[vlc-devel] Re: TS mux generates stream which doesn't play smoothly on Amino 103 STBs

[Marian Durkovic]

Hi all,

  I did some testing on the broadcast TS streams, and it seems the 
FixPES() is trying to normalize sizes ranging between 10 and 700 bytes
into exactly 170-byte chunks.

  Before I get into further debugging in order to verify accuracy of
interpolated PTSs and/or testing wheter the BlockNew and BlockRealloc
operations don't cause processing delays for audio ES and thus
desynchronization with video ES, could someone please comment if it 
won't be easier to simply ommit FixPES and use audio ES as it was 
packetized by broadcast-quality TS mux at TV studio? 
Or if needed for some specific low-bitrate audio streams (as the changeset
10591 implies) shouldn't this be optional feature only?

	Thanks & kind regards,

		M.


On Tue, May 30, 2006 at 06:36:58PM +0200, Derk-Jan Hartman wrote:
> On 30-mei-2006, at 14:17, Gregor Fuis wrote:
> >Hello,
> >
> >I was testing it again and right now i can say that --sout-ts- 
> >alignment did the job properly!
> >Now i don't see any skips of frames on aminet and in cisco video  
> >analyzer i get proper audio buffer!
> 
> Well if everything is aligned, it's logical that it will work. I  
> wonder if Mans his analysis is of interest here. What is the A/V  
> interleaving desynchronization on your stream? can the cisco tool  
> show that ?
> 
> BTW just to get this correctly quoted:
> 
> 2.4.3.6 ISO-13818-1
> In the case of audio, if a PTS is present in PES packet header it  
> shall refer to the first access unit commencing
> in the PES packet. An audio access unit commences in a PES packet if  
> the first byte of the audio access unit is
> present in the PES packet
> 
> I wonder what the amino people think of this.
> 
> DJ
> 
> 
> >Derk-Jan Hartman wrote:
> >>Could you please try the attached patch ? Perhaps it's as simple  
> >>as that...
> >>
> >>DJ
> >>
> >>
> >>
> >>On 28-mei-2006, at 19:26, Marian Durkovic wrote:
> >>
> >>>On Sun, May 28, 2006 at 06:30:29PM +0200, Gregor Fuis wrote:
> >>>>I found picture that would explain it more, it is on
> >>>>http://freeweb.siol.net/gfuis/pic.htm .
> >>>>
> >>>>You will notice the audio buffer, it is empty, but if i remove  
> >>>>FixPES
> >>>>from vlc ts.c then audio buffer rises  to the same level as  
> >>>>video buffer
> >>>>is in this picture.
> >>>
> >>>In other words, it seems that PTSs in audio stream don't properly
> >>>corelate to the PTSs of the video stream - and as a result, there's
> >>>continual buffer underrun on audio ES.
> >>>
> >>>Please don't close ticket #669 until this is properly  
> >>>investigated & fixed.
> >>
> >
> 


--------------------------------------------------------------------------
----                                                                  ----
----   Marian Durkovic                       network  manager         ----
----                                                                  ----
----   Slovak Technical University           Tel: +421 2 524 51 301   ----
----   Computer Centre, Nam. Slobody 17      Fax: +421 2 524 94 351   ----
----   812 43 Bratislava, Slovak Republic    E-mail/sip: md at bts.sk    ----
----                                                                  ----
--------------------------------------------------------------------------

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html