Autopopulation of ARTISTSORT / ALBUMARTISTSORT tags
Most file formats allow for an artist sort (and album artist sort) tag which music apps can use to organise tracks.
E.g., I have:
.. in my ogg files.
It would be good to force these tags (data might be extracted from musicbrainz).
Thanks for that detective work! You're right - blindly using this would be problematic.
In your first comment in 2014 you pointed out that MusicBrainz does not have a Sort Name every artist. As you can clearly see from the following search, this is still a significant problem:
There is a now a Sort Name for each artist, but for a significant number of artists in this test search the value is just plain wrong, it's simply the artist name. So automatically setting ARTISTSORT from MusicBrainz would lead to a lot of duplication (all the artists where Sort Name is actually the same as artist name), and erroneous data in the cases where Sort Name would actually be different.
If Bliss were to check with MusicBrainz and point out any differences between Sort Name and ARTISTSORT in the local file, nothing is gained, because those differences might be exactly because the data on MusicBrainz is wrong. You'd have to fix it in MusicBrains first.
Since there's no way to fix this automatically, I suggest autopopulation of ARTISTSORT/ALBUMARTISTSORT from Musicbrains if and only if Sort Name is different from artist name, and keeping it empty otherwise, so that we don't end up with erroneous data. Correctness trumps consistency, doesn't it?
For that to be of use Bliss might introduce a computed tag "RELEASE ARTIST" which evaluates tags in the following order and takes the first one that is not empty
* ARTIST SORT
Note that some online sources have sort names, e.g. https://musicbrainz.org/artist/d5da1841-9bc8-4813-9f89-11098090148e
I doubt that you could have an automatic fallback, because I don't see how to prevent mangling a lot of band names erroneously along the way. But a manual fix would work here, and maybe even a selection of replacements using regex might make sense, because that would also have
to apply to "The Fall", and "A Flock of Seagulls", and "Die erste weibliche Fleischergesellin nach 1945".
To keep the number of manual fixes manageable, you'd have apply every manual fix to all instances of an artist's name in the collection automatically.
Yeah, this data is on MusicBrainz although that doesn't cover all artists. I wonder if we could have fallbacks to analyse the artist string and try to guess what the sort tags should be? E.g. two words without a comma: reverse them and insert a comma (depending on how you want to sort).