News for hackers. Stuff that matters.
Re:Show HN: Tired of the non-portability of my playlists, I ... by lucasgonze on Wednesday March 01, 2017 @03:31PM (#13768017)
There is no standardized canonical fingerprint and never will be, because the audio features covered by the fingerprint vary with the use case.

For example a fingerprint might or might not need to draw these distinctions:

- Clean ("Walmart") and dirty (profanity) mixes of the same recording

- Amount of opening and closing silence (often varies across releases)

- Different masterings

- Fidelity of the source recording (e.g. 64K Real Audio vs uncompressed PCM)

In addition, fingerprint developers are constantly improving on their algorithms.

There are no comments on this comment yet.