Exporting match metadata


I am running a music informatics research project that involves scraping some audio data from the internet. I get the track metadata from Spotify and then seek the audio from YouTube. (I hope we can set aside the ethical discussion for now :slight_smile: ).

Since the mapping from Spotify metadata to YouTube URL is error prone, I would like to use MusicBrainz Picard to (1) fingerprint the tracks and (2) verify that they match (approximately) the Spotify metadata.

I would like to generate something like a CSV or JSON file that has the original values and the new values from the fingerprinting lookups. Is this possible? It seems like the UI has all metadata I need.

1 Like

I don’t think that MB Picard can return a CSV or JSON for your use case.

But you could fingerprint your songs with the existing fingerprinting tool chromaprint/fpcalc from AcoustID.org
With the locally created fingerprint, you can query the AcoustID WebService. In best case, you
will get a (hopefully matching) ID from AccoustID.

With this ID you can query the API from MusicBrainz. This API return a result in XML or JSON.

1 Like