Your speakers are analog. They sound different when they are dusty. Your ears are analog. Things sound different when you have dirty ears. Nyquist-Shannon sampling theorem only applies when there are no frequencies outside of the sample range, which doesn't happen in real life. None of this matters, because like I said it's trivial to have orders of magnitude more accuracy than you need. Digital is just way cheaper to copy accurately, so that's why it has become dominant, and that's fine, but the idea that it's inherently more representative of reality is just gibberish.