You just have to have a binary ID and maybe "pronunciation" fields too, for each song/album/artist entities in the database schema, that are on UI sanitized/nor

You just have to have a binary ID and maybe "pronunciation" fields too, for each song/album/artist entities in the database schema, that are on UI sanitized/normalized by searches and loose matches. It's not worth dealing with regularization and mapping rules between ASCII and whatever artists insist on using, like slightly different style of T used only in Western East Antarctic. It'll be just a String whatever. That way each files may keep whatever emojis used specifically for any specific distribution channels without troubles.