and say cut it in half in a video editor, and put it anywhere in your video, the "ID" for the file itself is still their, and easily found via the auto scanner.
Think of it like serial numbers.
I'm not a computer guru, nor do I even attempt to pretend I am.
Youtube claims they detect via file ids. what that means in the technical world is up for debate.
Thanks for sharing that, it was a good read.