I’m just going to offer one other final (yes! haha) piece of feedback, based on yet more reading of the docs and further experimentation.
I couldn’t work out why, having manged to perform a simple
scrub on an album, I would reopen said album in a GUI editor and notice that while all the id3v1 tags had gone, at the same time I suddenly had included loads and loads of ‘default’ tags such as
Arranger, Original Date, Composer, BPM, etc. but NONE of which - I repeat NONE of which were included on the files before the
Ah well I said… maybe the database it built isn’t what I thought it was. It may well be wiping and rewriting tags from its .db file, but that isn’t information that was included in the files in the first place.
So I decided to just delete the database and
state.pickle file, and begin again with a new import, and this time I would allow Beets to build its own database based on its advanced heuristics (I read that page - impressive).
I’ve now given up completely. Why? I collect quite a lot of African, ME and Latin music, some of it on compilations, some of it ripped from my extensive CD collection that began years ago etc. and much of which has been sourced abroad.
Off went beet, busily scanning (I used the recommended options including
-ql) and it stopped at about the third album in with some options for me. Long story short - opening the musicbrainz webpage allowed me to search manually for those same release, artist that beet was asking me about but not offering, only to discover that probably less than half the albums just in that one directory (for Africa) aren’t even in Musicbrainz. Sometimes the artist’s name appears, but in context of one track in a compilation. By the time I’d used the
-U option to just ‘leave it as it is’ I felt I was defeating the point of the entire exercise.
My point here is this is clearly excellent software, carefully thought out. But I don’t have particularly obscure or difficult tastes yet it cannot make even a 40% match with the first couple of dozen albums it comes across due to the fact that Musicbrainz, while good, is very ‘euro-centric’ and doesn’t cater at all well for non English releases. For me to have to babysit the creation of a database of close to 5000 albums, probably half of which is going to stop and ask me for options, just isn’t realistic I’m afraid. And since I haven’t been able to work out how/if I can simply edit a database created from a pure import read without any auto-tagging I think I’m going to wish you well with this excellent project and uninstall it. At the end of the day I could have been halfway through my collection by now using manual editing - the time it’s taken to learn, test and discover to be not performing as hoped.
I may just go back to plain old
find and some regex
Thanks for all your helpful replies thus far.