By “character” and “identity” I presume you mean more specifically timbre, natural envelope (attack, decay etc) and EQ fingerprint of the sound? As I’m sure you are aware sound can be manipulated beyond all recognition and anything in between. It all depends on your goal. If you merely want to clean a piano note to remove hiss for example, then careful NR in good restoration software can retain the sound attributes of the note very transparently. If your goal is to do something interesting with the sound by taking it to new sonic areas (which is half the fun of sampling) then the attributes will obviously change and it is just a question of how drastically that change occurs. One way to have the best of both worlds is to allow the user a certain amount of control about how much the sound is manipulated in the UI with the original sound or close to it still being able to be played back.
Of course, strictly speaking the sound is already compromised once it’s been recorded even if you are using the best equipment. It has already started its transformative journey so to speak. The ones and zeros on the hard drive are only a close approximation of the sound that happened in nature (or generated electronically).