The Japanese are quite used to mojibake [1], so they would've understood immediately that the mismatch between your ticket and passport was caused by encoding issues.
Interestingly, I've had problems in Korea (Gimpo Airport) because my name contains an "ö", and the canonical spelling in the passport for this is "oe". This was cause for much confusion among the airport staff.
I would have thought that people from CJK-countries were more understanding of encoding-to-latin weirdness than most, but apparently not.
I think their understanding would be focused on the encoding for their language and a relatively narrow set of problems. I've encountered name issues in CJK countries that keep names in native encoding due to an assumption that full names fit within a couple of characters with no need for any spaces or punctuation. Some systems might be designed to be "accommodating" and take even up to 8 or 10 characters! There was one train system where my name had at least four different iterations through the tickets I collected, with different ordering of first and last names and truncating.
In defense of the Korean airport staff, they might have been more accommodating if the "ö" was completely and obviously broken, like "£‡�". Spelling it as "oe" makes it look like there are no encoding issues, in which case strict checking makes more sense.
It's much easier to identify mojibake (they tend to be extremely obvious in CJK encodings) than to remember canonical spellings and other variations in a whole bunch of different languages. Airport staff probably know that "oe" and "œ" are interchangeable, but that's about it.
I would have thought that people from CJK-countries were more understanding of encoding-to-latin weirdness than most, but apparently not.