[mythtv] Closed Captioning & Teletext/Videotext

[Erik Arendse]

At 15-1-03 07:55, Henk Poley wrote:
> > I am busy implementing Closed Captioning support and support for
> > European Teletext/VideoText (used for 1000 sorts of info, including
> > Closed Captioning).
>
>Actually max. 899 sorts of things ;-)
>
>Pages lower than 100 and higher than 999 don't exist.
You forget sub-pages... :-)

> > The data will probably be integrated into the recorded file into a newly
> > added frame, you could generate a searchlist from a recorded file or
> > search for strings inside the frames themselves.
>
>The question would be, can you re-generate the Teletext (CeeFax) with your
>Video-Out?
Not possible with any driver I've seen. Not sure if lowlevel hacking would 
be sufficient, or that the PAL/NTSC encoding chip would get in the way.
But not an interesting question for me, MythTV will just show subtitles 
itself, with the added benefit of a transparent background instead of a 
surrounding black box. This way all users with a external PAL/NTSC 
converter will be happy as well.

 > Other possibility would be to bitmask or alphablend it to the displayed 
videostream, but that would need
 > a Teletext viewer/interpreter.
Trivial and thus included. AleVT's only function _is_ teletext 
interpreting, and the mythtv OSD was only a small class away from a 
teletext viewer. Still looking or a nice truetype teletext font by the way, 
bitmapped ones galore...

> > This way CC can be turned on/off (and selected if more than one
> > language) afterwards on timeshifted TV and recorded shows.
> >
> > As a by-product mythtv will have 100% buffered teletext/videotext
> > support from the current selected channel, not timeshifted, but that
> > would be a bit too much effort only to display a timeshifted clock :-)
>Uh, how are you going to sync subtitles (?Closed Captioning?) then?
Close Captioning is separate from the rest of my engine, it's just that as 
I included the teletext decoder it  will be only so much work to show all 
pages. Not fully implemented yet, but it will be.
CC is added to the recorded file in a new type of CC-frame, it's only a few 
lines every few seconds. Adding Teletext to the recorded file would be a 
load of data for no use (why would anyone read shifted or recorded teletext?)
The Teletext decoder is just one way of feeding CC data to the recorded 
stream, I presume somebody on the other side of the pond will add a US-CC 
decoder and feed the outcoming text the same way.

> > I use AleVT as engine for the actual VBI decoding, if somebody wants to
> > strip the zvbi lib to include US based CC with the same engine perhaps
> > that's better but I don't like the lib (far too bulky).
>Ah, that probably is a Teletext viewer?
AleVT is small and efficient teletext decoder & viewer, from which I use 
the decoder engine. The zvlib is a bulky universal vbi decoder, supposedly 
incorporating teletext, vps, american CC, and such.
I don't like it and can't test US-CC anyway, so I used AleVT. It would be 
rather trivial to upgrade to a stripped zvlib if somebody likes to hack & 
strip it's internals. Maybe we could even link directly to it, would be 
probably better as AleVT is not a lib (well, until I ran into it anyway :-)

I'll post a note if somebody want to help with US-CC.

Erik