[gst-devel] mpeg transport stream "stuck"

[Dan Taylor]

More new data:

Tuning live ATSC, I see that the elemental streams within a
multi-program
transport stream have, or do not have, PES start flags "randomly".  I
left a
program playing overnight, and when I stopped/restarted it, there was
nothing.  Checking the stream, I found that the PES start flags were no
longer present in its streams.  Several times today, I have had PES
start
flags come and go from various elemental streams within the same
transport
stream.

Also, it appears that if there ever was a PES start flag in a program's
elemental streams, that the program will continue indefinitely, whether
the PES start flags continue to be supplied, or no.


-----Original Message-----
From: Zaheer Merali [mailto:zaheermerali at gmail.com] 
Sent: Tuesday, October 14, 2008 7:08 AM
To: Discussion of the development of GStreamer
Subject: Re: [gst-devel] mpeg transport stream "stuck"

On Tue, Oct 14, 2008 at 3:35 AM, Dan Taylor <dtaylor at startrac.com>
wrote:
>> The top-post is because I have new data.
>>
>> In my captured stream are several programs.  I found one that plays.
>>
>> The difference between the one that plays and the ones that do not
play
>> is that the one that plays has PES start flags in some (or all) of
the
>> headers, and the ones that do not play have NO PES start flags
anywhere
>> in the stream.
>>
>> "mplayer" can play (at some level) all of the programs.
>>
>> Is the lack of PES start flags (found by reading through a debug dump
>> of the entire stream) something that should be "handled" by
flutsdemux
>> or some other part of the GStreamer pipeline?

> Hi Dan,
> 
> Could you provide me say a 50-100MB snippet of the transport stream
you have.

> In response to your previous section, the transport stream clock is
> only used when it is a live pipeline which this is not.

> Zaheer

File sent.  Hope it has some info you need.

Thanks.