Well I’m not sure it keeps me up at night, but since working on music industry databases years ago, I have put a lot of thought into how best to make iTunes work for me and now, how to integrate that with Plex.
In the case you mention, I use ‘Tom PETTY’ for Album Artist. ‘PETTY•Tom’ for SortAlbumArtist and whatever is appropriate for the Artist. So in some cases that will be ‘Tom PETTY & THE HEARTBREAKERS’ as Artist and ‘PETTY & THE HEARTBREAKERS•Tom’ as SortArtist, but in others they will be blank, meaning it’s the same as the AlbumArtist. No need to store it twice as that always makes original entry and any later updating more work. Golden rule of database design - don’t duplicate data. In any case, I rely on what is specified on the album cover as to the artist for each track. Some artists it has varied over time the name they use on an Album. Not sure about Tom Petty, but an album could be just by Tom Petty, but others might be …& The Heartbreakers. I prefer to keep them together, so in such cases I use the name of the actual artist alone as the AlbumArtist because that’s shorter and easier to spot in the list.
I use upper case to denote where to sort and have a script that creates the correct Sort version according to the case used in the main tag. If the Sort version would be the same (or differs just by 'The ’ or 'A ’ or 'An ’ at the beginning) I leave it blank as iTunes and Plex will just use the basic tag for sorting (after stripping off 'The ’ etc).
Out of the nearly 600 artists in the latest import I tried, I have not found a single anomaly. Plex has done a great job.
