[mythtv] transform.c (was: DVB subtitles and TS->PES decoder)

[Jesper Sörensen]

Marcus Metzler wrote:

>    >> First problem I run into is that hte current ts->pes parser in
>    >> transform.c does not seem to parse subtitle streams, at least I
>    >> cannot get it to output video, audio and subtitle streams, it
>    >> seems to ignore the subtitle stream. Does anyone know if this
>    >> is possible at all or should I just use my own ts->pes parser
>    >> and replace the one in transform.c ?
>    >> 
>    >> 
>
>    Jesper> I think the plan is to eventually drop transform.c
>    Jesper> entirely and use a "stripped" (filtered) TS for DVB
>    Jesper> instead of a PS. I think that's the way it's done in the
>    Jesper> ATSC/HDTV recorder, and it would mean we could share some
>    Jesper> code between DVB/ATSC. Using a TS also has some other
>    Jesper> benefits, e.g. allowing the use of multiple audio audio
>    Jesper> streams (e.g. both PCM & AC3).
>
>I mean you could drop transport.c right now, most of the code in there
>is not needed (only about 3-4 routines). Except for section handling,
>there should be no difference between DVB and ATSC and for recording
>yuo don't even need to look into the TS packets, you just need to
>filter for the PIDs. Why are there two code bases for that.
>  
>

I don't think there is any good reason. I guess it's something that 
happens when there are many different chefs poking in the same pot. I 
don't know how big of a change it would be and what other parts of Myth 
would be affected but if someone who knows had the time to do it it 
would be great. :)

>By the way, you can also have more than one audio or video stream in a
>program stream. 
>

Ok, I guess it's only a limitation in the implementation in Myth then.

>That would be no reason to switch to TS
>recording. Much more compelling is the fact that the PS would have to
>be properly remuxed. The TS from digital TV sources is already muxed
>correctly.
>  
>

Yeah, I agree.