[linux-dvb] Re: Question about PTS in TS

[Nicola Sabbi <sabbi@students.cs.unibo.it>]

On Wed, 14 May 2003, Gregor Lawatscheck wrote:

> At 17:23 14/05/2003, you wrote:
> >Hi,
> >
> >often when parsing a TS coming from DVB drivers, I read an audio PTS
> >roughly 4 seconds lower than the corresponding video PTS closest to it.
> 
> This is exactly what I am seeing here and it appears to make problems with 
> demuxing some channels with new drivers / firmware on a av7110.
> 
> Are the streams you're seeing on this statmuxed (using VBR?) by any chance?
> 
> http://members.aol.com/mpucoder/DVD/pes-hdr.html might be helpful. In fact 
> I'm looking at it myself at the moment to work out what pvastrumento fixes 
> when it sets so called "VBR flags" to the PES header.
> 
> It's almost as if in the original header it indicates that there were no 
> PTS when there actually is. It's almost as if ts2pes (code of which is 
> employed in almostt all other apps outputting pes) writes b0rked headers or 
> something - but too early to blame anything (and I tend to be wrong when I 
> do anyway heh).
> 
> - Gregor 
> 

I don't know if the stream was stat-muxed. I read data from the dvr after
having set two PES_FILTER
but it's a problem present in almost all TS samples I have, as Arpi wrote
before me.

And I forgot to mention that sometimes I find a regressing PTS belonging
to the same PID!

Don't the drivers do a remuxing from pes to ts when feeding the dvr?



Nico



-- 
Info:
To unsubscribe send a mail to ecartis@linuxtv.org with "unsubscribe linux-dvb" as subject.