what would you say is the difference between fingerprint and using something like OpenAI's whisper approach (visual spectrogram ML) for finding the music tange

what would you say is the difference between fingerprint and using something like OpenAI's whisper approach (visual spectrogram ML) for finding the music tangent: I'm also thinking about some fast way to search text algo maybe related to Spotify damn that was a long time ago read that article