[mythtv] OpenCable stream type & #2556

Daniel Kristjansson danielk at cuymedia.net
Mon Oct 16 13:35:30 UTC 2006


On Mon, 2006-10-16 at 09:35 +0100, Stuart Auchterlonie wrote:
> On Sun, Oct 15, 2006 at 07:05:00PM -0400, Daniel Kristjansson wrote:
> > On Sun, 2006-10-15 at 17:06 +0000, MythTV wrote:
> > > #2579: Make libavformat recognize OpenCable video streams in MPEG-TS files
> > > 
> > >  If the opencable video stream identifier is actually 0x80 and not 0x128
> > >  then this solves two problems in one go.
> > But, I don't think 0x128 is actually a valid stream id, you are
> > probably seeing this as a result of broken VERBOSE formatting.
> > In that case 128 == 0x80, and the problem is not solved yet.
> > We do store the SI table type when we scan for channels, so we
> > could ignore the OpenCable(R) stream id's whenever "dvb" is the
> > SI type...
> Yep it is broken VERBOSE(). I'll fix that shortly.

> One way to handle #2556 might be to ignore alternative video
> streams once we have found a valid stream. Perhaps using
> the stream type as a preference.
I think we should only remap 0x80 to 0x02 if the sistandard is "atsc",
a longer term fix is to have the channel scanner set the sistandard
type to "opencable" for OpenCable transports and only do the remap
when the sistandard is "opencable". Hopefully transports won't mix
sistandards, if they do we will have to move this column from the
dtv_multiplex table to the channel table.

> The PMT debugging printout says that it'll use the first video
> stream found, which is contrary to what we do. Is there a situation
> in which things would break if we chose the video stream in this
> way?
I don't know if this would break MHEG. But this debugging was written
when we only supported a single video stream. We now partially support
multiple video streams for MHEG, and in the future we will support
it for ATSC and OpenCable Interactive TV applications.

-- Daniel



More information about the mythtv-dev mailing list