[mythtv-users] Why do those mpeg files not play? (capture with a ceton infiniTV 4)

[Daniel Kristjansson]

On Sat, 2011-07-09 at 16:08 -0400, Ronald Frazier wrote:
> > but it does and should send the full tables through unmolested. It's
> > up to the recorder to trim those down so that they only describe the
> > recorded program.
> 
> You say the device SHOULD send the full tables, but the thing is that
> the HDHomeRun doesn't seem to behave that way. If I tell it (via
> hdhomerun_config) to select a channel and program, and then use netcat
> to dump the data directly into a file, the HDHR rewrites the PAT to
> remove all but the channel I'm interested in, and it removes any PMTs
> (and packets from their associated PIDs) for any program I didn't
> tune. So according to your suggestion, then the HDHR is currently
> doing it the wrong way.

The HDHomeRun is a special case, it actually has two modes. One where
it passes the tables unmolested, and one where it tries it's best to
give you a ready to play file. The second mode was added later as an
additional feature and is possible because the device has a fairly
powerful processor on board. The second mode is useful mostly for
debugging since you can just feed the stream directly into a player.
But if it were the only mode it would be difficult to match up 
additional metadata like EIT with channels.

> > MythTV has a class for doing this for you called MPEGStreamData
> > you just need to set the desired program either in the constructor
> 
> Yeah, I figured this out already. As I mentioned in my other email
> yesterday (in the other Ceton thread), I was stumbling over this issue
> for several days because there were actually 2 different problems I
> was dealing with, but I didn't realize it at first. The other issue
> was that there is apparently a bug in the MPEGStreamData class. It
> must not be a common occurrence, since nobody seemed to notice the bug
> before, but the mpeg-ts stream that's coming from my cable provider is
> tripping it up. The bug appears to also be present in trunk code, so
> I'll be opening a ticket for it soon.

Cool, it would be great if you could share the code you have so far.

-- Daniel