EBU Subtitling data exchange format
The EBU (European Broadcasting Union) Subtitling data exchange format is a standardised format for exchanging subtitles between TV broadcast stations and other organisations.
The file name extension is STL. (The format is also specified to be exchanged on 3.5" floppy disks formatted for use with MS-DOS 3.3 and newer, but that can probably be disregarded.)
The format was defined in 1991 and has a number of idiosyncrasies due to the requirements and the time it was designed in. The format has a 1024 byte header (general information block) designed to be readable with a plain text editor, and after that follows text and timing information blocks coded in a binary format. The header block is defined to be stored in one of a short list of MS-DOS codepages, and the subtitle text data are defined to be stored in one of a short list of ISO text coding standards. There is no support for Unicode. The subtitle text can contain various formatting codes, depending on whether the data are specified to be for teletext captioning or for closed captioning.
Standard document
The standards document can be downloaded from the EBU: http://tech.ebu.ch/docs/tech/tech3264.pdf
Thoughts on implementing support in Aegisub
Initially, a limited write support is likely the best option. This will allow users to use Aegisub to create subtitles that can be used with various applications used in professional broadcast media.
The main hurdle for an initial write support is making sure the user can pick only the allowed text encodings/codepages, and possibly allowing the user to pick both the GSI and TTI encodings separately. That would require new UI.
Another hurdle to write support is the fact that the format is based around a TV screen with a limited number of lines, and a limited number of character cells per line, and we would have to hard break subtitles to fit the screen. This is made simpler by the fact that the format doesn't support Unicode, and so several assumptions can be made about line breaking.
Some of the advanced formatting supported by SSA and ASS can be translated into EBU STL, to a degree depending on whether teletext or CC format is being used. Formatting that could possibly be converted includes emphasis, colours and alignment. Comment lines can also be retained as comments.
Unclear information in the standard
One thing that isn't entirely clear from the standard is what the timecode format is, when Disk Format Code STL30.01 is selected, whether it's clean 30 fps, or whether it's actually 29.976 drop-frame. This would need to be confirmed by a user of the format.
