All text in a Matroska™ file is encoded in UTF-8. This means that mkvmerge(1) has to convert every text file it reads as well as every text given on the command line from one character set into UTF-8. In return this also means that mkvmerge(1)'s output has to be converted back to that character set from UTF-8, e.g. if a non-English translation is used with --ui-language or for text originating from a Matroska™ file
以上是 mkvtoolnix 的 man page 說明,也是為什麼無法撥放字幕的原因,merge 字幕的時候還要加上 --ui-language [code] 的參數,讓解碼器可以由UTF-8轉回你能支援的編碼格式。