[MPlayer-dev-eng] ripping subtitles

[Arpi]

Hi,

> On Friday 14 June 2002 04:09, Kim Minh Kaplan wrote:
> > I am still not sure how subtitles ripping should be done.  For the
> > moment I have some pieces of code that does it inside mencoder (at the
> > same time that it is burning the subtitles inside the encoded movie).
> > But this does not feel right.  For one MPlayer's code can only read
> > one subtitle language at a time so it can only do one language
> > subtitle ripping at a time (it is vobsub ripping so it should be
> > possible to put multiple languages into a single file).  So here is
> > what I still can't answer: should subtitle ripping happen inside
> > mencoder or in a separate program?
> Hmm, Arpi talked about advancing the muxers/demuxers to be able to handle 
> multiple audio streams to be able to generate multiple language AVI.

yes

> I think it would be a great thing, if all demuxers can handle multiple audio
> streams as it opens a lot of interesting possibilities, like switching audio
> streams on the fly (DVD, Multi-language (S)VCD/AVI). Also switching subtitle
> streams on the fly would be cool.

yes

> To achieve this MPlayer/MEncoder IMHO needs some not-so-easy structural 
> changes (eg. otf audio reinit), which will surely happen sooner or later, bu
> t 
> I think not in shortterm.

why?

video is no problem, imho neither subtitles
audio will be ok when audiofilters are moved to libmpcodecs, ending by af_ao
wrapper, so audio init/uninit otf will be easy

> Btw. I'm interested in Arpi's thoughts/ideas about this.

yes

> But for now you could get started with only dumping one sub stream, it's sti
> ll 
> useful (eg. usually you only dump english or native language sub).

one or all - user option
or a list of streams, like -dumpdvdsub -slang hu,en,de would dump these 3

> You could even do very crazy things like encoding the speech of the movie in
> different languages at low-bitrate/different codec and having music/effects 
> at higher bitrate and at playback downmix the wanted dialogs to music/sfx 
> channel (this of course is only practical, if source media supplies seperate
> streams for dialogs/sfx etc, this approach is a bit familiar to mpeg4's idea

isit possible to separate speech if you have speech+fx and fx-only channels?
(but no speech-only one)

> of having video objects, so motion and backgrounds can be encoded 
> independently). Well I guess this will happen in the next few years anyways 
> =)


A'rpi / Astral & ESP-team

--
Developer of MPlayer, the Movie Player for Linux - http://www.MPlayerHQ.hu