2015-09-30, 13:31
(2015-09-30, 12:48)DarkHelmet Wrote: Dave, while this is not directly related to classical music and the database logic, I still want to bring this up. It has more to do with scraping/identifying the tracks. I've thought about a shazam like solution for cases where the results of scraping is not 100% consistent to recognize the track.
https://acoustid.org
This is an open source database using an acoustic fingerprint and from the website they claim to include or relate their data to musicbrainz. Might be worth a look, because in the end I think an acoustic fingerprint is the future and the only reliable way to identify a track.
This might indeed be a good way to go. But I think it still won't be able to match correctly, as a single might have the same fingerprint as the album version. But the data will be better in both cases.
Theres an C library for the calculation part:
https://acoustid.org/chromaprint
And here is a example using FFmpeg (which we also use) to generate the prints:
https://bitbucket.org/acoustid/chromapri...ew-default
