[MPlayer-dev-eng] PATCH: Fixes seeking based on percentage for mpg's with timestamps not starting at 0

Nico Sabbi nsabbi at tiscali.it
Sat Apr 2 18:44:39 CEST 2005


Oded Shimon wrote:

>On Sat, Apr 02, 2005 at 06:22:35PM +0200, Nico Sabbi wrote:
>  
>
>>unless someone implements -idx (any volunteer?)
>>    
>>
>
>Hmm, I'm actually quite interested and I still have a little free time left.. 
>I know absoloutely nothing about MPEG's though (except playing them...), so 
>I'll probably need help.
>
>I think the main thing to be gained from this -idx is proper seeking in VBR 
>files (with today's demux_mpg, you can end up in almost 10x more seek than you 
>asked), a real and correct total time, and good percentage seeking...
>
>The best thing to possibly be gained from this is smarter MEncoder handling 
>with broken MPG files, where you have "blank spots" of audio.. they cause 
>MEncoder to desync and gradually remove frames and fix it.. With this, it 
>could just remove the whole chunk or make empty audio pakcets or similar... 
>But this is much later. :)
>
>I hope I'm up to the challenge, I'm gonna start reading demux_mpg.c .
>Where should I post my questions BTW?.. Usually I ask on IRC, but you're not 
>there...
>
>- ods15
>
>  
>

if you explain how to get to irc (that I never used) I will come; in any 
case you can ask
here or to me personally.
You shouldn't try to understand how mpeg works reading demux_mpg, 
because that demuxer
was written to be synthetic, not explanatory.
You can find plenty of 13818-1 drafts around.

    Nico




More information about the MPlayer-dev-eng mailing list