support better metadata

[ArjixWasTaken]

I noticed that someone updated the lrc for Otonoke to have both japanese/romaji/english, this kinda sucks.
I get why they did that, but it would be preferable to have separate lrc files for each language.

The reason I say it sucks, is because this would conflict with automatic romanization, e.g.
th-ch/youtube-music#2790

It would be nice if extra metadata was attached to indicate if the lrc is not pure and has modifications/translations to it.

[ArjixWasTaken]

An alternative solution would be to make your own lrc format for variants.

e.g.

[00:01.00] 必死で這い出た先で霧は晴れた
[00:01.00][v:jp_ro] Hisshi de hai deta saki de kiri wa hareta
[00:01.00][v:en] The fog lifted as I crawled desperately out

[ArjixWasTaken]

Possibly related to #22

[tranxuanthang]

lrclib.net follows MXM guidelines, one of which is to keep CJK lyrics in their original script rather than transcribing them into a romanized version.

If you come across lyrics like this, feel free to overwrite them to ensure they are correct. To make this easier, you can use the new LRCLIB browser available in LRCGET v0.9.0 and later.

I admit I haven't made enough effort to communicate this clearly, so many lyrics contributors don't get this. We definitely need to create a contribution guidelines page as soon as possible to prevent issues like this from spreading further in the lrclib.net database.