Well, I received a response to my MKVToolNix bug tracker ticket:

(for my FULL original post, please visit https://trac.bunkus.org/ticket/911)

moritz (MKVToolNix tech support) said:

The special characters are not a property of the language, but of the character set that's been used to create the file. All languages on this planet can actually be expressed in more than one character set: at least in UTF-8, UTF-16 and UTF-32. Then you also have all the legacy charset (e.g. ISO-8859-1 or Windows-Latin1, CP1450 etc. etc). You'll have to tell mkvmerge that as well. You can use mmg's "subtitle charset" option for it.

There is generally no way to auto-detect the character set used by text file. Programs can take guesses with some measure of certainty, but that's always prone to a lot of false positive -- which is why mkvmerge simply opts not to do that.

In you case:

•The files with "UTF-8" in their names are encoded in UTF-8 (duh!).

•The Polish file is encoded in UTF-8 as well.

•The French and Spanish files, however, are encoded in ISO-8859-15.

•The English file doesn't use any non-ASCII characters, therefore you don't have to select anything special for it.

So, what I take away from this is that for improved compatibility with MKVToolNix (and perhaps other applications) the French and Spanish subs would likely benefit from an ISO-8859-15 to UTF-8 conversion. A quick Google search, a few lines in Terminal under MacOS X (using the "iconv" command), and I just sent off (what appears to me to be) 100% working, MKVToolNix compatible, UTF-8 format French and Spanish versions of johnlocke2342 and Vali's subs to HanDuet for proper vetting…

That's about it for now! ;)

-A.S.