Le 07/08/2014 18:09, Rémi Denis-Courmont a écrit : > Of course, that assumes that this is really a plain text and Unicode > subset encoding. Otherwise, none of your or my proposals work. > > Alas as far as I recall, ARIB is _not_ a subset of Unicode. > Or have a library, whatever it outputs, used by both demuxer and codec. Francois