[MPlayer-dev-eng] [PATCH] Automate mov to mpegps demuxer switch

[John Koleszar]

Nico Sabbi wrote:
> John Koleszar wrote:
> 
>> Nico Sabbi wrote:
>>  
>>
>>> unconditionally searching an mpegps stream as a fallback  when no audio
>>> and no video
>>> streams were found is wrong. Mpeg in mov is qualified by appropriate
>>> identifiers
>>> in the atoms (that thread should describe them).
>>>
>>>   
>>
>> Ok. I didn't recognize the 1ba and m1s atoms identified in the other
>> thread as
>> being QT atoms, and was under the impression you were suggesting
>> searching the
>> payload to try and identify it.. Is keying off the media handler atom
>> (mhlr)
>> subtype (MPEG) sufficient?
>>
>> MOV:   Handler header: mhlr/MPEG (appl) Apple MPEG Media Handler
>>
>>  
>>
> 
> I think so, but when dealing with MOV who can be sure of anything?
> my discussion in irc with the author of the other patch leads me to believe
> that the presence of that atom is enough

Here's another cut at it, this time checking for a track with its media handler
set to MPEG.

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: mov_mpegps_demux.diff
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20060803/4c82f401/attachment.txt>