Why w*fstream classes converts unicode characters to single byte before storing? As it well known, Standard Template Library has wstring type, which operates with the unicode characters. Also there is the corresponding steams types, such as for example wofstream. When I use the code below, resulting text file contains single byte characters (ASCII), i.e. the stream converts string from multibyte presentation to single byte. Why? I am absolutely sure that wstring contains natural unicode symbols. std::wofstream outfile ("somefile.txt") ; std::wstring str1 = L"Some string" ; //L translates char to wide char (MSVC) outfile << str1 ; outfile.close() ;