Kanji not being romanized and wrong romanization

[EstebanOD]

I noticed a few times, a resilient kanji does not get romanized; for example in the last sentence of this song, but also this one at 02:25.

But sometimes it's due to other reasons: In this song at 01:09 "つれてって" is romanized 'tsure te tsu te', because つ is tsu but っ indicates repetition so it isn't 'tsure te tsu te' but "tsutre te tte" Sometimes it's the particles for example "は" which is written with the 'ha' hiragana but should be pronounced "wa" and this isn't something a hiragana-latin converter can guess, same for o e and other particles.

Sometimes there's the issue of kanjis having multiple meanings therefore multiple ways of being pronounced and this means they can't be romanized without understanding which word the kanji corresponds to.

I know how difficult it is to romanize Japanese, and the extension does it pretty well 99% of the time, but to have a perfect romanization the only way would be to use a tool that knows Japanese, for example, Google Translate knows Japanese so when it displays romanization it's accurate (but sending the full lyrics to Google Translate just to get back the romanization would take much more longer than the time taken by this extension). If you visit this link and then do CTRL+F to search "tsuretette" you will see Google was able to understand it's not a hiragana used to add a new syllable but to change the following "te".

I know it would be incredibly hard to implement anything like that to a spotify extension, but I ought to let you know about this minor issue. Hope you have a nice day!