Because of the way those captions are stored VLC has to use OCR to convert the .SRT file (which basically stores low resolution b/w images I assume to easier allow for different alphabets) to normal text. I don’t know why the open source solutions are so bad at this (especially considering how good the proprietary solutions seem to be) but I had similar problems ripping a DVD. I would assume that had he turned off the special font VLC uses for the subtitles and instead just seen the raw data there wouldn’t have been a problem. Why VLC doesn’t enable this by default (/ have this) I don’t know.