While tagging my music collection using Picard, a few titles were identified as being part of compilations.
I would prefer to use the canonical data, i.e. the first album a title was released on instead, and the MusicBrainz Canonical Data Dumps would help with that:
However, this doesn’t seem to be used by Picard. When I use the option “Search for similar tracks…”, dozens of entries are shown, but no canonical version.
Will this be integrated into Picard at some point, or is the only use case writing one’s own tagger (or using the example provided on GitHub) ?
@IvanDobsky’s answer is an excellent one, but this situation can also arise when you use a Scan-base workflow rather than a Lookup-based one.
In addition, here is a more technical take on this…
The canonical data is available only in *.csv form, and cannot therefore be accessed by Picard online through the MB API.
However, I can see the possibility of someone writing a Picard Plugin that:
Downloads the CSV files as needed.
Loads the CSV files into an SQLite database (likely needed for performance reasons).
Provides conversion of Releases or NAT lookups from the Picard/MB matched one to the Canonical version. This might involve triggering the download of the canonical release from MB, and when the metadata download was complete, then moving the files over and removing the old release.
You would probably need to check that the canonical version had matching tracks for all the files you have (since e.g. a canonical version of a release can have less tracks than e.g. a special version which has extra CDs or extra tracks) and possibly check the length of the tracks also.
Also, whilst you might like this to happen automatically when you first load or lookup the files, you would have to guard against it happening when e.g. you manually select an alternative release from the context menu.
