I had observed this first a few weeks ago, expecting a transient mess-up, but it is still present.
Looking at Firefox's page-info (see file Page-info.png)the encoding is stated as UTF-8. However, whereever Umlaute (and other special chracters) show up, the text is sprinkled with black triangles with a white question mark in in (see files umlaute.png and umlaute2.png).
This propagates into xbmc/kodi (see file xbmc-umlaute.png). Here the "Kln" should be "Köln", and the "Eisbr" should be an "Eisbär".
Please, review and correct the encoding problem.