It seems like it's doing this because there's rich text tags inbetween each and every Japanese character.
It's able to see each and every character that's coming out of a dialogue box, but it can't piece it all together as a single sentence.
So it's reading "粉雪の代わりに、柔らかい日差しが大地を覆う。" as individual words like "粉" "雪" "の" "代" "わ", so on and so forth because of the <alpha=#FF> between all of them.
XUnity's not exactly built for that and apparently making it handle such a case that isn't just a flurry of <alpha=#FF>'s
You must be registered to see the links
View attachment 4208280